Characterization and Clinical Relevance of ALDH Populations In

Published OnlineFirst August 22, 2013; DOI: 10.1158/1078-0432.CCR-12-2857

Clinical Cancer Human Cancer Biology Research

Characterization and Clinical Relevance of ALDHbright Populations in Prostate Cancer

Clementine Le Magnen1, Lukas Bubendorf3, Cyrill A. Rentsch2, Chantal Mengus1, Joel Gsponer3, Tobias Zellweger4, Malte Rieken1,2, George N. Thalmann5, Marco G. Cecchini5, Markus Germann5, Alexander Bachmann2, Stephen Wyler2, Michael Heberer1, and Giulio C. Spagnoli1

Abstract Purpose: High aldehyde dehydrogenase (ALDH) has been suggested to selectively mark cells with high tumorigenic potential in established prostate cancer cell lines. However, the existence of cells with high ALDH activity (ALDHbright) in primary prostate cancer specimens has not been shown so far. We investigated the presence, phenotype, and clinical significance of ALDHbright populations in clinical prostate cancer specimens. Experimental Design: We used ALDEFLUOR technology and fluorescence-activated cell-sorting (FACS) staining to identify and characterize ALDHbright populations in cells freshly isolated from clinical prostate cancer specimens. Expression of genes encoding ALDH-specific isoforms was evaluated by quantitative real- time PCR in normal prostate, benign prostatic hyperplasia (BPH), and prostate cancer tissues. ALDH1A1- specific expression and prognostic significance were assessed by staining two tissue microarrays that included more than 500 samples of BPH, prostatic intraepithelial neoplasia (PIN), and multistage prostate cancer. Results: ALDHbright cells were detectable in freshly excised prostate cancer specimens (n ¼ 39) and were þ þ mainly included within the EpCAM( ) and Trop2( ) cell populations. Although several ALDH isoforms were expressed to high extents in prostate cancer, only ALDH1A1 gene expression significantly correlated with ALDH activity (P < 0.01) and was increased in cancers with high Gleason scores (P ¼ 0.03). Most importantly, ALDH1A1 protein was expressed significantly more frequently and at higher levels in advanced-stage than in low-stage prostate cancer and BPH. Notably, ALDH1A1 positivity was associated with poor survival (P ¼ 0.02) in hormone-na€ve patients. Conclusions: Our data indicate that ALDH contributes to the identification of subsets of prostate cancer cells of potentially high clinical relevance. Clin Cancer Res; 1–11. 2013 AACR.

Introduction tiation states. However, the clinical significance and the Despite the availability of several therapeutic options, prognostic relevance of specific cancer cell subpopulations prostate cancer remains a leading cause of cancer-related are still unclear (2). death in men (1). Prostate cancer is characterized by Aldehyde dehydrogenase (ALDH) enzyme is responsible remarkable cellular heterogeneity and includes cells with for the oxidization of cellular aldehydes resulting in the different phenotypes, proliferative capacities, and differen- production of retinoic acid (3). Notably, ALDH has been shown to be involved in stem cell protection and differentiation, and high levels of ALDH activity have been found in Authors' Affiliations: 1ICFS, Departments of Surgery and Biomedicine, several stem cell populations (4, 5). Thus, high ALDH 2Department of Urology, 3Institute for Pathology, Basel University Hospital; activity has been used to select and identify normal hemato- 4Division of Urology, St Claraspital, Basel; and 5Department of Urology, University of Bern, Bern, Switzerland poietic stem cells (4, 6) and tumor-initiating cells (TIC) in hematopoietic malignancies (7). Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/). TICs, functionally defined as cells capable of initiating tumors in immunodeficient mice (8), have also been Current address for C. Le Magnen: Department of Urology, Herbert Irving identified in a variety of human solid tumors (8–11). In Comprehensive Cancer Center, Columbia University Medical Center, New York, NY 10032. a number of cancers of diverse histologic origin, TICs have been reported to exhibit high levels of ALDH activity Corresponding Author: Clementine Le Magnen, Basel University Hospi- tal, ICFS, Departments of Surgery and Biomedicine, Laboratory 401, (5, 12–14). Hebelstrasse 20, CH-4031 Basel, Switzerland. Phone: 41612652378; Fax: In the prostate, high ALDH activity has been shown to 41612653090; E-mail: [email protected] represent a functional marker for murine normal progen- doi: 10.1158/1078-0432.CCR-12-2857 itor/stem cells (15). More recently, cells exhibiting high bright 2013 American Association for Cancer Research. ALDH activity, referred to as "ALDH " cells, have been

www.aacrjournals.org OF1

Le Magnen et al.

referred for treatment at the Department of Urology of the Translational Relevance University Hospital of Basel (Switzerland) from 2008 to High aldehyde dehydrogenase (ALDH) activity has 2012 was studied. Patients with BPH underwent conven- been shown to be associated with tumorigenesis and tional transurethral resection (TUR-P), whereas patients proposed to represent a functional marker for tumor- with prostate cancer underwent radical prostatectomy (RP). initiating cells (TIC) in various tumor types, including Written informed consent was obtained from all patients in prostate cancer. As previous studies were mainly based accordance with the requirements of the local Ethical Com- on the use of established cell lines, clinical relevance of mittee (EKBB, Ref.Nr.EK: 176/07). Clinical and pathologic bright ALDH populations has not been investigated in data of the patients with prostate cancer included in the detail so far in prostate cancer. Here, using a large cohort present study are summarized in Supplementary Table S1. of multistage prostate cancer specimens, we show that bright ALDH populations are present and heterogeneously Isolation of primary cells freshly derived from prostate distributed in prostate cancer tissues. Furthermore, we cancer surgical specimens show that expression of the ALDH1A1-specific isoform, Prostate specimens were examined for the presence of at both the gene and protein levels, is associated with cancer tissues by experienced pathologists. Prostate cancer advanced clinical stage and unfavorable prognosis in samples were chopped, washed, and digested in a mixture hormone-na€ve prostate cancer. Our findings highlight containing Dulbecco’s Modified Eagle’s Medium (DMEM), the potential importance of ALDH in prostate cancer 5% knockout serum replacement (KO serum, Gibco), 1% pathogenesis and suggest that its clinical significance Pen/Strep, and 200 IU/ml of type I collagenase (Worthing- might be specific for patient subgroups. These results ton). After an overnight incubation at 37C, digested tissues might help improve stratification and identification of were washed and centrifuged as previously described to high-risk patients with prostate cancer. separate the epithelial and stromal fractions (22). Cell pellets enriched in the epithelial fraction were resuspended in PBS and passed through a 100 mm cell strainer to obtain single-cell suspensions. Cells were immediately used for fluorescence-activated cell-sorting (FACS) analysis to assess identified in human prostate cancer cell lines and expanded ALDH activity and surface markers expression (see below). primary prostate cancer cultures (16). ALDHbright cells isolated from established prostate cancer cell lines were Quantification of gene expression by quantitative found to be associated with increased clonogenicity, inva- real-time PCR siveness, as well as high tumorigenic and metastatic capac- A normal human prostate RNA pool was purchased from ities (16, 17). However, culture conditions might modulate Clontech (Clontech Laboratories Inc.). cell characteristics, potentially favoring selection of specific Total cellular RNA was extracted from tissues, and DNase phenotypes, and prostate cancer cell lines may therefore be treated by using NucleoSpin RNA II (Macherey-Nagel). RNA inadequate to reflect the biology of human prostate cancer was reverse-transcribed by using Moloney Murine Leukemia (18–21). Virus Reverse Transcriptase (M-MLV RT, Invitrogen). Quan- To date, there is no evidence of the existence and func- bright titative real-time PCR (qRT-PCR) assays were conducted and tional characterization of ALDH populations in uncul- specific gene expression was analyzed as previously described tured clinical prostate cancer specimens. In addition, it is (23). TMPRSS2-ERG primers and probe sequences were still unclear whether such cells have clinical and prognostic derived from existing literature (24). Primers and probes for relevance in prostate cancer. ALDH1A1, ALDH1A3, ALDH3A1, ALDH4A1, ALDH7A1, In this study, we have identified, quantified, and charac- ALDH9A1 ALDH18A1 bright ,and specific isoforms were provided terized ALDH populations in cells isolated from freshly by Assays-on-Demand, Gene Expression Products (Applied excised clinical prostate cancer specimens. Our results indi- Biosystems). cate that ALDHbright subsets are detectable to various extents in all clinical prostate cancer samples. Moreover, expression Identification and isolation of cells with high ALDH ALDH of several specific isoforms was increased in cancers activity (ALDHbright cells) as compared with benign samples. ALDH activity was assessed by using the ALDEFLUOR Furthermore, we have assessed localization and clinical Assay System (StemCell Technologies) according to the relevance of cells expressing ALDH1A1 isoform using 2 manufacturer’s recommendations. Single cells obtained distinct tissue microarrays (TMA). Our results support an from prostate cancer specimens were resuspended in association between ALDH1A1 positivity and poor prog- ALDEFLUOR buffer and incubated with the ALDH substrate, nosis in prostate cancer. biodipy-aminoacetaldehyde (BAAA). As negative control, an aliquot of the treated cells was also incubated with the ALDH Materials and Methods inhibitor, diethylaminobenzaldehyde (DEAB). Following 35 Clinical specimens to 40 minutes of incubation at 37C, cells were washed and A series of 38 patients with benign prostatic hyperplasia analyzed using a dual laser BD FACS Calibur (BD Bio- (BPH) and 71 patients with pT2a-pT3b stage prostate cancer sciences). Dead cells were excluded on the basis ofpropidium

OF2 Clin Cancer Res; 2013 Clinical Cancer Research

ALDH Activity in Prostate Cancer

Table 1. Immunohistochemical analysis of ALDH1A1 expression on TMAs

Total interpretable Negative Positive n (%) n (%) n (%) Average score Progression TMA BPH 54 (100%) 52 (96%) 2 (4%) 0.3 PIN 38 (100%) 35 (92%) 3 (8%) 8.2 T1a/b 49 (100%) 47 (96%) 2 (4%) 9.0 RP (pT2-pT3b) 69 (100%) 57 (83%) 12 (17%) 12.4 CR 80 (100%) 65 (82%) 15 (18%) 18.3 Castration resistance TMA BPH 11 (100%) 10 (91%) 1 (9%) 1.7 HN 100 (100%) 60 (60%) 40 (40%) 40.5 CR 107 (100%) 67 (63%) 40 (37%) 33.7 iodide (PI) incorporation. Results are presented as percen- (0, 1, 2, or 3) by the percentage of positive cells, leading to tages of ALDHbright cells or as ratios of mean fluorescence an H-score ranging from 0 to 300 (25). A score greater than 0 intensities (MFI) as compared with control cells incubated was considered positive. Stainings were independently eval- with DEAB. uated by 2 experienced members of the team (L. Bubendorf and J.R. Gsponer) with excellent correlation between paired Phenotypic characterization of ALDHbright cells measurements. Following treatment with ALDEFLUOR assay system, cells were resuspended in buffer and incubated with phycoery- Statistical analysis thrin (PE)-labeled anti-CD44 (BD Biosciences), allophyco- Statistical analyses were conducted using GraphPad cyanin (APC)-labeled anti-EpCAM (BD Biosciences), and Prism5.0 (GraphPad Software Inc.) and SPSS softwares t APC-labeled anti-Trop2 (R&D Systems) antibodies or iso- (IBM). To assess the equality of means, parametric test or type-matched immunoglobulins, at concentrations recom- nonparametric Mann–Whitney test were used. Correlation mended by the manufacturers. of specific expression between 2 groups was assessed using Pearson or Spearman correlation tests. Fisher and c2 tests Tissue microarrays (TMAs) were used to compare ALDH1A1 positivity frequency in The 2 prostate TMAs used in this study were constructed different groups of patients. Survival curves were con- as previously described (25). The "progression TMA" constructed according to Kaplan–Meier and compared using P < tained single tissue cores from prostate cancers from all log-rank (Mantel–Cox) tests. 0.05 was considered sta- stages (25). The "castration-resistance TMA" has been tistically significant. recently described (26) and addresses the progression from hormone-na€ve (HN) to lethal castration-resistant (CR) Results prostate cancer (27). Characteristics and number of samples Epithelial cells freshly isolated from prostate cancer included in the 2 TMAs are given in Table 1. clinical specimens contain a heterogeneous ALDHbright population Immunohistochemistry Presence of ALDHbright cells was investigated in cells Immunohistochemical analyses were conducted accord- freshly isolated from 39 prostate cancer specimens directly ing to standard indirect immunoperoxidase procedures as retrieved following surgery. Representative cytograms are previously reported (25). Expression of ALDH1A1 in TMAs shown in Fig. 1A. As shown in Fig. 1B, ALDHbright popula- was assessed using a specific rabbit polyclonal antibody tions could be detected in all cell preparations obtained (ab51028, Abcam) at 1:200 (pretreatment: heat-mediated from clinical prostate cancer samples. Percentages of antigen retrieval at 100C). Other immunohistochemical ALDHbright cells were highly variable and differed from procedures are described in detail in the Supplementary patient to patient (average SE ¼ 1.56% 0.24% bright Information. cells; MFI ratio ¼ 257.2 40.87; n ¼ 39). When gated on þ EpCAM( ) cells to identify the bulk of the epithelial pop- TMA analysis ulation, percentages of ALDHbright cells were significantly þ Staining was visually scored and stratified into 4 groups: higher and exceeded 15% of EpCAM( ) cells in several negative (absence of staining), weak (weak but distinct specimens (average SE ¼ 9.06% 0.97% bright cells; immunoreactivity), moderate (between weak and strong), MFI ratio ¼ 292.3 32.99; n ¼ 31). As shown in Fig. 1C, a and strong intensity (apparent even at low magnification: trend toward a higher percentage of ALDHbright cells in high 2.5 objective). As previously described, a histoscore (H- (G8-9) as compared with medium (G7) Gleason score was score) was calculated by multiplying the staining intensity evident (P ¼ 0.09).

www.aacrjournals.org Clin Cancer Res; 2013 OF3

Le Magnen et al.

A All live cells EpCam (+)

+ DEAB – DEAB + DEAB – DEAB EpCam SSC SSC PCa patient 1

ALDH activity ALDH activity ALDH activity EpCam SSC SSC PCa patient 2 ALDH activity ALDH activity ALDH activity EpCam SSC SSC PCa patient 3 ALDH activity ALDH activity ALDH activity

BC20 2,000 20 P = 0.09 1,500 15 1,000 15 600

10 10 400

5 200 5 MFI ratio to control within EpCAM(+) cells % of ALDH bright cells % of ALDH bright cells 0 0 0 All EpC(+) All EpC(+) G7 G8-9 n = 39 n = 31 n = 39 n = 31 n = 20 n = 7

Figure 1. Identification of ALDHbright populations in cells freshly isolated from prostate cancer (PCa) surgical specimens. A, flow cytometric analysis of cells from 3 representative patients with prostate cancer (#1–3). Single-cell suspensions were obtained after digestion of prostate cancer tissues. ALDH activity was tested on live cells (left) using the ALDEFLUOR technology. Cells were also tested for EpCam-specific expression to identify ALDHbright cells within the EpCam(þ) population (right). B, cell suspensions freshly derived from 39 prostate cancer specimens were assessed for ALDH activity ("All"). For 31 specimens, ALDH activity was also tested in the EpCam(þ) population ["EpC(þ)"]. Left, percentages of ALDHbright cells. Right, MFI ratios to the DEAB control. C, percentage of ALDHbright cells in the EpCam(þ) population in patients with Gleason 7 (G7) as compared with patients with high (G8-9) Gleason score. Four patients with Gleason 5–6 were not included in the analysis.

þ Phenotypic characterization of ALDHbright cells in for EpCAM (average SE ¼ 75.47% 4.06% of EpCAM( ) prostate cancer clinical specimens cells within the ALDHbright subset; Fig. 2A and B). Expres- The epithelial cell–speciﬁc marker EpCAM has been sion of Trop2, previously shown to be expressed by epithe- shown to be highly expressed in a majority of carcinomas lial prostate cells (30) and to be overexpressed in human including prostate cancer (28) and has also been proposed prostate cancer, was also tested (31, 32). Likewise, a high as TIC marker in a variety of solid tumors (29). Interestingly, proportion of ALDHbright cells was also positive for Trop2 þ the majority of the ALDHbright population was also positive (average SE ¼ 81.41% 4.94% Trop2( ) cells within the

OF4 Clin Cancer Res; 2013 Clinical Cancer Research

ALDH Activity in Prostate Cancer

AB 100 ALDH bright 80

20 EpCam Trop2 SSC the ALDH br population

% of positive cells within 0 ALDH activity ALDH activity ALDH activity EpC Trop2 n = 31 n = 7 C low med bright All cells ALDH bright 1 2

3 PCa 1 EpCam EpCam SSC 4

ALDH activity CD44 EpCam CD44 CD44

1 2 PCa 2 EpCam EpCam SSC

ALDH activity CD44 EpCam CD44 CD44

Figure 2. Characterization of ALDHbright populations derived from prostate cancer (PCa) surgical specimens. In prostate cancer tissues, ALDHbright subsets are mainly comprised within the EpCam(þ) and the Trop2(þ) populations. A, expression of EpCam and Trop2 markers in the ALDHbright population derived from one representative patient. B, percentages of EpCam-positive (n ¼ 31) and Trop2-positive (n ¼ 7) cells in the ALDHbright populations from a panel of prostate cancer samples. C, left, expression of CD44 or EpCam in cell populations with different levels of ALDH activity (blue, low; red, medium; green, bright). Right, expression of CD44 and EpCam in all cells and gated on ALDHbright cells. Gate 1: EpCamhighCD44(–); Gate 2: EpCamhighCD44(þ); Gate 3: EpCamlow/CD44high; Gate 4: EpCam(–)CD44high.

ALDHbright subset; Fig. 2B). To gain additional insights into lower gene expression in prostate cancer than in BPH (Fig. the phenotype of ALDHbright cells in prostate cancer, ALDE- 3A). In contrast, a trend toward higher expression of FLUOR-treated prostate cancer cells were co-stained ALDH1A1, ALDH4A1, and ALDH9A1 was observed in with antibodies recognizing EpCAM and CD44 (n ¼ 4). In prostate cancer as compared with BPH and normal tissues all cases, ALDHbright phenotype was heterogeneous, as (Supplementary Table S2 and Fig. 3A). The ALDH1A3, illustrated by 2 representative staining profiles in Fig. 2C. ALDH7A1, and ALDH18A1 isoforms were expressed at Two populations of EpCAMhigh cells were present within significantly higher levels in prostate cancer than in BPH the ALDHbright population, either positive or negative for specimens and normal tissues (Supplementary Table S2 CD44. In one particular prostate cancer case, an EpCAM( ) and Fig. 3A). The same trends were observed when selected CD44high subset was detected in the ALDHbright population. prostate cancer samples found to be positive for the tumor- specific TMPRSS2-ERG gene fusion (33) were used (Sup- ALDH-specific isoforms are highly expressed in plementary Table S2). prostate cancer clinical specimens Expression of ALDH isoform genes was then compara- We next investigated the expression of selected ALDH- tively analyzed in prostate cancer with different Gleason specific isoform genes. High expression of all tested ALDH scores. Interestingly, expression of ALDH1A1 was upregu- isoforms was found in all prostate cancer tissues (Fig. 3A, lated in high-grade (G8-9) as compared with low-grade left). Expression of these genes was then comparatively (G5-6) cancers (P ¼ 0.03; Fig. 3B). The other isoforms did evaluated in prostate cancer, BPH, and normal prostate. not show any differential gene expression in high-grade as ALDH3A1 was the only isoform displaying a significantly opposed to low-grade cancers (data not shown).

www.aacrjournals.org Clin Cancer Res; 2013 OF5

Le Magnen et al.

A PCa tissues ALDH1A1 ALDH1A3 ALDH3A1 101 101 101 101

100 100 100 100 10–1 10–1 –1 –1 10–2 10 10 10–2

10–3 10–3 10–2 10–2 10–4 10–4 Relative gene expression Relative gene expression –3 –3 Relative gene expression 10–5 10 10 Relative gene expression 10–5 1A1 1A3 3A1 4A1 7A1 9A1 18A1 N BPH PCa N BPH PCa N BPH PCa

ALDH4A1 ALDH7A1 ALDH9A1 ALDH18A1 1 1 101 10 10 101

0 0 100 10 10 100

–1 –1 10–1 10 10 10–1

–2 –2 10–2 10 10 10–2 Relative gene expression –3 –3 Relative gene expression –3 Relative gene expression Relative gene expression 10 10 10 10–3 N BPH PCa N BPH PCa NNBPH PCa BPH PCa

BCALDH1A1

101 P = 0.12 0.6 0.5 P ≤ 0.001 P = 0.006 r = 0.69 r = 0.59 0.4 100 n = 20 n = 20 0.4 0.3 10–1 0.2 0.2 10–2 0.1 1A1 relative expression 1A1 relative expression Relative gene expression 10–3 0.0 0.0 G5-6 G7 G8-9 0 2 4 6 8 0 200 600400 1,000800 n = 13 n = 32 n = 11 % of ALDH bright MFI ratio to control

Figure 3. Comparative expression of ALDH-speciﬁc isoform genes in prostate cancer (PCa) and BPH samples. A, expression of genes encoding ALDH1A1, ALDH1A3, ALDH3A1, ALDH4A1, ALDH7A1, ALDH9A1, and ALDH18A1 isoforms in clinical prostate cancer specimens (left) and in these clinical prostate cancer specimens (^) as compared with BPH samples (*) (other panels). Expression levels are reported as relative values as compared with the housekeeping gene GAPDH. P values and number of samples are listed in Supplementary Table S3. As control, cDNA from a normal prostate pool (N) was tested (~). Mean values are indicated by horizontal bars. , P 0.01; , P 0.001. B, prostate cancer samples were divided into 3 groups according to their Gleason grade (5–9) and expression of ALDH1A1 was compared using a Mann–Whitney test. Horizontal lines represent the medians, whereas outer limits of the whiskers show the minimal and maximal values. , P 0.05. C, Spearman test was used to analyze the correlation between ALDH1A1 (1A1) gene expression and percentages of ALDHbright cells (left) or MFI ratio to control (right) in clinical prostate cancer samples.

ALDEFLUOR reagent is generally thought to act as a gain insights into its potential clinical relevance, we inves- substrate for the ALDH1A1 isoform (34). However, in tigated the expression of ALDH1A1 protein in a series of breast cancer, ALDEFLUOR-dependent ALDH activity has BPH, prostatic intraepithelial neoplasia (PIN), and multi- been attributed to the ALDH1A3 isoform (35). In prostate stage prostate cancer specimens included in 2 TMAs. Num- cancer, ALDH1A1 was the only isoform, whose gene expres- bers and characteristics of samples represented in the TMAs sion correlated with levels of ALDH activity as detected in are given in Table 1. the same tissue specimens (Fig. 3C). These data indicate that To investigate ALDH1A1 protein expression and locali- ALDH1A1 is the main isoform contributing to measurable zation, we ﬁrst used a prostate "progression TMA" compris- ALDH activity in primary prostate cancer. ing 290 evaluable samples including BPH, PIN, early-stage prostate cancer (T1a/b), radical prostatectomy (RP) speci- Expression and localization of ALDH1A1 protein in mens, and castration-resistant (CR) prostate cancer prostate tissues samples. ALDH1A1 gene expression appears to be higher in tumors In BPH, ALDH1A1 protein was detectable in 3 of 65 of higher grade and correlates with enzymatic activity. To specimens (4.6%, Table 1). In the few positive BPH cases,

OF6 Clin Cancer Res; 2013 Clinical Cancer Research

ALDH Activity in Prostate Cancer

ALDH1A1 protein expression was clearly restricted to a few between ALDH1A1 expression and patients overall survival cells in the basal layer of the epithelium (Fig. 4A–C). In (P ¼ 0.89; Fig. 5B). contrast, ALDH7A1 protein in BPH was uniformly expressed in both basal and luminal cells (Supplementary Fig. S1). Discussion To strengthen these data, additional staining of larger ALDHs are detoxification enzymes that catalyze the oxi- sections from BPH specimens was analyzed. This analysis dation of various aldehydes. Their dysfunction is involved confirmed that in positive BPH samples, ALDH1A1 protein in several types of disease including cancer (3). High ALDH expression is usually limited to a few epithelial basal cells activity was shown to characterize hematopoietic stem (Supplementary Fig. S2B). In addition, ALDH1A1 protein cells and to select for TICs in various types of tumors was also detectable in nonepithelial cells, such as histiocytes (5, 12–14), supporting a link between ALDH expression and peripheral nerve cells (Supplementary Fig. S2C and and carcinogenesis. S2D). High ALDH activity was also proposed to select for highly In PIN and low-stage T1a/b prostate cancer samples, tumorigenic cells in prostate cancer cell lines (16, 17, 36). ALDH1A1 positivity was detectable in 3 of 38 (8%) and To date, however, there is no evidence for the existence of 2 of 47 samples (4%), respectively (Table 1). For the few ALDHbright populations in primary prostate cancer. In the positive PIN samples, ALDH1A1 protein expression was present work, we therefore investigated presence, preva- specifically localized in PIN premalignant cells but not in lence, characteristics, and clinical relevance of ALDHbright benign glands (Fig. 4D and E). populations in primary prostate cancer. Consistent with the PIN data, ALDH1A1 positivity was First, we determined whether ALDHbright cell populations specifically detectable in cancerous cells of RP samples (Fig. were present in cells freshly isolated from clinical prostate 4F and G). In CR samples, ALDH1A1 expression showed cancer specimens collected after surgery. We successfully large variations, ranging from fully negative (Fig. 4H), to identified and quantified ALDHbright populations among focally or diffusely positive tumors (Fig. 4I). Interestingly, cell suspensions freshly isolated from 39 clinical prostate þ ALDH1A1( ) cells were also present in the peri-epithelial cancer specimens. To our knowledge, this is the first study stroma (Fig. 4D). reporting the identification of cells with high functional Importantly, ALDH1A1 protein expression was more fre- ALDH activity in uncultured prostate cancer specimens. quent in RP (12 of 69, 17%) and CR (15 of 80, 18%) than in Overall, ALDHbright cells could be detected with variable BPH (P ¼ 0.02 and P ¼ 0.01, respectively) or low-stage and relatively high frequency in most prostate cancer sam- samples (P ¼ 0.04 and P ¼ 0.02, respectively; Fig. 4A ples with a trend toward higher percentages in cancers with and Table 1). higher histologic grade. Using ALDH1A1 scoring, the same trends were observed We found that high ALDH activity, in prostate cancer, by comparing RP and CR with BPH (P ¼ 0.01 and P ¼ 0.003, mainly localized to cells highly positive for the epithelial respectively; Fig. 4A). However, we did not find significant markers EpCAM and Trop2, suggesting an epithelial origin þ differences in percentages of ALDH1A1( ) tissues or global for these cells. Notably, both these markers have been score among samples with different Gleason scores or Ki67 reported to be highly expressed in prostate cancer (28, 31). levels (data not shown). We also tested whether cells with high ALDH activity expressed CD44, a mostly basal cell–specific marker. Inter- estingly, both CD44-positive and -negative phenotypes ALDH1A1 expression is a predictor of poor prognosis were observed within the ALDHbright population. Our pre- in hormone-na€ve patients liminary analysis paves the way toward a thorough pheno- To evaluate the prognostic relevance of ALDH1A1 in typic characterization of ALDHbright populations in primary prostate cancer, ALDH1A1 expression was examined in the prostate cancer. "castration resistance TMA" which contained hormone- The ALDEFLUOR assay is the sole accurate method to na€ve (HN) and CR samples with complete follow-up data. assess functional ALDH activity but is unsuitable to inves- Consistent with the results from the "progression TMA," tigate the clinical and prognostic relevance of ALDHbright ALDH1A1 protein was detectable with significantly higher populations in large cohorts of archived specimens (34). frequency in HN (40 of 100, 40%) and CR (40 of 107, 37%) To date, 19 ALDH isoforms with different substrate prostate cancer than in BPH (Table 1). Score data were also specificities have been identified in the human genome comparably different (HN vs. BPH: P ¼ 0.03; CR vs. BPH: (3). Among these, ALDH1A1 is generally considered to be P ¼ 0.04). the main isoform contributing to ALDH activity, as mea- In the HN patient set, the median survival (MS) was 34 sured by ALDEFLUOR (34). However, evidence for the months for patients with ALDH1A1-positive tumors and 56 contribution of other isoforms in solid tumors has recently months for patients with ALDH1A1-negative tumors (P ¼ been obtained. In fact, ALDH1A3 rather than ALDH1A1 has 0.02; Fig. 5A). been identified as the main isoform involved in ALDE- In the CR group, however, MS for patients with FLUOR-measured ALDH activity and a strong predictor of ALDH1A1-positive or -negative tumors was similar (13 and metastasis in breast cancer (35). Moreover, ALDH7A1 has 10 months, respectively), and no correlation was observed recently been shown to contribute to ALDH activity in a

www.aacrjournals.org Clin Cancer Res; 2013 OF7

Downloaded from CR PCa PIN BPH % ALDH1A1 (+) H F D B A 10 20 30 40 50 0 BPH Published OnlineFirstAugust22,2013;DOI:10.1158/1078-0432.CCR-12-2857 PIN clincancerres.aacrjournals.org T1a/b PCR RP E G C ALDH1A1 score I 10 20 30 0 Cancer Research. P PIN BPH on September 26, 2021. © 2013American Association for T1a/b

PCR RP

CR PCa PIN BPH ,I: H, Magni inserts. in shown and enlarged areas positive indicate Brown arrows positivity. diffuse with CR I, cancer. CR negative H, arrow). (right glands benign negative and arrow) (brown glands cancer positive F G, and lesions. PIN positive E, positivity. of degree some showing cells stromal interstitial D, cells. basal positive few a exhibiting C, glands BPH glands. BPH negative B, prostate tissues. of panel a in expression ALDH1A1 of pictures representative (T1a/b). cancer prostate low-stage BPH and in than CR (PCa) and cancer RP prostate in (right) levels higher at and (left) frequently more expressed is ALDH1A1 A, tissues. prostate cancerous and benign in protein 4. Figure lnclCne Research Cancer Clinical ﬁ 400. ain ,F: B, cation: , P xrsino ALDH1A1 of Expression < 0.05; 0.C ,E G, E, D, C, 200. , P .1 B 0.01. – I, Published OnlineFirst August 22, 2013; DOI: 10.1158/1078-0432.CCR-12-2857

ALDH Activity in Prostate Cancer

Figure 5. ALDH1A1 expression is ABHN PCa CR PCa predictive of poor prognosis in 100 hormone-naïve patients. Kaplan– 100 Meier curves of patient overall 80 80 survival according to ALDH1A1 protein expression. Log-rank 60 60 (Mantel–Cox) tests were used to P = 0.89 P = 0.02 compare ALDH1A1-positive and 40 40 -negative tumors in hormone- ALDH1A1(-) ALDH1A1(-) 20 20 naïve patients (HN; A) and Overall survival ALDH1A1(+) Overall survival ALDH1A1(+) castration-resistant patients (CR; 0 0 P B). , 0.05. PCa, prostate 0 51015 024 6 8 cancer. Time (y) Time (y)

PC3-derived cell line and to be involved in bone metastasis early identification of patients at high risk of developing in prostate cancer (37). Thus, it appears that the nature of prostate cancer. the isoforms contributing to high ALDH activity, as detected Next, we assessed the prognostic relevance of ALDH1A1 by the ALDEFLUOR assay, is tissue- and cell-specific (34). expression using a "castration resistance" TMA. In the To identify prostate cancer relevant isoforms, we have tested hormone-na€ve cohort, we found a significant correlation the expression of seven ALDH isoforms, previously shown between ALDH1A1 positivity and poor patient outcome. to be implicated in ALDH activity, as detected by ALDE- These results are consistent with a study by Li and collea- FLUOR, and in cancer initiation (reviewed in ref. 34). gues, who evaluated a smaller and less diversified patient Except for ALDH3A1, we observed a trend toward higher cohort (38). In contrast, in patients with castration-resistant expression of ALDH isoforms in prostate cancer than in disease, no significant correlation was found between BPH and normal samples. In particular, ALDH1A3, ALDH1A1 positivity and clinical outcome. These data may ALDH7A1, and ALDH18A1 were expressed to significantly suggest a prognostic relevance of ALDH activity limited to higher extents in prostate cancer than BPH. ALDH1A1 gene early phases of prostate cancer. was not expressed to significantly higher extents in prostate ALDH1A1þ cells have been proposed to be responsible cancer than BPH (P ¼ 0.12) but we observed a significantly for tumor re-initiation after castration (40, 41). Our data, higher gene expression in high-grade than in low-grade however, indicate that these cells are not likely associated prostate cancer. Importantly, ALDH1A1 was the sole iso- with prognostic significance in CR prostate cancer, suggest- form whose gene expression was correlated with high ALDH ing that ALDH1A1-negative cells might acquire a predom- activity detected in the same specimens, suggesting its major inant clinical relevance following cancer progression to contribution to functional ALDH activity in prostate cancer. androgen independence. These results led us to investigate expression of ALDH1A1 High ALDH activity has previously been shown to iden- at the protein level, in a large cohort of benign and cancer- tify TICs in solid cancers (5, 12, 13) and established prostate ous samples. In the 2 TMAs analyzed, we found a minor cancer cell lines (16). On the basis of this background, it þ (<5%) proportion of BPH samples with ALDH1A1 positiv- would be tempting to speculate that ALDHbright/ primary ity restricted to a few cells in the basal layer of the epithe- prostate cancer cells might possess tumor-initiating capac- lium, in agreement with a recent study (38). This was ity. However, a formal demonstration of the TIC nature of confirmed by staining of large tissue sections. these ALDHbright primary prostate cancer cells would In contrast, ALDH1A1 expression was specifically require direct testing of their tumorigenicity in vivo. These detected in luminal cancer cells and showed broad varia- studies are hampered by a lack of appropriate experimental tions, ranging from full negativity to focal, diffuse, or models and the difficulties inherent in the generation of strong positivity. These expression patterns highlight the primary cultures and xenografts derived from prostate can- patient-to-patient heterogeneity common in prostate cancer primary cells (19, 20). Recent successful attempts are cer. ALDH1A1 expression was higher and more frequently based on the implantation of tumor fragments under the found in RP and CR than in early-stage prostate cancer. renal capsules of immunodeficient mice in the presence of We also observed intratissue heterogeneity with concom- androgen-releasing pellets (42). Yet generation of cell sus- itant presence of negative and positive glands within the pensions from clinical samples and cell sorting are known same samples. In these cases, ALDH1A1 expression was to affect viability of prostate cancer cells (42, 43), therefore convincingly restricted to cancerous glands. Interestingly, preventing successful tumor transplantation (reviewed in ALDH1A1 positivity was also found in a subgroup of PIN ref. 19). To the best of our knowledge, tumor transplanta- (8%). While these samples usually contain a mixture of tion of cells derived from primary human prostate cancer, as benign and PIN glands, ALDH1A1 expression was restrict- selected by the expression of putative TIC markers, has never ed to PIN lesions. Because PIN is a precursor for prostate been reported so far (44). This prevented us from testing the cancer (39), these findings may have implications for tumorigenicity of ALDHbright primary prostate cancer cells.

www.aacrjournals.org Clin Cancer Res; 2013 OF9

Le Magnen et al.

Alternatively, considering its expression in rare benign Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L. Bubendorf, C.A. Rentsch, J.R. Gsponer, basal cells and in some untransformed interstitial cells, in T. Zellweger, G.C. Spagnoli addition to tumor cells, ALDH1A1 could represent a marker Analysis and interpretation of data (e.g., statistical analysis, biosta- associated with undefined cell differentiation stages, irre- tistics, computational analysis): C. Le Magnen, C.A. Rentsch, J.R. Gsponer, G.N. Thalmann, G.C. Spagnoli spective of a putative role in cancer initiation. Nevertheless, Writing, review, and/or revision of the manuscript: C. Le Magnen, our data clearly document a prognostic relevance of L. Bubendorf, C.A. Rentsch, C. Mengus, T. Zellweger, M. Rieken, G.N. Thalmann, M.G. Cecchini, A. Bachmann, S. Wyler, M. Heberer, G.C. Spagnoli ALDH1A1 expression in untreated prostate cancer. This Administrative, technical, or material support (i.e., reporting or orga- might have important implications for the identification nizing data, constructing databases): T. Zellweger, M. Rieken, S. Wyler, of patients at high risk for progression to castration-resistant G.C. Spagnoli Study supervision: M. Heberer, G.C. Spagnoli disease. In summary, we provide novel evidence of cells with high Acknowledgments ALDH functional activity in primary prostate cancer. Detec- The authors thank Paul Zajac, Beatrice Tonnarelli, Emanuele Trella, and tion at functional, mRNA, and protein levels suggests that Manuele Giuseppe Muraro for helpful discussions and technical support. ALDH might be involved in prostate cancer outgrowth and progression. However, this involvement might be specific Grant Support for subgroups of patients as suggested by the high variability This study was partially supported by unrestricted grants from Astra observed between prostate cancer specimens. Zeneca, the Freiwillige Akademische Gesellschaft Basel, the Lichtenstein Stiftung Basel, Novartis Research Foundation (formerly Ciba-Geigy Jubilee fl Foundation) and the Department of Surgery of the University Hospital Basel. Disclosure of Potential Con icts of Interest The costs of publication of this article were defrayed in part by the No potential conflicts of interest were disclosed. payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate Authors' Contributions this fact. Conception and design: C. Le Magnen, C. Mengus, G.N. Thalmann, M.G. Cecchini, M. Germann, G.C. Spagnoli Received September 4, 2012; revised July 23, 2013; accepted August 3, Development of methodology: C. Le Magnen, C. Mengus, G.C. Spagnoli 2013; published OnlineFirst August 22, 2013.

References 1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer human colonic stem cells (SC) and tracks SC overpopulation during statistics. CA Cancer J Clin 2011;61:69–90. colon tumorigenesis. Cancer Res 2009;69:3382–9. 2. Liu AY, Roudier MP, True LD. Heterogeneity in primary and metastatic 13. Chen YC, Chen YW, Hsu HS, Tseng LM, Huang PI, Lu KH, et al. prostate cancer as defined by cell surface CD profile. Am J Pathol Aldehyde dehydrogenase 1 is a putative marker for cancer stem cells in 2004;165:1543–56. head and neck squamous cancer. Biochem Biophys Res Commun 3. Marchitti SA, Brocker C, Stagos D, Vasiliou V. Non-P450 aldehyde 2009;385:307–13. oxidizing enzymes: the aldehyde dehydrogenase superfamily. Expert 14. Alison MR, Guppy NJ, Lim SM, Nicholson LJ. Finding cancer stem Opin Drug Metab Toxicol 2008;4:697–720. cells: are aldehyde dehydrogenases fit for purpose? J Pathol 4. Kastan MB, Schlaffer E, Russo JE, Colvin OM, Civin CI, Hilton J. Direct 2010;222:335–44. demonstration of elevated aldehyde dehydrogenase in human 15. Burger PE, Gupta R, Xiong X, Ontiveros CS, Salm SN, Moscatelli D, hematopoietic progenitor cells. Blood 1990;75:1947–50. et al. High aldehyde dehydrogenase activity: a novel functional marker 5. Ginestier C, Hur MH, Charafe-Jauffret E, Monville F, Dutcher J, Brown of murine prostate stem/progenitor cells. Stem Cells 2009;27:2220–8. M, et al. ALDH1 is a marker of normal and malignant human mammary 16. van den Hoogen C, van der Horst G, Cheung H, Buijs JT, Lippitt JM, stem cells and a predictor of poor clinical outcome. Cell Stem Cell Guzman-Ramirez N, et al. High aldehyde dehydrogenase activity 2007;1:555–67. identifies tumor-initiating and metastasis-initiating cells in human 6. Storms RW, Trujillo AP, Springer JB, Shah L, Colvin OM, Ludeman SM, prostate cancer. Cancer Res 2010;70:5163–73. et al. Isolation of primitive human hematopoietic progenitors on the 17. Nishida S, Hirohashi Y, Torigoe T, Kitamura H, Takahashi A, Masumori basis of aldehyde dehydrogenase activity. Proc Natl Acad Sci USA N, et al. Gene expression profiles of prostate cancer stem cells isolated 1999;96:9118–23. by aldehyde dehydrogenase activity assay. J Urol 2012;188:294–9. 7. Pearce DJ, Taussig D, Simpson C, Allen K, Rohatiner AZ, Lister TA, 18. Pan C, Kumar C, Bohl S, Klingmueller U, Mann M. Comparative et al. Characterization of cells with a high aldehyde dehydrogenase proteomic phenotyping of cell lines and primary cells to assess activity from cord blood and acute myeloid leukemia samples. Stem preservation of cell type-specific functions. Mol Cell Proteomics Cells 2005;23:752–60. 2009;8:443–50. 8. Dieter SM, Ball CR, Hoffmann CM, Nowrouzi A, Herbst F, Zavidij O, 19. Toivanen R, Taylor RA, Pook DW, Ellem SJ, Risbridger GP. Breaking et al. Distinct types of tumor-initiating cells from human colon cancer through a roadblock in prostate cancer research: an update on human tumors and metastases. Cell Stem Cell 2011;9:357–65. model systems. J Steroid Biochem Mol Biol 2012;131:122–31. 9. Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF. 20. Peehl DM. Primary cell cultures as models of prostate cancer devel- Prospective identification of tumorigenic breast cancer cells. Proc Natl opment. Endocr Relat Cancer 2005;12:19–47. Acad Sci USA 2003;100:3983–8. 21. Ertel A, Verghese A, Byers SW, Ochs M, Tozeren A. Pathway-specific 10. Collins AT, Berry PA, Hyde C, Stower MJ, Maitland NJ. Prospective differences between tumor cell lines and normal and tumor tissue cells. identification of tumorigenic prostate cancer stem cells. Cancer Res Mol cancer 2006;5:55. 2005;65:10946–51. 22. Collins AT, Habib FK, Maitland NJ, Neal DE. Identification and isolation 11. Quintana E, Shackleton M, Sabel MS, Fullen DR, Johnson TM, Morri- of human prostate epithelial stem cells based on alpha(2)beta(1)- son SJ. Efficient tumour formation by single human melanoma cells. integrin expression. J Cell Sci 2001;114:3865–72. Nature 2008;456:593–8. 23. Le Magnen C, Bubendorf L, Ruiz C, Zlobec I, Bachmann A, Heberer M, 12. Huang EH, Hynes MJ, Zhang T, Ginestier C, Dontu G, Appelman H, et al. Klf4 transcription factor is expressed in the cytoplasm of prostate et al. Aldehyde dehydrogenase 1 is a marker for normal and malignant cancer cells. Eur J Cancer 2013;49:955–63.

OF10 Clin Cancer Res; 2013 Clinical Cancer Research

ALDH Activity in Prostate Cancer

24. Mertz KD, Setlur SR, Dhanasekaran SM, Demichelis F, Perner S, 34. Marcato P, Dean CA, Giacomantonio CA, Lee PW. Aldehyde dehy- Tomlins S, et al. Molecular characterization of TMPRSS2-ERG gene drogenase: its role as a cancer stem cell marker comes down to the fusion in the NCI-H660 prostate cancer cell line: a new perspective for specific isoform. Cell Cycle 2011;10:1378–84. an old model. Neoplasia 2007;9:200–6. 35. Marcato P, Dean CA, Pan D, Araslanova R, Gillis M, Joshi M, et al. 25. Ruiz C, Holz DR, Oeggerli M, Schneider S, Gonzales IM, Kiefer JM, Aldehyde dehydrogenase activity of breast cancer stem cells is pri- et al. Amplification and overexpression of vinculin are associated with marily due to isoform ALDH1A3 and its expression is predictive of increased tumour cell proliferation and progression in advanced pros- metastasis. Stem Cells 2011;29:32–45. tate cancer. J Pathol 2011;223:543–52. 36. Yu C, Yao Z, Dai J, Zhang H, Escara-Wilke J, Zhang X, et al. ALDH 26. Zellweger T, Sturm S, Rey S, Zlobec I, Gsponer JR, Rentsch CA, et al. activity indicates increased tumorigenic cells, but not cancer stem Estrogen receptor beta expression and androgen receptor phosphor- cells, in prostate cancer cell lines. In Vivo 2011;25:69–76. ylation correlate with a poor clinical outcome in hormone-naive pros- 37. van den Hoogen C, van der Horst G, Cheung H, Buijs JT, Pelger RC, tate cancer and are elevated in castration-resistant disease. Endocr van der Pluijm G. The aldehyde dehydrogenase enzyme 7A1 is func- Relat Cancer 2013;20:403–13. tionally involved in prostate cancer bone metastasis. Clin Exp Metas- 27. Mottet N, Bellmunt J, Bolla M, Joniau S, Mason M, Matveev V, et al. tasis 2011;28:615–25. EAU guidelines on prostate cancer. Part II: Treatment of advanced, 38. Li T, Su Y, Mei Y, Leng Q, Leng B, Liu Z, et al. ALDH1A1 is a marker for relapsing, and castration-resistant prostate cancer. Eur Urol 2011;59: malignant prostate stem cells and predictor of prostate cancer 572–83. patients' outcome. Lab Invest 2010;90:234–44. 28. Went P, Vasei M, Bubendorf L, Terracciano L, Tornillo L, Riede U, et al. 39. Bostwick DG, Cheng L. Precursors of prostate cancer. Histopathology Frequent high-level expression of the immunotherapeutic target Ep- 2012;60:4–27. CAM in colon, stomach, prostate and lung cancers. Br J Cancer 40. Germann M, Wetterwald A, Guzman-Ramirez N, van der Pluijm G, 2006;94:128–35. Culig Z, Cecchini MG, et al. Stem-like cells with luminal progenitor 29. Gires O, Klein CA, Baeuerle PA. On the abundance of EpCAM on phenotype survive castration in human prostate cancer. Stem Cells cancer stem cells. Nat Rev Cancer 2009;9:143; author reply 143. 2012;30:1076–86. 30. Goldstein AS, Huang J, Guo C, Garraway IP, Witte ON. Identification 41. Qin J, Liu X, Laffin B, Chen X, Choy G, Jeter CR, et al. The PSA(-/lo) of a cell of origin for human prostate cancer. Science 2010;329: prostate cancer cell population harbors self-renewing long-term 568–71. tumor-propagating cells that resist castration. Cell Stem Cell 2012;10: 31. Trerotola M, Cantanelli P, Guerra E, Tripaldi R, Aloisi AL, Bonasera V, 556–69. et al. Upregulation of Trop-2 quantitatively stimulates human cancer 42. Priolo C, Agostini M, Vena N, Ligon AH, Fiorentino M, Shin E, et al. growth. Oncogene 2013;32:222–33. Establishment and genomic characterization of mouse xenografts of 32. Trerotola M, Jernigan DL, Liu Q, Siddiqui J, Fatatis A, Languino LR. human primary prostate tumors. Am J Pathol 2010;176:1901–13. Trop-2 promotes prostate cancer metastasis by modulating beta1 43. Chen S, Principessa L, Isaacs JT. Human prostate cancer initiating integrin functions. Cancer Res 2013;73:3155–67. cells isolated directly from localized cancer do not form prostaspheres 33. Tomlins SA, Rhodes DR, Perner S, Dhanasekaran SM, Mehra R, Sun in primary culture. Prostate 2012;72:1478–89. XW, et al. Recurrent fusion of TMPRSS2 and ETS transcription factor 44. Garber K. A tale of two cells: discovering the origin of prostate cancer. J genes in prostate cancer. Science 2005;310:644–8. Natl Cancer Inst 2010;102:1528–9, 1535.

www.aacrjournals.org Clin Cancer Res; 2013 OF11

Characterization and Clinical Relevance of ALDHbright Populations in Prostate Cancer

Clémentine Le Magnen, Lukas Bubendorf, Cyrill A. Rentsch, et al.

Clin Cancer Res Published OnlineFirst August 22, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-12-2857

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2013/09/05/1078-0432.CCR-12-2857.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/early/2013/09/11/1078-0432.CCR-12-2857. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.