The stem cell-associated Hiwi gene in human adenocarcinoma of the pancreas: expression and risk of tumour-related death

Piwi proteins and their interaction with piRNAs have rapidly emerged as important contributors to gene regulation, indicating their crucial function in germline and stem cell development. However, data on the Hiwi 1 (Hiwi) gene, one of the four human Piwi homologues, are still scarce. Therefore, we investigated the Hiwi mRNA expression in microdissected PDAC tissues from patients with ductal adenocarcinoma of the pancreas (PDAC) by quantitative real-time PCR and the protein expression by immunohistochemistry. Elevated levels of Hiwi mRNA transcripts were measured in 40 out of 56 tissues and a positive immunostaining of Hiwi was detected in tumours of 21 out of 78 patients. There was no general impact of elevated Hiwi mRNA transcript levels or protein expression on survival, as tested by multivariate Cox regression and Kaplan–Meier analysis. However, men showed a significantly increased risk for tumour-related death in case of down- or upregulated expression of Hiwi mRNA (relative risk (RR)=2.78; P=0.034). In summary, we report the first analysis of Hiwi expression in PDAC and its impact on prognosis. We suggest that alterations in mRNA expression of Hiwi can increase the risk of tumour-related death in male PDAC patients.

Current stem cell genetic research provides only scarce data on Hiwi 1 (Hiwi) -one of four human homologues of the Piwi gene family -despite growing focus of studies on the latter and its interacting RNAs (piRNAs), for which a key function in transcriptional gene silencing and germline/stem cell development in flies and mammals has been demonstrated (Aravin et al, 2007;Seto et al, 2007;Farazi et al, 2008). Data on Piwi proteins, a subfamily of Argonaute proteins, highlight their impressive biological capacities, such as biogenesis/regulation of small RNAs, control of protein synthesis and mRNA stability, showing conserved functions in maintenance and self-renewal of stem cells (Seto et al, 2007;Hutvagner et al, 2008).
It has been suggested that deregulation of stem cell selfrenewal may cause the development of malignancies because normal stem cells share several similarities with so-called tumour stem cells, for example, the ability to self-renew and a relative resistance to drugs (Soltysova et al, 2005;Chen et al, 2007;Taubert et al, 2007a). Besides being expressed abundantly in germline cells of human testes, Hiwi's enhanced expression was detected in testicular seminomas, which originate from embryonic germ cells with retention of a germ cell phenotype (Qiao et al, 2002). This has given rise to the assumption that Hiwi overexpression may be involved in the development of germ cell malignancy, prompting further investigation on other human malignancies. Consequently, it has been shown that expression of the Hiwi gene in human gastric cancer was associated with the proliferation of cancer cells (Liu et al, 2006). Moreover, in patients with soft-tissue sarcoma, elevated Hiwi mRNA transcript levels were associated with a significantly increased risk of tumour-related death (Taubert et al, 2007a, b). However, Hiwi was not detectable in leukaemia cell lines (Sharma et al, 2001).
Ductal adenocarcinoma of the pancreas (PDAC), a dismal disease with a late clinical presentation and a very poor overall prognosis (Yeo et al, 2002), seems to be an interesting subject to investigate with regards to the involvement of stem cell-associated genes and Hiwi in particular. The main features of PDACaggressive biological phenotype, early local invasion with high metastatic potential and a high resistance to radiation and chemotherapy (Yeo et al, 2002) -suggest the involvement of cells with stem cell characteristics in PDAC. Analysis of the stem cellassociated gene Hiwi in this malignancy might provide important clues to the understanding of this disease.
Besides the above listed tumour entities, no studies on other malignancies, particularly on PDAC, on the involvement of Hiwi expression in tumour development have been reported, and the impact of Hiwi expression on patients' prognosis has only been studied in soft-tissue sarcomas yet (Taubert et al, 2007a, b). Therefore, the aim of our study was to investigate the expression profile of Hiwi and its impact on prognosis in PDAC patients.

Patients
We analysed a cohort of patients who underwent primary surgery for PDAC in the years 2001 -2005 in our clinic (Department of Surgery 1, University of Ulm, Germany). Paraffin-embedded tissue of 78 patients (31 females and 47 males, age range 34 -80 years; mean age 61.1 years) was obtained. Furthermore, fresh-frozen tissue of 56 patients (22 females and 34 males, age range 34 -80 years; mean age 61.7 years) was conserved. All patients from whom fresh-frozen tissue suitable for microdissection (sufficient amount of neoplastic cells per sample (total of 44000 cells per patient) and intact RNA (as tested with Agilent Bioanalyzer, see below) was obtained were included in the study. The mean observation time was 15.99 (range 1 -61) months and the median survival rate was 15.26 (range 1 -52) months. Four patients died from non-tumourrelated causes. None of the patients has received neoadjuvant chemo-or radiotherapeutic treatment. All patients included in the study gave written informed consent. The approval of the local ethics committee was obtained.

Microdissection and QRT -PCR analysis
Fresh-frozen tissue samples from 56 PDAC patients were microdissected to exclude stromal tissue and highly enrich neoplastic cells (Burgemeister et al, 2003). Surgical pancreatic resection specimens were immediately placed on ice and subsequently snapfrozen, then stored at À801C. The integrity of the isolated RNA (using Innuprep RNA mini kit, AJ Innuscreen GmbH, Berlin, Germany) was confirmed with the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). The cryostat tissues were cut into 8 -10-mm sections. Microdissection of selected areas (containing approximately 50 -300 neoplastic ductal epithelial cells per dissected area, 44000 cells per patient per tissue) was carried out by Laser microdissection and pressure catapulting technique (PALM Microlaser Technologies, Bernried, Germany) on cresyl violet-stained sections (Supplementary Figures). Subsequently, we performed quantitative real-time PCR analysis for the expression of Hiwi mRNA. The mRNA expression of Hiwi was quantified by a commercially available TaqMan gene expression assay (Assay-ID: Hs01041737_ml; Applied Biosystems, Foster City, CA, USA) and standardised to the transcript level of hypoxanthinphospho-ribosyl-transferase (HPRT), using TaqMan gene expression assay (Assay-ID: Hs99999909_ml; Applied Biosystems).

Immunohistochemistry
Analysis of 78 PDAC tissue specimens by immunohistochemistry (IHC) was performed. All specimens were fixed in buffered formalin and embedded in paraffin. Sections of 4-mm thickness were deparaffinised in xylene and rehydrated. After microwaveheating of the sections in citrate buffer at 450 W for 2 min and subsequently at 80 W for 20 min, slides were cooled to room temperature, then treated with peroxidase-blocking reagent (DakoCytomation, Dako, Hamburg, Germany) and serum-blocking reagent (VECTASTAIN Elite ABC Kit; Vector Lab., Burlingame, CA, USA). Sections were incubated overnight at 41C with an 1 : 250 diluted goat anti-human polyclonal Hiwi antibody (HIWI N-17:sc-22685; Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA). The slides were further treated with the VECTASTAIN Elite ABC Kit, according to the manufacturer's instructions using a secondary biotinylated Anti-goat-IgG antibody. Diaminobenzidine was used as chromogen and the sections were counterstained with Mayer haematoxylin. All washing steps were performed in Tris-buffer. Immunohistochemical evaluation was carried out by at least two independent investigators. At least 500 neoplastic ductal cells were counted/examined per tissue section. Positive staining required a minimum of 10% of neoplastic cells exhibiting a distinct staining. Three patterns of staining intensity were defined: strong, intermediate and weak staining. Testicular seminomas served as positive control.

Statistical analysis
Kaplan -Meier and multivariate survival analysis according to the Cox's proportional hazards regression model (adjusted for tumour staging, type of tumour resection, patients' age and gender) was performed for evaluation of Hiwi mRNA expression and IHC results. First, two categories of Hiwi levels were set up for analysis of mRNA transcript levels: (1) 'No expression' (o0.001 attogram (ag) Hiwi mRNA per (femtogram) fg HPRT mRNA); (2) 'Elevated expression' (X0.001 ag Hiwi mRNA per fg HPRT mRNA). Second, analyses using the mean and median values as cutoff levels were performed. Accordingly, threshold levels were set at 1.12 and 0.62 ag Hiwi mRNA per fg HPRT mRNA, respectively. A third analysis applied the quartile of Hiwi mRNA expression and resulted in four groups, that is, two groups with low expression

Pattern of Hiwi expression
Elevated transcript levels of Hiwi mRNA were measured in 40 out of 56 PDAC tissues (71.4%), whereas the remaining patients showed no detectable transcripts. The mean expression level of Hiwi transcripts was 1.12 (range: 0.00 -28.24; median: 0.62) ag Hiwi mRNA per fg HPRT mRNA. Patients whose tumours showed an expression below or above the mean or median, revealed an equal distribution between tumour stages, resection types (R-status), age and gender (Table 1).
In IHC, positive immunostaining with the anti-HIWI N-17 antibody was detected in 21 out of 78 samples (26.9%), whereas the remaining 57 tumours showed no positive staining in more than 10% of the cells (Table 2). Of the 21 positively stained tumours, 11 (14.1%) showed a simultaneous immunostaining in the nucleus and cytoplasm of the neoplastic ductal cells and 10 (12.8%) demonstrated a sole cytoplasmatic staining pattern (Figure 1). A total of 17 tumours showed a strong and 4 a moderate staining intensity, whereas a weak staining was not observed among the tumour samples. Patients with a Hiwi expression on protein level showed an equal distribution between the tumour stages, resection types, age and gender.

Clinical outcome and Hiwi expression
In Kaplan -Meier and multivariate Cox's regression hazard analysis (adjusted to tumour resection type and tumour stage) no impact of Hiwi mRNA expression on survival was determined in a comparison of groups of patients whose tumours expressed Hiwi either below or above the median/mean level of expression. Furthermore, separation of patients in two groups with tumours expressing undetectable levels (o0.001 ag Hiwi mRNA per fg HPRT mRNA) or detectable levels of Hiwi transcripts (X0.001 ag Hiwi mRNA per fg HPRT mRNA) did not show an impact of Hiwi expression on survival (RR ¼ 1.401; P ¼ 0.328). These findings were independent of sex or age. In a former study involving patients with soft-tissue sarcoma, we identified three groups of patients, where patients with an elevated or a reduced Hiwi expression in their tumours showed a poor prognosis compared to patients with an intermediate expression level (Taubert et al, 2007a). Therefore, we separated patients into a group where tumours carried a high or a low expression of Hiwi transcripts and a group with intermediate Hiwi expression. We performed this separately for female and for male patients. In a multivariate Cox's regression analysis in male patients, we detected a 2.78-fold increased risk of tumour-related death for patients (P ¼ 0.034) whose tumours showed an elevated or a decreased expression of Hiwi compared to patients with an intermediate expression of Hiwi (Figure 2). Interestingly, for female patients no different impact on prognosis could be found comparing both patient groups. Evaluating the results obtained from immunohistochemistry, no impact of Hiwi protein expression on tumour-related death was found. No effect on patients' prognosis was observed either for isolated cytoplasmatic staining or simultaneous staining of the nucleus and cytoplasm. There was no significant impact on the patients' gender and Hiwi protein expression on survival, as evaluated by IHC.

DISCUSSION
Here, we report the first analysis of Hiwi mRNA and protein expression in PDAC and its relation to prognosis. Consistent with  findings on other tumour entities, our results indicate that Hiwi might not only be relevant in germ cell malignancies but also in solid tumours of epithelial or mesenchymal origin (Qiao et al, 2002;Liu et al, 2006;Taubert et al, 2007a, b). Besides testicular tumours, so far the expression of the human homologue of the Piwi gene has only been investigated in leukaemia cell lines (Sharma et al, 2001), gastric cancer (Liu et al, 2006) and soft-tissue sarcomas (Taubert et al, 2007a, b), demonstrating elevated protein and/or mRNA transcript levels in the two latter kinds of tumours. Elevated and reduced Hiwi mRNA message were significantly associated with a poor outcome in soft-tissue sarcoma patients, outlining its possible key function in non-germ cell malignancies (Taubert et al, 2007a, b). Accordingly, we observed a high detection rate of Hiwi mRNA transcripts in our cohort, with elevated mRNA levels detectable in 40 out of 56 patients.
Regarding survival analysis, no impact of Hiwi mRNA expression on tumour-related death was ascertained for all PDAC patients. However, when only male patients were considered, a 2.78-fold increased risk of tumour-related death was detected for patients whose tumour showed an altered, that is, down-or upregulated expression of Hiwi mRNA compared to those with an intermediated Hiwi transcript level. The gender-dependent impact of Hiwi mRNA expression on prognosis is not easy to explain. Hiwi expression has been detected in normal human testes and male germline malignancies, but was not detectable in ovaries (Qiao et al, 2002). Moreover, mice bearing targeted mutations in the murine-homologues Miwi , Mili (Kuramochi-Miyagawa et al, 2004) and Miwi2 (Girard et al, 2006) are male-sterile with distinct defects in spermatogenesis. However, in lower eukaryotic organisms, such as Drosophila, mutations of the Piwi gene caused failure in germline stem cells in both females and males (Cox et al, 1998). Hormonal aspects might have a function in PDAC as investigations on the expression of androgen and oestrogen/progesterone receptors have demonstrated (Andren-Sandberg et al, 1999;Lam et al, 2008). It would be interesting to investigate the expression of the Hiwi gene -especially in tumours of hormonally regulated or active organs -as it might give additional clues to its function and regulative processes in humans.
As far as immunohistochemistry for Hiwi protein expression is concerned, a strong positive staining in 26.9% of the carcinomas (21 out of 78 patients) was observed, a rate substantially lower than for mRNA transcript levels. These differences in the rates of elevated Hiwi expression between mRNA and protein levels, however, are very likely to be because of the much higher sensitivity of the real-time PCR technique than that of IHC, thus not allowing for detection of relevant Hiwi protein levels. Furthermore, a wide range of modified protein variants (e.g. posttranslational or other downstream modifications) might remain undetected by IHC.
The detection rate of Hiwi mRNA transcripts and the association with a gender-dependent risk pattern underscores the involvement of stem cell-associated genes in carcinogenesis and cancer development in PDAC. Given the tumour's aggressive potential and frequent changes in the Hiwi mRNA levels, Hiwi as a stem cellassociated gene may have general and early impact on tumourigenesis in PDAC. A comparison of expression levels in PDAC and in the normal pancreas could help to relate expression levels to tumour biological behaviour. Altogether, the gene might be a potential important target for further investigation in carcinogenesis and therapy of this dismal disease.
In summary, we demonstrate for the first time the expression of the stem cell-associated Hiwi gene -a human homologue of the Piwi family -in human PDAC and identify Hiwi as a potential target for further investigation in carcinogenesis and therapy of this dismal disease. The detection rate of Hiwi mRNA in clinically apparent advanced PDAC and the lack of the gene's significant impact on tumour-related death, in general, might be indicative of the tumour's aggressive potential and hints at a broad and early impact of the Hiwi gene in carcinogenesis of PDAC. Intriguingly, alterations in mRNA expression levels of Hiwi can increase the risk of tumour-related death in male patients, pointing at gender differences in the prognostic impact of Hiwi mRNA expression in PDAC.