Gene expression profiling of human prostate cancer stem cells reveals a pro-inflammatory phenotype and the importance of extracellular matrix interactions
© Birnie et al.; licensee BioMed Central Ltd. 2008
Received: 20 December 2007
Accepted: 20 May 2008
Published: 20 May 2008
The tumor-initiating capacity of many cancers is considered to reside in a small subpopulation of cells (cancer stem cells). We have previously shown that rare prostate epithelial cells with a CD133+/α2β1hi phenotype have the properties of prostate cancer stem cells. We have compared gene expression in these cells relative to their normal and differentiated (CD133-/α2β1low) counterparts, resulting in an informative cancer stem cell gene-expression signature.
Cell cultures were generated from specimens of human prostate cancers (n = 12) and non-malignant control tissues (n = 7). Affymetrix gene-expression arrays were used to analyze total cell RNA from sorted cell populations, and expression changes were selectively validated by quantitative RT-PCR, flow cytometry and immunocytochemistry. Differential expression of multiple genes associated with inflammation, cellular adhesion, and metastasis was observed. Functional studies, using an inhibitor of nuclear factor κB (NF-κB), revealed preferential targeting of the cancer stem cell and progenitor population for apoptosis whilst sparing normal stem cells. NF-κB is a major factor controlling the ability of tumor cells to resist apoptosis and provides an attractive target for new chemopreventative and chemotherapeutic approaches.
We describe an expression signature of 581 genes whose levels are significantly different in prostate cancer stem cells. Functional annotation of this signature identified the JAK-STAT pathway and focal adhesion signaling as key processes in the biology of cancer stem cells.
The concept of a cancer stem cell within a more differentiated tumor mass, as an aberrant form of normal differentiation, is now gaining acceptance over the current stochastic model of oncogenesis, in which all tumor cells are equivalent both in growth and tumor-initiating capacity [1, 2]. For example, in leukaemia, the ability to initiate new tumor growth resides in a rare phenotypically distinct subset of tumor cells  that are defined by the expression of CD34+CD38- surface antigens and have been termed leukemic stem cells. Similar tumor-initiating cells have also been found in 'solid' cancers, such as prostate , breast , brain , lung  colon [8, 9] and gastric cancers . We have recently shown that a rare cell population in human prostate cancer, defined by the phenotype CD133+/α2β1hi (high expression of α2β1 integrin) and comprising less than 0.1% of the tumor mass, has many of the properties of cancer stem cells . In particular, self renewal, extended lifespan (compared to normal stem cells), a high invasive capacity, a primitive epithelial phenotype and an ability to differentiate to recapitulate the phenotypes seen in prostate tumors. The cancer stem cell content was not, however, dependent on prostate tumor clinical stage or grade.
Numerous groups have profiled prostate cancer using DNA microarrays (reviewed in ). Despite this, the genetic changes associated with initiation and progression of this disease remains undefined. Traditionally, expression profiling has focused on sampling the tumor cell mass, but this does not take into account the genetic and phenotypic heterogeneity of tumors. Moreover, individual genes are identified rather than sets of genes that share a biological function. Here we report the first expression profile of a stem cell population from human prostate cancers. By further analyzing this expression signature in the context of biological function, key pathways have been identified that are associated with inflammation, extracellular matrix interactions and stem cell self-renewal.
Identification of gene products associated with a cancer stem cell phenotype
Cancer stem cells express known prostate cancer-associated genes
A panel of genes was selected to confirm the reproducibility of the array data by real time PCR. Comparison of stem versus committed populations demonstrated expression changes in the same direction as the array data, in 10 out of the 12 genes studied (83%), but variations in magnitude were observed (Figure 3b). Similar results were obtained when comparing benign and malignant samples, although some genes (4 out of 12; 33%) did display inconsistencies between microarray and PCR assays (Figure 3c).
Gene expression signature associated with the cancer stem cell phenotype
Candidate genes whose expression is altered in the cancer stem cell population
Stem versus committed*
Malignant versus benign*
Pregnancy-associated plasma protein A, pappalysin 1
Nuclear factor of kappa light polypeptide gene enhancer in B-cells 1 (p105)
Tight junction protein 2 (zona occludens 2)
B-cell translocation gene 1, anti-proliferative
Interleukin 6 (interferon, beta 2)
CASP8 and FADD-like apoptosis regulator
Smu-1 suppressor of mec-8 and unc-52 homolog (C. elegans)
S100 calcium binding protein A3
Chromosome 17 open reading frame 27
RAS and EF-hand domain containing
Integrin, alpha V (vitronectin receptor, alpha polypeptide, antigen CD51)
Interferon gamma receptor 1
Insulin growth factor-like family member 1
Microseminoprotein, beta- (PSP94)
Prostate stem cell antigen
Carcinoembryonic antigen-related cell adhesion molecule 5
S100 calcium binding protein A7 (psoriasin 1)
Hydroxyprostaglandin dehydrogenase 15-(NAD)
Carcinoembryonic antigen-related cell adhesion molecule 7
Trefoil factor 1 (breast cancer, estrogen-inducible sequence expressed in)
Parthenolide treatment affects cancer stem cells but not normal progenitor and stem cell activity
Functional annotation of the cancer stem cell signature
We then extended our search to look for components of these pathways that were present in our gene expression signature, but were not identified by GSEA. This search returned a total of 8 members of the JAK-STAT pathway, 7 components of the extracellular matrix-receptor system and 15 components of the focal adhesion signaling pathway. It is worth noting that five members of the focal adhesion pathway and the extracellular matrix-receptor system overlap, as the focal adhesion pathway is activated by extracellular matrix-receptor interaction.
Despite advances in both screening and in surgical treatment, the long-term prognosis for patients with hormone relapsed prostate cancer remains disappointingly poor . Current tumor targeting strategies for therapy are largely based on differentiation antigens, such as prostate specific antigen and androgen receptor, but our previous studies have shown that the cells that self-renew are a population of primitive cells with the phenotype α2β1hi/CD133+, which are most likely unaffected by current chemotherapeutic regimes . Accordingly, previous expression array studies of prostate have been dominated by androgen receptor-regulated gene products derived from more abundant differentiated cells and the higher average gene expression in these cells is likely to have masked more subtle expression changes in rare cancer stem cells.
Recent advances in microarray technology and target labeling methods have opened up the possibility of performing whole genome transcription profiling experiments from small amounts of starting material, such as rare stem cells . Dumur and colleagues  showed that the GeneChip Two-cycle sample labeling method produced similar results to the standard One-cycle method on 11 out of 12 quality control parameters tested. There was a small bias in the 3'/5' ratio of some genes caused by the generation of shorter products from the Two-cycle labeling method. However, hierarchical clustering showed that each Two-cycle labeled sample was most closely associated with its One-cycle counterpart.
The most striking conclusion from studying highly purified subpopulations from human prostate cancers was the ability of the combined tumor/differentiation cancer stem cell 'signature' to distinguish benign epithelium from tumors with a Gleason 4 morphology . Interestingly, not all Gleason score 7+ cultures expressed the TMPRSS2:ERG fusion , including one lymph node metastasis, yet they clearly clustered away from Gleason 6 cultures (one of which expressed TMPRSS2:ERG). Recently, expression array analysis of micro-dissected prostate tumors has confirmed the hypothesis that the transition to Gleason pattern 4 is associated with significant shifts in gene expression patterns . Lymph node metastases segregated with primary tumors based on the expression signature, but preliminary results indicated that hormone-refractory tumors form a distinct (and possibly more heterogeneous) subgroup in terms of gene expression, as do the Gleason 6 tumors. As the TMPRSS2:ERG gene fusion was detected in one out of two Gleason 6 cultures tested, and is associated with lethal prostate cancer , further study of larger samples of prostate cancer stem cells from different classes of therapy-resistant and Gleason 6 tumors is warranted.
Despite short-term culturing, to expand the stem cell population, the cancer signature was validated by confirming the expression levels of several established prostate cancer markers. Alpha-methylacyl-CoA racemase, a phenotypic marker identified in the first microarray experiments on prostate cancer , was significantly over-expressed in cancer samples, as was MMP9. High MMP expression is consistent with matrix degradation and high invasive capacity previously reported in cancer stem cell cultures . As expected, PTEN showed a modest down-regulation in malignant and stem populations, consistent with the haplo-insufficiency proposed on the basis of transgenic mouse experiments  and in recent studies of hematopoetic tumor stem cells .
Several studies have investigated the differences in gene expression profiles between samples isolated directly from tissue and those from cells cultured in vitro [27–29]. Wick et al.  compared transcriptional profiles from ex vivo and in vitro cultured samples of human dermal lymphatic endothelial cells and blood endothelial cells. These authors found that 2.1% and 4.0% of transcripts were affected by culture in lymphatic endothelial cells and blood endothelial cells, respectively. It is worth noting that this study employed different labeling methods for in vitro and ex vivo samples, which may partially account for the discrepancy. A similar study on hepatic stellate cells highlighted the importance of culture microenvironment and the appropriate use of feeder cells in co-culture. Comparison of transcriptional profiles from hepatic stellate cells cultured in vitro or from cells isolated directly from tissues found substantial differences in the lists of genes found to be differentially expressed. It was shown that co-culture of hepatic stellate cells with Kupffer cells in vitro (acting as feeders) shifted the gene expression profile to a pattern that was consistent with that found in vivo . This suggests that the use of feeder cells in our cultures of cancer stem cells is likely to be important for maintaining gene expression patterns similar to cancer stem cells in vivo.
Expression of multiple genes associated with cell-cell communication and adhesion was associated with the cancer stem cell population. These expression products have been implicated in tissue integrity  and the normal stem cell 'niche' [31, 32]. The gene showing the highest differential expression in the cancer stem cell population was that encoding PAPPA . This pregnancy-associated plasma protein specifically cleaves insulin-like growth factor binding protein (IGFBP)-4 and IGFBP-5. Proteolysis of IGFBPs regulates the bioavailability of IGFs, and because of the association between IGF levels and prostate cancer , strategies for the direct inhibition of IGF signaling, by inhibiting proteolytic activity, is a potential therapeutic strategy and would likely not interfere with insulin signaling .
We used a panel of genes, based on their known association with prostate biology and cancer, to confirm the reproducibility of the array data. Most genes were consistent, but we did note discrepancies, particularly between the malignant and benign RT-PCR results, which may be due to patient variability. In all cases where discrepancy exists, the fold change in expression as measured by RT-PCR was less than two and these small differences are difficult to reproduce accurately. In some cases the absolute expression levels of the genes were quite low, which makes them more sensitive to small fluctuations. The discrepancy could also be caused by the use of probes targeted to different regions of the transcript. Real-time PCR probes are commonly designed against the consensus sequence of the known transcripts for the target gene. Microarrays carry multiple probes against the same gene distributed throughout the length of the transcript, some of which detect only a subset of the known transcripts for the target gene.
Despite this, our data suggest that the transcription factor NF-κB may be a promising therapeutic target as PTL, which acts directly on NF-κB and prevents it entering the nucleus, appeared to promote selective cell death of the cancer-specific CD133 population. Similar results have been demonstrated for leukemic CD34+ stem cells, with normal CD34 cells spared from apoptosis .
Functional annotation of the cancer stem cell signature by GSEA led us to four main pathways: JAK-STAT signaling; cell adhesion and extracellular matrix-interactions; focal adhesion signaling; and WNT signaling. There is a substantial body of work linking Wnt signaling with stemness and malignant behavior (reviewed in ). With respect to prostate cancer, Wnt signaling has been linked to progression to androgen-independence and bone metastasis [37, 38].
Extracellular matrix-receptor signaling and the focal adhesion pathway can be considered part of the same system, as the focal adhesion pathway is activated by extracellular matrix-receptor interaction. Changes in extracellular matrix and associated proteins have been reported in the metastatic progression of prostate cancer , and activation of Focal adhesion kinase through α5β1 integrin/fibronectin has previously been implicated in regulating the invasiveness of prostate cancer cells via activation of phosphatidylinositol-3,4,5-trisphosphate kinase . The JAK-STAT pathway could also be considered to overlap with this system since focal adhesion signaling, as defined in the KEGG database, can be activated by cytokine-cytokine receptor interaction, which is also the major activation method of the JAK-STAT pathway. In addition, JAK-STAT and focal adhesion signaling share several common components, such as the GRB-SOS (growth factor receptor-bound protein 2-son of sevenless) complex and the phosphatidylinositol-3,4,5-trisphosphate kinase/Akt axis. The involvement of IL6 and the JAK-STAT pathway in advanced prostate cancer is well known [41, 42]. More recently, STAT1 has emerged as a potential mediator of drug resistance in prostate cancer  and may present a potential therapeutic target.
Our ability to select and culture stem cell populations will now allow us to determine the genotype of these cells for permanent (mutagenic) changes, such as characteristic translocations  and the presence of epigenetic control . We should also now be able to monitor the effects of novel therapeutics on the cancer stem cell population. Advances in viable cell separation technology and the first detailed expression signature reported here now provide the means to update and ultimately test the cancer stem cell hypothesis in a common non-hematological tumor.
Materials and methods
Tissue collection, isolation, and culture of tumor stem cells
Summary of patient population and invasive characteristics of corresponding stem cell cultures in vitro
% Invasion in vitro*
101 ± 21
99 ± 56
105 ± 29
35 ± 9.5
75 ± 9
119 ± 21
63 ± 18.4
61 ± 4.2
Nested RT-PCR for the detection of the TMPRSS2:ERGfusion
RNA was extracted from prostate tissue using the Qiagen RNeasy kit (Qiagen, Crawley, UK) following the manufacturer's instructions. The RNA was reverse transcribed using random hexamers and reverse transcriptase (Superscript III, Invitrogen, Paisley, UK).
Specific primers were used to detect the presence of the TMPRSS2:ERG fusion by nested RT-PCR (first step, forward 5'-CGC GAG CTA AGC AGG AGG C-3' and reverse 5'-GGC GTT GTA GCT GGG GGT GAG-3'; 2nd step, forward 5'-GGA GCG CCG CCT GGA G-3' and reverse 5'-CCA TAT TCT TTC ACC GCC CAC TCC-3'; Invitrogen). Each PCR reaction contained 1 μM of the respective forward and reverse primers, 1.5 mM MgCl2, 0.2 mM dNTPs and 1 U Taq polymerase (GoTaq, Promega, Southampton, UK). The PCR conditions were adapted from those of Clarke et al. . Briefly, the first step PCR conditions were 94°C for 30 s followed by 35 cycles of 94°C for 20 s and an extension step of 68°C for 1 minute. There was no annealing step as the region amplified is very GC rich. The second step conditions were 94°C for 30 s, 35 cycles of 94°C for 20 s, 66°C for 10 s and 68°C for 1 minute followed by 68°C for 7 minutes.
PCR products were separated by electrophoresis through a 1.5% agarose GelRed (Invitrogen) stained gel for 1 h at 80 V. PCR products were visualized using a Gene Genius bio-imaging system.
Array sample and data processing
Total RNA extraction
Total RNA was extracted from up to 1 × 104 CD133+/α2β1hi selected cells from malignant and non-malignant cultures using Qiagen RNeasy micro-columns according to the manufacturer's protocol. For CD133-/α2β1low cells, total RNA was extracted from between 1 × 105 and 1 × 106 selected cells using Qiagen RNeasy mini-columns. RNA yields were determined spectrophotometrically at 260 nm and RNA integrity checked by capillary electrophoresis using an Agilent 2100 bioanalyzer (Agilent, South Queensferry, UK).
Production of fragmented labeled cRNA
Total RNA (10-50 ng) was amplified using two rounds of cDNA synthesis and in vitro transcription, and biotin labeled by following the Affymetrix small scale labeling protocol VII , omitting the T4 DNA polymerase steps in the two second strand cDNA synthesis reactions and using the Affymetrix GeneChip in vitro transcription labeling kit for the second cycle in vitro transcription for cRNA amplification and labeling. The Affymetrix eukaryotic sample and array processing standard protocol was followed at this stage and the quality of first and second round cRNA products and fragmented cRNA was checked by capillary electrophoresis using an Agilent 2100 bioanalyzer.
Labeled fragmented cRNA (10 μg) was hybridized to oligonucleotide probes on an Affymetrix HG-U133plus2 GeneChip, according to the hybridization, washing, staining and scanning procedure in the Affymetrix eukaryotic sample and array processing standard protocol (Affymetrix Fluidics Station 450 using the EukGE-WS2v5 protocol). Final scanning of the arrays was carried out with an Affymetrix Gene Scanner 3000. The raw data are available in the ArrayExpress Database (accession E-MEXP-993).
Scanned GeneChip images were processed using Affymetrix GCOS 1.2 software to derive an intensity value and flag (present, marginal or absent) for each probe. Probe intensities were derived using the MAS5.0 algorithm. Comparisons between different sample datasets were conducted using Agilent GeneSpring GX software. Datasets to be compared were first normalized using three steps (consecutively applied in the order given): by transforming values <0.01 to 0.01; normalizing each chip to the median of the measurements taken for that chip; and finally normalizing each probe to the median of the measurements for that probe. Low quality or uninformative data were removed using three selections (consecutively applied as follows): probes flagged 'absent' in all samples; probes with standard deviation within a parameter class of >1 in at least three of the four conditions; and probes with less than a two-fold overall change in normalized expression value between all four of the conditions.
The gene expression profile of CD133+/α2β1hi and CD133-/α2β1low prostate cancer cells were compared with benign CD133+/α2β1hi and CD133-/α2β1low prostate epithelial cells. Statistical analysis of the transcription profiles was derived from patients with Gleason score 7 cancers and above. Gleason score 6 biopsies, and one Gleason score 7 biopsy (from a patient who had received hormone therapy) were excluded (Table 1).
Following removal of low quality or uninformative data (see 'Data processing') samples were subjected to a two-way ANOVA test to identify significant (p < 0.05) changes between malignant and benign populations and between stem (CD133+/α2β1hi) and committed (CD133-/α2β1low) populations. Gene expression changes in benign versus malignant cells (within the stem cell population) was compared using a Welch t test. A second Welch t test was used to compare stem and committed populations independent of their disease status. To define signature probesets for the cancer stem cell population, the Benjamini and Hochberg false discovery rate multiple testing correction was applied to the results of Welch t tests between the cell populations, resulting in a corrected critical value of p < 0.035. This value was used in the comparison of stem and committed populations, independent of their disease status, to define a stem cell-specific expression signature. When comparing malignant against benign samples very few probesets were significantly different at p < 0.035, resulting in a combined cell type/malignancy signature that was biased in favor of cell differentiation characteristics. To compensate for this, the critical value was adjusted to p < 0.1 for the comparison of benign and malignant components within the stem cell population. Those genes found to be significantly over-expressed in stem cells were combined with genes significantly over-expressed in malignant samples to generate a malignant stem cell signature.
Quantitative reverse transcriptase PCR
Reverse transcription was carried out on cDNA generated from 50 ng of fractionated cell RNA purified as described above. This was either prepared freshly from RNA or taken from the second round cDNA synthesis for Affymetrix arrays (see above) where starting material was limiting. cDNA generated from the cell lines P4E6  and PC346C (kindly provided by Nefkens Institute, Erasmus University, Rotterdam) was combined in a 1:1 ratio and used to generate the standard curve for each assay. Real time PCR was carried out using TaqMan gene expression pre-synthesized reagents and master mix (Applied Biosystems, Warrington, UK). Reactions were prepared following the manufacturers protocol except that a reduced total volume of 25 μl was used. All reactions were carried out in triplicate on 96-well PCR plates (ABgene, Epsom, UK) in an ABI PRISM 7000 sequence detection system (Applied Biosystems). Standard thermal cycling conditions included a hot start of 5 minutes at 50°C, 10 minutes at 95°C, followed by up to 50 cycles of: 95°C 15 s, 60°C for 1 minute. Data analysis was carried out using ABI SDS software and Microsoft Excel. Expression values are presented relative to the geometric mean of the measurements for three endogenous control genes (GAPDH, ITGB1 and PPIA) in the corresponding samples.
Functional annotation of the prostate cancer stem cell signature
Genes found to be differentially expressed were analyzed for over representation of GO terms to identify important functional categories for further study . Analysis was performed using the PGSEA package for the R environment available through the Bioconductor project [18, 48, 49]. Our analysis was designed to identify GO terms that were significantly over-represented (p < 0.01) in cancer versus benign samples within the stem cell population or within the committed population. We then mapped significant GO terms back to the cancer stem cell signature to identify the individual genes involved. These genes were then searched against the KEGG pathway database  to identify the critical pathways.
Validation by immunocytochemistry and flow cytometry
CD133+/α2β1hi cells were selected from cultured cells before processing for dual-color imaging under confocal microscopy by fixation in a 50:50 mix of ice-cold methanol/acetone or 4% paraformaldehyde in phosphate-buffered saline. After blocking with 20% normal serum in Tris buffered saline, cells were incubated with monoclonal antibodies against the NF-κB p65 subunit (Chemicon International, Hampshire, UK) or a non-specfic isotype control. Appropriate positive control cells were stained in parallel for each antibody. After washing (3 × Tris buffered saline), cells were labeled with Alexa Fluor® 488-tagged secondary antibody (Invitrogen). Cells were mounted in the anti-photobleaching medium Vectashield containing 4',6-diamino-2-phenylindole (DAPI; Vector Laboratories, Peterborough, UK). Cultured cells were processed for dual-color staining flow cytometry as described previously . Cells were co-stained with CD133 (clone 293C; Miltenyi Biotec Ltd, Bisley, UK) and antibodies to Pappalysin 1A (a kind gift from Dr Claus Oxvig, University of Aarhus, Denmark) or ZO-1 (clone ZO1-1A12), ZO-2 (clone 3E8D9; Zymed Laboratories Inc., San Franscisco, CA, USA) and the NF-κB p65 subunit (Chemicon International). Cells were separated on a DakoCytomation CyAn high-performance flow cytometer and analyzed using DakoCytomation Summit version 3.3 software.
Unselected cells were treated for 18 h with increasing concentrations of PTL in the presence of TNFα. Cells were subsequently stained with anti-CD133-APC (anti-CD133-allophycocyanin; Miltenyi Biotec Ltd) for 10 minutes on ice. Cells were then washed in cold magnetic assisted cell sorting (MACS) buffer and resuspended in annexin binding buffer (10 mM HEPES/NaOH, pH 7.4, 140 mM NaCl, 2.5 mM CaCl2). Annexin V-FITC (Pharmingen, Oxford, UK) and 0.25 μg/ml DAPI were then added for 15 minutes before analysis by flow cytometry. The percent viable cells was defined as annexin-V-/DAPI- cells on total (ungated) cells and on gates set for CD133+ populations. The total number of events collected was between 1 × 105 to 1 × 106 depending on the CD133 content of the sample.
benign prostatic hyperplasia
gene set enrichment analysis
insulin-like growth factor
IGF binding protein
Janus activated kinase
Kyoto Encyclopedia of Genes and Genomes
nuclear factor κB
phosphatase and tensin homolog
signal transducer and activator of transcription
tight junction protein
ERG, transmembrane protease, serine 2:v-ets erythroblastosis virus E26 oncogene homolog fusion product
tumor necrosis factor.
This research was supported by program support from Yorkshire Cancer Research and the UK National Cancer Research Institute (NCRI). Further program support was provided by the US Department of Defense (New ideas grant W81xWH-06-1-0082). We would also like to thank Dr Claus Oxvig, University of Aarhus, Denmark, for his gift of antisera against the PAPPA protein and the staff of the Biology Technology Facility, University of York.
- Hamburger AW, Salmon SE: Primary bioassay of human tumor stem cells. Science. 1977, 197: 461-463. 10.1126/science.560061.PubMedView ArticleGoogle Scholar
- Pardal R, Clarke MF, Morrison SJ: Applying the principles of stem-cell biology to cancer. Nat Rev Cancer. 2003, 3: 895-902. 10.1038/nrc1232.PubMedView ArticleGoogle Scholar
- Bonnet D, Dick JE: Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat Med. 1997, 3: 730-737. 10.1038/nm0797-730.PubMedView ArticleGoogle Scholar
- Collins AT, Berry PA, Hyde C, Stower MJ, Maitland NJ: Prospective identification of tumorigenic prostate cancer stem cells. Cancer Res. 2005, 65: 10946-10951. 10.1158/0008-5472.CAN-05-2018.PubMedView ArticleGoogle Scholar
- Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF: Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci USA. 2003, 100: 3983-3988. 10.1073/pnas.0530291100.PubMedPubMed CentralView ArticleGoogle Scholar
- Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, Henkelman RM, Cusimano MD, Dirks PB: Identification of human brain tumor initiating cells. Nature. 2004, 432: 396-401. 10.1038/nature03128.PubMedView ArticleGoogle Scholar
- Kim CF, Jackson EL, Woolfenden AE, Lawrence S, Babar I, Vogel S, Crowley D, Bronson RT, Jacks T: Identification of bronchioalveolar stem cells in normal lung and lung cancer. Cell. 2005, 121: 823-835. 10.1016/j.cell.2005.03.032.PubMedView ArticleGoogle Scholar
- O'Brien CA, Pollett A, Gallinger S, Dick JE: A human colon cancer cell capable of initiating tumor growth in immunodeficient mice. Nature. 2007, 445: 106-110. 10.1038/nature05372.PubMedView ArticleGoogle Scholar
- Ricci-Vitiani L, Lombardi DG, Pilozzi E, Biffoni M, Todaro M, Peschle C, De Maria R: Identification and expansion of human colon-cancer-initiating cells. Nature. 2007, 445: 111-115. 10.1038/nature05384.PubMedView ArticleGoogle Scholar
- Houghton J, Stoicov C, Nomura S, Rogers AB, Carlson J, Li H, Cai X, Fox JG, Goldenring JR, Wang TC: Gastric cancer originating from bone marrow-derived cells. Science. 2004, 306: 1568-1571. 10.1126/science.1099513.PubMedView ArticleGoogle Scholar
- Nelson PS: Predicting prostate cancer behavior using transcript profiles. J Urol. 2004, 172 (Suppl): S28-S32. 10.1097/01.ju.0000142067.17181.68.PubMedView ArticleGoogle Scholar
- Tomlins SA, Rhodes DR, Perner S, Dhanasekaran SM, Mehra R, Sun X-W, Varambally S, Cao X, Tchinda J, Kuefer R, Lee C, Montie JEB, Shah RB, Pienta KJ, Rubin MA, Chinnaiyan AM: Recurrent fusion of TMPRSS2 and ETS transcription factor genes in prostate cancer. Science. 2005, 310: 644-648. 10.1126/science.1117679.PubMedView ArticleGoogle Scholar
- Rubin MA, Zhou M, Dhanasekaran SM, Varambally S, Barrette TR, Sanda MG, Pienta KJ, Ghosh D, Chinnaiyan AM: α-Methylacyl coenzyme A racemase as a tissue biomarker for prostate cancer. JAMA. 2002, 287: 1662-1670. 10.1001/jama.287.13.1662.PubMedView ArticleGoogle Scholar
- Stuart RO, Wachsman W, Berry CC, Wang-Rodriguez J, Wasserman L, Klacansky I, Masys D, Arden K, Goodison S, McClelland M, Wang Y, Sawyers A, Kalcheva I, Tarin D, Mercola D: In silico dissection of cell-type-associated patterns of gene expression in prostate cancer. Proc Natl Acad Sci USA. 2004, 101: 615-620. 10.1073/pnas.2536479100.PubMedPubMed CentralView ArticleGoogle Scholar
- Oxvig C, Sand O, Kristensen T, Gleich GJ, Sottrup-Jensen L: Circulating human pregnancy-associated plasma protein-A is disulfide-bridged to the proform of eosinophil major basic protein. J Biol Chem. 1993, 268: 12243-12246.PubMedGoogle Scholar
- Luo J-L, Kamata H, Karin M: IKK/NF-κB signaling: balancing life and death - a new approach to cancer therapy. J Clin Invest. 2005, 115: 2625-2632. 10.1172/JCI26322.PubMedPubMed CentralView ArticleGoogle Scholar
- Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, Harris MA, Hill DP, Issel-Tarver L, Kasarskis A, Lewis S, Matese JC, Richardson JE, Ringwald M, Rubin GM, Sherlock G: Gene Ontology: tool for the unification of biology. Nature Genet. 2000, 25: 25-29. 10.1038/75556.PubMedPubMed CentralView ArticleGoogle Scholar
- Furge KA, Chen J, Koeman J, Swiatek P, Dykema K, Lucin K, Kahnoski R, Yang XJ, Teh BT: Detection of DNA copy number changes and oncogenic signaling abnormalities from gene expression data reveals MYC activation in high-grade papillary renal cell carcinoma. Cancer Res. 2007, 67: 3171-3176. 10.1158/0008-5472.CAN-06-4571.PubMedView ArticleGoogle Scholar
- Kanehisa M, Goto S, Hattori M, Aoki-Kinoshita KF, Itoh M, Kawashima S, Katayama T, Araki M, Hirakawa M: From genomics to chemical genomics: new developments in KEGG. Nucleic Acids Res. 2006, 34 (Database issue): D354-D357. 10.1093/nar/gkj102.PubMedPubMed CentralView ArticleGoogle Scholar
- Tannock IF, de Wit R, Berry WR, Horti J, Pluzanska A, Chi KN, Oudard S, Théodore C, James ND, Turesson I, Rosethal MA, Eisenberger MA, TAX 327 Investigators: Docetaxel plus prednisone or mitoxantrone plus prednisone for advanced prostate cancer. N Engl J Med. 2004, 351: 1502-1512. 10.1056/NEJMoa040720.PubMedView ArticleGoogle Scholar
- Dumur CI, Garrett CT, Archer KJ, Nasim S, Wilkinson DS, Ferreira-Gonzalez A: Evaluation of a linear amplification method for small samples used on high-density oligonucleotide microarray analysis. Anal Biochem. 2004, 331: 314-321. 10.1016/j.ab.2004.03.040.PubMedView ArticleGoogle Scholar
- Gleason D: Classification of prostate carcinomas. Cancer Chemother Rep. 1996, 50: 125-128.Google Scholar
- True L, Coleman I, Hawley S, Huang C-Y, Gifford D, Coleman R, Beer TM, Gelmann E, Datta M, Mostaghel E, Knudsen B, Lange P, Vessella R, Lin D, Hood L, Nelson PS: A molecular correlate to the Gleason grading system for prostate adenocarcinoma. Proc Natl Acad Sci USA. 2006, 103: 10991-10996. 10.1073/pnas.0603678103.PubMedPubMed CentralView ArticleGoogle Scholar
- Demichelis F, Fall K, Perner S, Andrén O, Schmidt F, Setlur SR, Hoshida Y, Mosquera J-M, Pawitan Y, Lee C, Adami H-O, Mucci LA, Kantoff PW, Andersson S-O, Chinnaiyan AM, Johansson J-E, Rubin MA: TMPRSS2:ERG gene fusion associated with lethal prostate cancer in a watchful waiting cohort. Oncogene. 2007, 26: 4596-4599. 10.1038/sj.onc.1210237.PubMedView ArticleGoogle Scholar
- Trotman LC, Niki M, Dotan ZA, Koutcher JA, Cordon-Cardo C, Pandolfi PP: Pten dose dictates cancer progression in the prostate. PLoS Biol. 2003, 1: e59-10.1371/journal.pbio.0000059.PubMedPubMed CentralView ArticleGoogle Scholar
- Yilmaz OH, Valdez R, Theisen BK, Guo W, Ferguson DO, Wu H, Morrison SJ: Pten dependence distinguishes haematopoietic stem cells from leukaemia-initiating cells. Nature. 2006, 441: 475-482. 10.1038/nature04703.PubMedView ArticleGoogle Scholar
- Liu XS, Zhang ZG, Zhang RL, Gregg SR, Meng H, Chopp M: Comparison of in vivo and in vitro gene expression profiles in subventricular zone neural progenitor cells from the adult mouse after middle cerebral artery occlusion. Neuroscience. 2007, 146: 1053-1061. 10.1016/j.neuroscience.2007.02.056.PubMedPubMed CentralView ArticleGoogle Scholar
- Wick N, Saharinen P, Saharinen J, Gurnhofer E, Steiner CW, Raab I, Stokic D, Giovanoli P, Buchsbaum S, Burchard A, Thurar S, Alitalo K, Kerjaschki D: Transcriptomal comparison of human dermal lymphatic endothelial cells ex vivo and in vitro. Physiol Genomics. 2007, 28: 179-192.PubMedView ArticleGoogle Scholar
- De Menicis S, Seki E, Uchinami H, Kluwe J, Zhang Y, Brenner DA, Schwabe RF: Gene expression profiles during hepatic stellate cell activation in culture and in vivo. Gastroenterology. 2007, 132: 1937-1946. 10.1053/j.gastro.2007.02.033.View ArticleGoogle Scholar
- Matter K, Balda MS: Signalling to and from tight junctions. Nat Rev Mol Cell Biol. 2003, 4: 225-236. 10.1038/nrm1055.PubMedView ArticleGoogle Scholar
- Jones PH, Harper S, Watt FM: Stem cell patterning and fate in human epidermis. Cell. 1995, 80: 83-93. 10.1016/0092-8674(95)90453-0.PubMedView ArticleGoogle Scholar
- Shinohara T, Avarbock MR, Brinster RL: β1- and α6-integrin are surface markers on mouse spermatogonial stem cells. Proc Natl Acad Sci USA. 1999, 96: 5504-5509. 10.1073/pnas.96.10.5504.PubMedPubMed CentralView ArticleGoogle Scholar
- Chan JM, Stampfer MJ, Giovannucci E, Gann PH, Ma J, Wilkinson P, Hennekens CH, Pollak M: Plasma insulin-like growth factor-I and prostate cancer risk: a prospective study. Science. 1998, 279: 563-566. 10.1126/science.279.5350.563.PubMedView ArticleGoogle Scholar
- Gyrup C, Oxvig C: Quantitative analysis of insulin-like growth factor-modulated proteolysis of insulin-like growth factor binding protein-4 and -5 by pregnancy-associated plasma protein-A. Biochemistry. 2007, 46: 1972-1980. 10.1021/bi062229i.PubMedView ArticleGoogle Scholar
- Guzman ML, Rossi RM, Karnischky L, Li X, Peterson DR, Howard DS, Jordan CT: The sequesterpene lactone parthenolide induces apoptosis of human acute myelogeneous leukaemia stem and progenitor cells. Blood. 2005, 105: 4163-4169. 10.1182/blood-2004-10-4135.PubMedPubMed CentralView ArticleGoogle Scholar
- Fodde R, Brabletz T: Wnt/beta-catenin signaling in cancer stemness and malignant behavior. Curr Opin Cell Biol. 2007, 19: 150-158. 10.1016/j.ceb.2007.02.007.PubMedView ArticleGoogle Scholar
- Verras M, Sun Z: Roles and regulation of Wnt signaling and beta-catenin in prostate cancer. Cancer Lett. 2006, 237: 22-32. 10.1016/j.canlet.2005.06.004.PubMedView ArticleGoogle Scholar
- Hall CL, Kang S, MacDougal OA, Keller ET: Role of Wnts in prostate cancer bone metastases. J Cell Biochem. 2006, 97: 661-672. 10.1002/jcb.20735.PubMedView ArticleGoogle Scholar
- Stewart DA, Cooper CR, Sikes RA: Changes in extracellular matrix (ECM) and ECM-associated proteins in the metastatic progression of prostate cancer. Reprod Biol Endocrinol. 2004, 2: 2-10.1186/1477-7827-2-2.PubMedPubMed CentralView ArticleGoogle Scholar
- Zeng ZZ, Jia Y, Hahn NJ, Markwart SM, Rockwood KF, Livant DL: Role of focal adhesion kinase and phosphatidylinositol 3'-kinase in integrin fibronectin receptor-mediated matrix metalloproteinase-1-dependent invasion by metastatic prostate cancer cells. Cancer Res. 2006, 66: 8091-8099. 10.1158/0008-5472.CAN-05-4400.PubMedView ArticleGoogle Scholar
- Culig Z, Steiner H, Bartsch G, Hobisch A: Interleukin-6 regulation of prostate cancer cell growth. J Cell Biochem. 2005, 95: 497-505. 10.1002/jcb.20477.PubMedView ArticleGoogle Scholar
- Mora LB, Buettner R, Seigne J, Diaz J, Ahmad N, Garcia R, Bowman T, Falcone R, Fairclough R, Cantor A, Muro-Cacho C, Livingston S, Karras J, Pow-Sang J, Jove R: Constitutive activation of Stat3 in human prostate tumors and cell lines: direct inhibition of Stat3 signaling induces apoptosis of prostate cancer cells. Cancer Res. 2002, 62: 6659-6666.PubMedGoogle Scholar
- Patterson SG, Wei S, Chen X, Sallman DA, Gilvary DL, Zhong B, Pow-Sang J, Yeatman T, Djeu JY: Novel role of Stat1 in the development of docetaxel resistance in prostate tumor cells. Oncogene. 2006, 25: 6113-6122. 10.1038/sj.onc.1209632.PubMedView ArticleGoogle Scholar
- Feinberg AP, Ohlsson R, Henikoff S: The epigenetic progenitor origin of human cancer. Nat Rev Genet. 2006, 7: 21-33. 10.1038/nrg1748.PubMedView ArticleGoogle Scholar
- Clark J, Merson S, Flohr P, Edwards S, Foster CS, Eeles R, Martin FL, Phillips DH, Crundwell M, Christmas T, Thompson A, Fisher C, Kovacs G, Cooper CS: Diversity of TMPRSS2-ERG fusion transcripts in the human prostate. Oncogene. 2007, 26: 2667-2673. 10.1038/sj.onc.1210070.PubMedView ArticleGoogle Scholar
- Affymetrix GeneChip Expression Analysis Technical Manual VII. [http://www.affymetrix.com/support/technical/manual/expression_manual.affx]
- Maitland NJ, Macintosh CA, Hall J, Sharrard M, Quinn G, Lang S: In vitro models of study cellular differentiation and function in human prostate cancers. Radiat Res. 2001, 155: 133-142. 10.1667/0033-7587(2001)155[0133:IVMTSC]2.0.CO;2.PubMedView ArticleGoogle Scholar
- R Development Core Team: R: a Language and Environment for Statistical Computing. 2006, Vienna, Austria: R Foundation for Statistical ComputingGoogle Scholar
- Gentleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, Dudoit S, Ellis B, Gautier L, Ge Y, Gentry J, Hornik K, Hothorn T, Huber W, Iacus S, Irizarry R, Leisch F, Li C, Maechler M, Rossini AJ, Sawitzki G, Smith C, Tierney L, Yang JY, Zhang J: Bioconductor: open software development for computational biology and bioinformatics. Genome Biol. 2004, 5: R80-10.1186/gb-2004-5-10-r80.PubMedPubMed CentralView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.