Background
PSCA has been discovered a decade ago and has been classified as a member of the Ly-6 family of GPI-anchored cell surface proteins [
1]. It is expressed in most prostate cancer specimens, including high-grade prostatic intraepithelial neoplasia, primary androgen-dependent tumors, and hormone-refractory metastases. PSCA levels are positively correlated with Gleason grade, tumor stage, and biochemical recurrence. Its expression is also particularly elevated in bone metastasis. Finally, PSCA is strongly expressed in other malignancies, including bladder and pancreatic cancers [
2‐
6]. Different immunotherapy approaches targeting PSCA have been tested in preclinical models including cancer vaccine, therapeutic monoclonal antibodies and antibody conjugated to toxic drugs [
7‐
10]. More recently, a human monoclonal antibody targeting PSCA has been evaluated in a phase I clinical trial in prostate cancer patients (AACR 2006).
Little information is available regarding the biological role of PSCA. Proteins belonging to the Ly6 family are involved in cell signaling events associated with thymic lymphocyte differentiation, maturation and activation. CD59, a member of this protein family, was shown to play a role in the protection against complement mediated lysis. In addition, it was found to be expressed in tumor cells where it may play a role in evading anti cancer immune response [
11]. Deletion of PSCA gene does not appear to interfere with normal development as PSCA knockout mice are viable. Additionally, crossing of the PSCA knockout mice with prostate tumor models driven by large T antigen did not increase primary tumor formation [
12,
13].
Here the biological role of human PSCA was evaluated using RNA interference and microarray analysis. To establish a pharmacologic control over gene expression a shRNA against PSCA was identified and expressed under the control of dox in a lentivirus system [
14]. Microarray analysis was utilized to identify genes coregulated with PSCA in tumor xenografts.
Methods
Cell culture and generation of lentivirus vectors
The 293T and SW780 cell lines were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. The lentivirus system for conditional gene suppression with dox-inducible shRNAs utilized has been described previously[
14]. Briefly, in the absence of dox, tTR-KRAB repressor binds to
tetO and suppresses H1-mediated shRNA transcription, thus allowing normal expression of the cellular target gene. In the presence of 10 μg/ml of dox, tTR-KRAB cannot bind to
tetO and hence shRNAs are produced, leading to downregulation of PSCA. The green fluorescent protein (GFP) cDNA contained in the shRNA vector provides a monitoring device, as it is switched on by dox treatment and GFPis expressed. All recombinant lentiviruses were produced by transient transfection in 293T cells. Briefly, 293T cells were cotransfected with 20 μg of pLUTHM-shPSCA3 plasmid, 15 μg of pCMV-ΔR8.91, and 5 μg of pMD2G-VSVG by calcium phosphate precipitation. After 16 h medium was changed, and recombinant lentivirus vectors were harvested 48 h later. FACS analysis was conducted as previously described [
15].
Viability assay
Cell viability was monitored using the CellTiter-Blue Viability. The assay is based on the ability of living cells to convert a redox dye (resazurin) into a fluorescent end product (resorufin); 1 × 103 SW780-shPSCA and SW780-shControl cells +/- Dox were plated in a 96 well plate in parallel with the parental cell line SW780. Cells were incubated at 37°c for 96 hrs, and fluorescence was subsequently monitored using a plate-reading fluorometer.
Tumor models
Six-week-old female CD-1 nude mice were purchased from Charles River Laboratories and maintained in accordance to Guidelines for the Care and Use of Laboratory Animals in IRBM's animal facility. This study, was submitted and approved by the IRBM ethical committee . Mice were injected subcutaneously (sc) in the right flank with 2 × 106 SW780-shPSCA cells resuspended into 100 μL phosphate-buffered saline (PBS) and Matrigel (1:1). Mice received 5% sucrose only or 5% sucrose plus 0.2 mg/ml of dox for control and knockdown cohorts, respectively. All water bottles were changed 3 times per week. Tumors were measured with calipers and mice weighed twice per week. At the end of the dosing study, or as indicated, appropriate tumor samples were taken.
Microarray experiments
Total RNA from cell in culture was isolated with RNAzolB and then dissolved in RNase-free water. 25 μg of total RNA was treated with DNase using the Qiagen RNase-free DNase kit and RNeasy spin columns. Then RNA was dissolved in RNase-free water to a final concentration of 0.2 μg/μl. cRNA was generated using T7 RNA polymerase on 5 μg of total RNA and labeled with Cy5 or Cy3 (Cy Dye, Amersham Pharmacia Biotech). From each sample, 5 μg of labeled-RNA were co-hybridized with 5 μg of a reference RNA (pool of two untreated SW780 cell lines). Labeled cRNAs were fragmented to an average size of 50-100 nucleotides by heating the samples to 60°C with 10 mM of zinc chloride and then adding an hybridization buffer containing 1 M NaCl, 0.5% sodium sarcosine, 50 mM MES, pH6.5, and formamide to a final concentration of 30%. The final volume was 3 ml at 40°C.
Samples were hybridized on a customized Agilent 44 k array containing ~40,000 unique probes mapping to ~21,000 human genes. Each sample was hybridized in duplicate with fluor reversal to systematically correct for dye bias. After hybridization, slides were washed and scanned using a confocal laser scanner (Agilent Technologies). The raw intensities obtained after scanning were quantified, background-corrected and lowness normalized. A weighted average ratio was computed for dye-swapped pairs of hybridizations. Tumors were collected in RNA later and processed for RNA extraction as described previously [
16]. Samples were hybridized on a customized Affymetrix array containing ~38,000 probes mapping to ~21,000 human genes [
17].
Hierarchical clustering
The microarray dataset was filtered before clustering in order to select the 2,000 most variable probes. probes with an absent call in more than 12 samples out of 14 were removed the 2,000 probes with the higher interquartile range were retained for subsequent analysis. Probes were hierarchically clustered using an average linkage algorithm based on Pearson correlation coefficients.
Identification of deregulated probes
Differences in average probe expression between the dox+ (PSCA silencing) and dox-samples were computed by 1-way ANOVA. Probes differentially expressed between the two classes were identified based on ANOVA p-value < 0.001.
Gene set enrichment analysis
Groups of genes identified by 1-way ANOVA were compared to a collection of annotated gene sets to identify the functional classes that were significantly over-represented. The enrichment p-values were computed according to the Fisher's exact test. The gene sets were obtained from public (Gene Ontology [
18], KEGG [
19], Interpro [
20], Panther [
21], oPOSSUM [
22] and commercial sources (GeneGo (GeneGo Inc., St Joseph, MI, USA), Ingenuity (Ingenuity Systems Inc, Mountain View, CA, USA), TRANSFAC [
23].
RT-PCR
Microarray findings were confirmed by real-time reverse-transcription PCR (RT-PCR) using the average value of mRNA from dox untreated SW780-shPSCA samples as calibrator. Analysis of mRNA expression of selected IFNα genes was performed using Applied Biosystems expression kit following manufacturers' instructions. Expression values were computed using the comparative CT method (ΔΔ CT) with GAPDH gene expression value serving as normaliser.
Discussion
In this study, we have identified a correlation between PSCA expression and tumor growth
in vivo. High level of PSCA mRNA and protein expression has been observed in most primary prostate and pancreatic human tumors and in particular in aggressive metastatic forms. In bladder cancer, higher levels of PSCA expression correlated with increasing tumor grade [
25] and more recently genetic variants in this gene were associated with cancer occurrence, although the biological implications of this observation remain to be elucidated[
26].
The observed tumor specific expression of PSCA has prompted the development of therapeutic antibodies specific for this membrane protein. Indeed, growth inhibition has been observed upon antibody treatment of cells transfected with vectors expressing PSCA or of naturally expressing cell lines such as SW780. The published studies are in agreement with the data reported here indicating that reduced PSCA expression is associated with lower viability [
7]. Thus, based on these published data, a better understanding of the biologic role played by PSCA in tumor growth is warranted.
To shed light on PSCA role in tumor biology we compared PSCA expression in different cell lines and found expression on cell surface in few of them or in explanted primary tumors. Thus, the bladder cell line SW780 was engineered to obtain a pharmacological controlled expression of PSCA and to study the impact of reduced PSCA gene expression on cell viability and tumor growth. Here, we report a direct correlation between PSCA expression and tumor growth. Of note is the observation that pretreatment with dox in vitro further reduced tumor growth. This latter result may be partially explained by the slow decay of PSCA displayed on the cell membrane as indicated by FACS analysis.
A role for PSCA in tumor proliferation is further supported by the biological pathway analysis of gene expression profiling where a statistical significant association with "negative regulation of apoptosis" and "growth" was observed. Many of the pathways activated upon PSCA downregulation control key immune functions. Activation of IFNα/β signal transduction pathway was observed only in SW780-shPSCA tumors (Figure
2B) and it was not evident in tumors obtained upon injection of control cells. In contrast, similar studies conducted with these cell lines
in vitro did not show induction of the pathway suggesting that in the tumor context additional factors such as those related to immune modulation may contribute to the observed microarray signature. A quantitative analysis of gene expression by qPCR confirmed overexpression of IFNα/β genes (Figure
4). The activation of IFNα/β pathway only in vivo further supports the idea that triggering of this pathway is not determined by an intracellular mechanism related to double strand RNA, as it has been previously reported. On the contrary, the data suggest that IFNα/β activation is related to environmental signals. The activation of IFNα/β pathway is consistent with a reduced tumor growth as previously shown with recombinant IFNα [
27,
28] .
Nonetheless, further experiments are required to better characterize the link between PSCA downregulation and activation of immune pathways such as IFNα/β. In this direction it is worth mentioning a potential physical interaction between PSCA and IFNα/β receptor. PSCA is a GPI-anchored protein located in lipid raft. IFNα/β receptor can be brought into this subcellular compartment upon interaction with its ligands [
29]. Thus, PSCA may counteract intracellular signaling exerted by the IFNα/β receptor playing a role as a defense protein. This biological role is reminiscent of the interfering role played by PSCA homolog, CD59, in the complement system [
30]. Activation of the complement system is tightly regulated and among the various regulators CD59 has been identified as the single membrane regulator of the terminal membrane attack complex. Moreover CD59 has been characterized as a negative regulator of the T-cell response in mice. Although we limited our qPCR confirmation only to IFNα/β pathway, a statistical significant up regulation was observed also for other immune pathways such as of MHC-1 pathway. Given the pleiotropic effects of IFNα/β pathway it is likely that some pathways, including MHC-1 presentation, may be affected downstream. A more in depth analysis would be required to answer this question. Finally we cannot exclude that the correlations observed only in tumors and not in cell culture are driven by undefined factors that are not captured in the expression profile and deserve further investigation.
Acknowledgements
The authors thank Aya Jakobovits and Jean Gudas for anti PSCA antibody, 4.117. The authors also thank Brendan Leeson and Ernest Coffey from Gene Expression Laboratories - Rosetta and LAR personnel for technical support. This work was partially supported by Progetto Integrato Oncologia, Ministero della Salute e Associazione Italiana per la Ricerca sul Cancro (AIRC).
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
EM generated cell lines expressing shRNA, analyzed PSCA expression by FACS analysis and performed experiments in vivo. PU performed microarray analysis and pathway analysis- VV generated cell lines expressing shRNA and performed IHC on tumor samples. VS performed qPCR analysis. ED contributed to the discussion of the data. EDR contributed in the discussion of the bioinformatics data and helped in drafting bioinformatics result section. AL performed the initial microarray analysis and contributed to data analysis. NLM, AN and GC contributed to the drafting and general discussion of the paper. FP conceived the study, analyzed the data and wrote the manuscript. All authors read and approved the final manuscript.