Abstract
Housekeeping (HK) genes are involved in basic cellular functions and tend to be constitutively expressed across various tissues and conditions. A number of studies have analyzed the value of HK genes as an internal standard for assessing gene expression, but the role of HK genes in cancer development has never been specifically addressed. In this study, we sought to evaluate the expression of HK genes during prostate tumorigenesis. We performed a meta-analysis of gene expression during the transition from normal prostate (NP) to localized prostate cancer (LPC) (i.e., NP>LPC) and from localized to metastatic prostate cancer (MPC) (i.e., LPC>MPC). We found that HK genes are more likely to be differentially expressed during prostate tumorigenesis than is the average gene in the human genome, suggesting that prostate tumorigenesis is driven by modulation of the expression of HK genes. Cell-cycle genes and proliferation markers were up-regulated in both NP>LPC and LPC>MPC transitions. We also found that the genes encoding ribosomal proteins were up-regulated in the NP>LPC and down-regulated in the LPC>MPC transition. The expression of heat shock proteins was up-regulated during the LPC>MPC transition, suggesting that in its advanced stages, prostate tumor is under cellular stress. The results of these analyses suggest that during prostate tumorigenesis, there is a period when the tumor is under cellular stress and therefore may be the most vulnerable and responsive to treatment.
Keywords: prostate cancer, gene expression, meta-analysis, housekeeping genes
Introduction
Although most genes are constitutively expressed in only a subset of tissues, some genes are required for the maintenance of basal cellular functions and tend to be constitutively expressed across various human tissues and conditions. These genes are called housekeeping (HK) genes 1-3. Several sets of HK genes have been proposed 1, 2, 4, 5.
Results from a number of studies suggest that modulation of the expression of HK genes is associated with cancer. For example, cytochromes are associated with the risk and progression of pancreatic cancer 6, 7, and translation initiation factors play a role in the development of colonic and esophageal cancers 8, 9. In addition, abnormal expression of ribosomal proteins was reported to be a hallmark of colorectal cancer 10. Although these and other results suggest that abnormal expression of the HK genes plays an important role in carcinogenesis, a systematic assessment of their role in cancer development had not been yet conducted.
We therefore performed a meta-analysis of publicly available gene expression data to evaluate the expression of HK genes during prostate cancer progression. Our results revealed an association between modulation of expression of HK genes and tumor progression and also suggest that the development of prostate cancer is driven by modulated expression of HK genes.
Material and methods
HK genes
We used 3 lists of HK genes for our meta-analysis. The first list comprises genes showing constant expression in various human tissues 4. Because that study was reported in Trends in Genetics, we called this set of HK genes TG_HK. The second list was obtained by using a naïve Bayes classifier with specific functional characteristics of HK genes 1; we called this set of HK genes NB_HK. The third list of genes came from an article by de Jonge et al. 2. The authors identified a set of 15 genes encoding for ribosomal proteins as the most stably expressed across multiple cell types and different experimental conditions. Thus we called this third set of HK genes 15_HK.
These 3 lists overlap substantially. We found that 124 (22%) of the HK genes were the same in the TG_HK and NB_HK lists. And among the 15 most stably expressed genes from the de Jonge et al. paper 2, 6 were also reported in the TG_HK list and 14, in the NB_HK list.
Gene expression datasets
The list of the gene expression datasets we used in this study is shown in Table I. Datasets were retrieved from the Oncomine database 11, accessed December 15, 2008. We used 18 datasets in total: 11 for the transition from normal prostate (NP) to localized prostate cancer (LPC) and 7 for the transition from LPC to metastatic prostate cancer (MPC).
Table I.
Datasets used for our meta-analysis
| Name | PMID | Sample description (number of samples and metastatic sites*) | Number of genes |
|---|---|---|---|
| Dhanasekaran_Prostate | 11518967 | Normal Prostate (22) | 9,956 |
| Primary Prostate Cancer (59) | |||
| Dhanasekaran_Prostate_2 | 15548588 | Normal Adjacent Prostate (12) | 19,650 |
| Prostate Cancer (25) | |||
| Holzbeierlein_Prostate | 14695335 | Normal Prostate (4) | 5,854 |
| Prostate Cancer (23) | |||
| Lapointe_Prostate | 14711987 | Normal Prostate (41) | 19,116 |
| Prostate Carcinoma (62) | |||
| Luo_Prostate | 11406537 | Benign Hyperplasia (9) | 6,500 |
| Prostate Carcinoma (16) | |||
| Nanni_Prostate | 16513839 | Normal Prostate (3) | 22,283 |
| Prostate Carcinoma (22) | |||
| Tomlins_Prostate | 17173048 | Benign Prostate (22) | 19,355 |
| Prostate Carcinoma (30) | |||
| Vanaja_Prostate | 12873976 | Normal Prostate (8) | 44,928 |
| Prostate Adenocarcinoma (27) | |||
| Varambally_Prostate | 16286247 | Benign Prostate (6) | 54,675 |
| Prostate Carcinoma (7) | |||
| Welsh_Prostate | 11507037 | Normal Prostate (9) | 11,138 |
| Prostate Carcinoma (25) | |||
| Yu_Prostate | 15254046 | Normal Prostate (23) | 12,625 |
| Prostate Carcinoma (64) | |||
| Dhanasekaran_Prostate | 11518967 | Primary Prostate Cancer (59) | 9,935 |
| Metastatic Prostate Cancer (20) | |||
| A (1), B (5), LN (5), LG (4), ST (5) | |||
| Dhanasekaran_Prostate_2 | 15548588 | Prostate Cancer (25) | 18,502 |
| Metastatic Prostate Cancer (6) | |||
| B (3), LG (3) | |||
| Tomlins_Prostate | 17173048 | Prostate Carcinoma (30) | 19,337 |
| Metastatic Prostate Cancer (19) | |||
| B (5), LG (5), LN (4), LI (5) | |||
| Yu_Prostate | 15254046 | Prostate Carcinoma (64) | 12,625 |
| Metastatic Prostate Cancer (25) | |||
| Distant sites, not specified | |||
| Vanaja_Prostate | 12873976 | Prostate Adenocarcinoma (27) | 44,928 |
| Metastatic Prostate Cancer (5) | |||
| B (2), LG (1), LN (1), LI (1) | |||
| Holzbeierlein_Prostate | 14695335 | Prostate Cancer (23) | 6,475 |
| Metastatic Prostate Cancer (9) | |||
| B (2), LG (5), ST (2) | |||
| LaTulippe_Prostate | 12154061 | Prostate Carcinoma (23) | 12,600 |
| Metastatic Prostate Cancer (9) | |||
| B (4), LG (3), ST (2) |
A, adrenal; B, bone; LG, lung; LN, lymph node; LI, liver.
Pretreatment samples from primary tumors were used for this study.
Meta-analysis
To estimate the overall statistical significance based on the statistical significance of the individual tests, we used an extension of Stouffer's method 12. This approach is based on estimating the standard normal deviation, Z. The individual probability, p, is first converted to a Z score, and the Z scores are summed up across studies. This sum is divided by the square root of the number of tests (k). The sum of normal deviates is itself a normal deviate and can be back-transformed into an overall p; the probability level associated with the sum of Z yields an overall level of significance. The advantage of this method lies in its increased power: if, for example, several tests consistently favored the research question but failed to reach the level of significance because of small sample size, the overall test would more easily become significant. This approach is similar to the approach recently proposed by Ochsner et al. 13. The selection of the method of meta-analysis for our study was dictated by the available data: for the majority of the datasets the raw gene expression data were not available, while t-tests and corresponding p values were easily accessible for individual probes.
Distribution of the genes by pathways
We used the Database for Annotation, Visualization, and Integrated Discovery (DAVID) 14 to assess the distribution of the differentially expressed genes across the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways 15, 16. DAVID tests the null hypothesis that the genes are uniformly distributed across the pathways. The p values characterize the statistical evidence for the clustering of the genes by the pathways; the lower the p value, the stronger the statistical evidence that the genes tend to cluster in a specific pathway. Student's t-test was used to compare mean expression values between primary tumors with various Gleason scores.
Results
Proportions of the up- and down-regulated genes in the NP>LPC and LPC>MPC transitions
We estimated the proportion of up- and down-regulated genes in the NP>LPC and LPC>MPC transitions. Genes with global p values of ≤10−6 were considered to be differentially expressed. We chose this threshold as being a conservative correction for multiple testing with 19,582 as the total number of genes in the NP>LPC transition and 17,902 as the total number in the LPC>MPC transition.
Table II shows the numbers of up- and down-regulated genes. We found that 2,415 (12.3%) of the genes were differentially expressed in the NP>LPC transition. In the LPC>MPC transition, the percentage of differentially expressed genes was lower: 1,574 genes (8.8%). Significant genes in the NP>LPC transition were largely up-regulated—55% of the differentially expressed genes—whereas in the LPC>MPC transition, the significant genes were largely down-regulated—67% of all significant genes.
Table II.
Number of up- and down-regulated genes in the NP>LPC and LPC>MPC transitions*
| Transition | Total no. of genes | Significant genes |
Total no. of significant genes | |
|---|---|---|---|---|
| Up-reg. | Down-reg. | |||
| NP>LPC | 19,583 | 1,337 (0.55) | 1,078 (0.45) | 2,415 |
| LPC>MPC | 17,902 | 520 (0.33) | 1 054 (0.67) | 1,574 |
Parenthetic numbers are the proportions of up- and down-regulated genes.
HK genes are more likely than the average gene in the human genome to be up- or down-regulated
On the basis of what we know about the expression of the HK genes, we expected them to be relatively stably expressed in prostate tumorigenesis. However, the results of our analysis revealed the opposite; the HK genes had a higher proportion of differentially expressed genes than the average gene in the human genome: 173/563 = 0.31 vs. 2,394/17,859 = 0.13 (χ2 = 135.1, df =1, p << 0.001) for TG_HK genes (Fig. 1). The results were similar for the 3 lists of HK genes. We found that in the NP>LPC transition, HK genes were more likely than the average gene in the human genome to be up-regulated: 122/563 = 0.22 vs.1,319/17,859 = 0.07 (χ2 = 152.5, df =1, p << 0.001). In the LPC>MPC transition, HK genes were more likely to be down-regulated than the average gene in the human genome: 65/563 = 0.12 vs. 1,052/17,859 = 0.06 (χ2 = 29.7, df =1, p << 0.001).
Figure 1.
The proportions of up- and down-regulated housekeeping genes in the normal prostate to localized prostate cancer (NP>LPC) and LPC to metastatic prostate cancer (LPC>MPC) transitions. The red horizontal lines represent the overall proportions of significant genes. Error bars indicate standard error (SE). We used TG_HK genes for this analysis. The results of the analyses of NP_HK and 15_HK genes were very similar.
Different functional categories of HK genes may have different expression patterns in prostate tumor progression
To look at the expression of the HK genes in more detail, we stratified them into 8 categories: 1) ribosomal proteins, 2) cell communication genes, 3) protein expression (translation) genes, 4) metabolism genes, 5) genes associated with transcription (gene expression), 6) cell structure and motility genes, 7) defense genes, and 8) cell division genes. This stratification was done according to Haverty et al. 3. (The lists of the genes in each category are in the Supporting Information.)
Table III shows the proportions of up- and down-regulated genes for each category. For most of the functional categories, the expression patterns (i.e., the direction and number of differentially expressed genes) were similar in the NP>LPC and LPC>MPC transitions. HK genes encoding for ribosomal proteins, however, had extremely different patterns of gene expression: in the NP>LPC transition, all ribosomal genes were up-regulated, but in the LPC>MPC transition, all differentially expressed ribosomal genes were down-regulated. We also found that the expression pattern of the genes involved in cell communication differed in the NP>LPC and LPC>MPC transitions. Cell communication genes were preferentially up-regulated in the NP>LPC transition and down-regulated in the LPC>MPC transition.
Table III.
Expression of the HK genes stratified by functional categories
| Type of HK genes | No. of genes | NP>LPC transition* | LPC>MPC transition* | χ 2 | p | ||||
|---|---|---|---|---|---|---|---|---|---|
| Unchanged | Up-reg. | Down-reg. | Unchanged | Up-reg. | Down-reg. | ||||
| Ribosomal protein | 59 | 29 (0.49) | 30 (0.51) | 0 | 45 (0.76) | 0 | 14 (0.24) | 25.5 | 1.2 × 10–7 |
| Cell communication | 64 | 46 (0.72) | 12 (0.19) | 6 (0.09) | 55 (0.86) | 0 | 9 (0.14) | 9.2 | 0.003 |
| Protein expression | 37 | 28 (0.75) | 8 (0.22) | 1 (0.03) | 32 (0.86) | 1 (0.03) | 4 (0.11) | 3.7 | 0.06 |
| Metabolism | 92 | 67 (0.73) | 15 (0.16) | 10 (0.11) | 72 (0.78) | 8 (0.09) | 12 (0.13) | 1.4 | 0.24 |
| Gene expression | 34 | 28 (0.82) | 5 (0.15) | 1 (0.03) | 28 (0.82) | 1 (0.03) | 5 (0.15) | 1.1 | 0.32 |
| Cell structure | 33 | 23 (0.7) | 3 (0.09) | 7 (0.21) | 28 (0.85) | 0 | 5 (0.15) | 1 | 0.31 |
| Defense | 35 | 28 (0.8) | 2 (0.06) | 5 (0.14) | 26 (0.74) | 0 | 9 (0.26) | 0.2 | 0.7 |
| Cell division | 14 | 11 (0.65) | 1 (0.14) | 2 (0.21) | 11 (0.89) | 1 (0.07) | 2 (0.14) | 0 | 1 |
| Total | 368 | 260 | 76 | 32 | 297 | 11 | 60 | ||
Parenthetic numbers are the proportions.
Gleason score and the expression of ribosomal genes
To pinpoint the time when the expression of ribosomal genes changes from up- to down-regulation, we analyzed their expression in clinically localized cancers with different Gleason scores. We used gene expression data from 3 studies: Luo 17, Singh et al. 18, and Welsh et al. 19. First, we found that the ribosomal genes are expressed at a higher level than the average gene in the human genome is, which is consistent with their housekeeping role. We also found that expression of the ribosomal genes increased as the Gleason sum score increased from 5 (2 + 3) to 8 (4 + 4) (Fig. 2). The average expression level of the ribosomal genes significantly dropped, however, as the Gleason sum score increased from 8 (4 + 4) to 9 (4 + 5) (t = 2.7, df =14, p = 0.02), suggesting that the turning point from up- to down-regulation lies between Gleason score 8 (4 + 4) and 9 (4 + 5).
Figure 2.
Dependence of the expression of ribosomal genes on Gleason score. Error bars indicate standard error (SE).
Clustering of the differentially expressed genes in KEGG pathways
According to the changes in their expression during the NP>LPC and LPC>MPC transitions, we subdivided all genes into 8 groups: 1) genes up-regulated in the NP>LPC transition, with no difference in the expression level in the LPC>MPC transition, 2) down-regulated in NP>LPC, with no difference in LPC>MPC, 3) up-regulated in LPC>MPC, with no difference in NP>LPC, 4) down-regulated in LPC>MPC, with no difference in NP>LPC, 5) up-regulated in both NP>LPC and LPC>MPC, 6) up- regulated in NP>LPC and down-regulated in LPC>MPC, 7) down-regulated in NP>LPC and up-regulated in LPC>MPC, and 8) down-regulated in both NP>LPC and LPC>MPC.
We then analyzed the distribution of the genes in each category by KEGG pathways. Table IV shows the results of the analysis with p values after Benjamini corrections (DAVID's adjustment on multiple testing) 14. We found that the genes that were significant in only 1 of the transitions (i.e., the first 4 groups) showed no significant clustering by pathway. However, the genes significant in both transitions showed strong clustering. More often than one would expect, the genes up-regulated in both transitions fell in the cell cycle pathway. The cell cycle genes BUB1B, MAD2L1, CDC2, PTTG1, E2F3, CDC6, CCNA2, ORC2L, MCM4, MDM2, CCNE2, GSK3B, CCNB2, BUB1, CDK4, CCNB1, and PRKDC include several cycling and cycling-dependent kinases, suggesting that the proliferation rate increases during prostate tumorigenesis. The genes that were down-regulated in both NP>LPC and LPC>MPC transitions were clustered in the focal-adhesion and actin cytoskeleton–regulation pathways. These genes include the integrins and cadherins ITGA2, ITGA5, ITGA7, ITGA8, ITGB1, ITGB8, CDH3, and CDH19.
Table IV.
Clustering of genes that are significantly up- or down-regulated in the NP>LPC and LPC>MPC transitions by the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways54
| NP>LPC* | LPC>MPC* | Significant KEGG pathway |
|---|---|---|
| Up (613) | ||
| Down (439) | ||
| Up (249) | ||
| Down (292) | ||
| Up (400) | Up (400) | Cell cycle (p = 3.8 × 10–5) |
| Up (684) | Down (684) | Ribosome (p = 2.2 × 10–4) |
| Down (266) | Up (266) | |
| Down (657) | Down (657) | Focal adhesion (p = 1.2 × 10–6) |
| Regulation of actin cytoskeleton (p = 1.6 × 10–4) |
Numbers of significant genes are in parentheses.
Expression of cell proliferation markers
To predict the cell proliferation rate, we estimated the expression of 13 cell proliferation markers: TOP2A, MKI67, PLK1, MCM7, CENPF, MCM6, PCNA, FEN1, MCM2, WT1, CLSPN, CCND1, and MIB1. There is published evidence that the expression of all these markers positively correlates with the proliferation rate 20-32.
We found that in the NP>LPC transition, all of the proliferation markers except MCM6 were up-regulated, and in the LPC>MPC transition, all proliferation markers, including MCM6, were up-regulated (Table V). This suggests that the cell proliferation rate increases during prostate tumor progression.
Table V.
Expression of cell proliferation markers during prostate tumor progression
| Markers | NP>LPC transition | LPC>MPC transition | ||||
|---|---|---|---|---|---|---|
| Z score | Direction | p | Z score | Direction | p | |
| TOP2A | 7.09 | Up | 9.8 × 10–13 | 12.19 | Up | 3.1 × 10–34 |
| CENPF | 5.96 | Up | 2 × 10–9 | 4.54 | Up | 5.6 × 10–6 |
| MKI67 | 4.09 | Up | 4.3 × 10–5 | 7.92 | Up | 1.9 × 10–15 |
| MCM7 | 2.45 | Up | 0.0142 | 4.57 | Up | 4.9 × 10–6 |
| FEN1 | 2.12 | Up | 0.0336 | 2.16 | Up | 0.0305 |
| MCM6 | –2.01 | Down | 0.0444 | 4.09 | Up | 4.2 × 10–5 |
| PLK1 | 1.85 | Up | 0.0644 | 7.71 | Up | 1.1 × 10–14 |
| PCNA | 1.74 | Up | 0.0816 | 3.25 | Up | 0.0012 |
| CCND1 | –1.58 | Down | 0.1132 | 1.77 | Up | 0.0771 |
| MCM2 | 1.55 | Up | 0.1218 | 2.06 | Up | 0.0392 |
| MIB1 | –0.91 | Down | 0.3647 | –0.84 | Down | 0.3983 |
| CLSPN | –0.13 | Down | 0.8993 | –1.96 | Down | 0.0500 |
| WT1 | 0.12 | Up | 0.9059 | 2.02 | Up | 0.0434 |
Heat shock genes
Numerous studies demonstrated that stress changes the expression of multiple genes, with most being down-regulated.33, 34 Because gene expression is mostly suppressed in the LPC>MPC transition, one can suppose that the prostate tumor is under stress during this time. We know that increased expression of the heat shock genes indicates cellular stress 35-37. We therefore assessed the expression of heat shock proteins in the NP>LPC and LPC>MPC transitions. The 34 heat shock genes were retrieved from the NCBI database (Supporting Information). We found that in the NP>LPC transition, 16 heat shock genes were differentially expressed: 9 (56%) were up-regulated, and 7 (44%) were down-regulated. These proportions are very similar to the overall proportions of up- and down-regulated genes in the NP>LPC transition: 55% and 45%, respectively. In the LPC>MPC transition, 14 genes were differentially expressed; 10 of them (71%) were up-regulated. The excess of up-regulated heat shock genes was statistically significant (χ2 = 4.3, df = 1, p = 0.01), suggesting that during the transition from localized to metastatic prostate cancer, the tumor cells are under stress.
Discussion
We found that HK genes are more likely than non–HK genes to be differentially expressed in the NP>LPC or LPC>MPC transition (Fig. 1). Because the HK genes are involved in essential cellular functions, it is unlikely that their differential expression is the result of a ripple effect induced by other genes in driving tumor development. We believe that modulation of the expression of HK genes is a driving force behind prostate tumorigenesis, and it is their differential expression that may induce a ripple effect on multiple downstream targets. This suggests the need to interpret the results of gene expression profiling with caution because the modulated expression of a given gene might be a consequence of the ripple effect induced by other genes rather than being directly related to the phenotype.
We found that the genes tended to be up-regulated in the NP>LPC transition and down-regulated in the LPC>MPC transition. The difference may be driven by overall changes in translation and transcription during prostate tumorigenesis. We also found that ribosomal proteins are up-regulated in the NP>LPC and down-regulated in the LPC>MPC transition: all 30 significant genes were up-regulated in NP>LPC, and all 14 significant genes were down-regulated in the LPC>MPC transition.
Our analysis suggested that the cell proliferation rate increases during prostate tumor progression: the cell proliferation markers were up-regulated in both the NP>LPC and LPC>MPC transitions. We also noted that the genes that were up-regulated in the transitions are clustered at the KEGG cell cycle pathway. Many of the cell cycle genes encoded for cyclins and cyclin-dependent kinases, suggesting that prostate tumor progression is associated with an increased proliferation rate 38, 39.
Cell proliferation requires an increased level of transcription and translation, 2 processes that provide building blocks for new cells. In the NP>LPC transition, expression of general transcription factors and ribosomal proteins was increased, with most of the genes differentially expressed in NP>LPC transition being up-regulated. We found, however, that despite the increased proliferation rate, transcription (expression level) and translation (biogenesis of ribosomes) were shut down in the LPC>MPC transition. This is counterintuitive because elevated cell proliferation requires an increase in the biogenesis of ribosomes 40-42.
Global suppression of gene expression in the advanced stages of prostate cancer may be a result of the increasing cell proliferation, which surpasses the ability of the tumor vascular system to provide oxygen. We found that expression of 2 of 3 hypoxia-inducible factors, HIF3A and HIF1AN, was significantly increased in the LPC>MPC transition, with p values 0.04 and 0.005, respectively, whereas no difference was found in the NP>LPC transition. Hypoxia and shortages of nutrients and growth factors can induce a stress reaction in advanced prostate tumors. As expected, we found that the expression of stress proteins, which are associated with inhibition of transcription and translation 43-45, is up-regulated in the LPC>MPC transition.
Of note, genes that were down-regulated in the NP>LPC and LPC>MPC transitions clustered into the cell adhesion– and actin cytoskeleton–regulatory pathways. These 2 pathways are mechanistically connected 46-48. The reports of a number of studies suggested that modulation of cell adhesion plays an important role in the progression and metastasis of prostate cancer 49-51.
Figure 3 summarizes our findings on the changes in transcription and/or translation, proliferation rate, and cellular stress during prostate tumorigenesis. On the basis of these findings, one can expect that an advanced tumor is under cellular stress and might be more vulnerable to environmental insults. Since translation is down-regulated at this stage, one can expect that its further suppression can kill tumor cells while causing little or no harm to normal cells with a normal proliferation rate. Currently the available chemotherapies used to treat advanced prostate cancer target DNA replication by introducing chemical links between DNA strands or by blocking DNA replication. Our results suggest that suppression of translation, alone or in combination with suppression of DNA replication, may be beneficial. It is interesting that the results from several studies suggest that inhibition of translation has an anticancer effect. For example, it was demonstrated that the anticancer effect of eicosapentaenoic acid is associated with inhibited initiation of translation 52. It has also been demonstrated that the anticancer effect of the tumor suppressor Pdcd4 is mediated through inhibition of translation 53. These findings suggest that inhibition of translation at a time when the tumor is under cellular stress, with already down-regulated transcription and translation, may improve survival and response to treatment.
Figure 3.
Changes in transcription and/or translation, proliferation rate, and cellular stress during prostate tumorigenesis. Gray bar represents a putative period of increased vulnerability of prostate cancer.
In conclusion, during prostate tumorigenesis, HK genes are more likely than the average gene in the human genome to be differentially expressed, suggesting that modulation of basic cellular functions drives cancer development. The level of transcription and translation increases during the initial stages of prostate tumorigeneses and decreases during the transition from localized to metastatic prostate cancer. During this stage, prostate cancer is under cellular stress and therefore might be more sensitive to treatment.
Acknowledgments
This study was supported by the David Koch Center for Applied Research in Genitourinary Cancer.
Abbreviations
- DAVID
Database for Annotation, Visualization, and Integrated Discovery
- HK
housekeeping
- KEGG
Kyoto Encyclopedia of Genes and Genomes
- LPC
localized prostate cancer
- MPC
metastatic prostate cancer
- NP
normal prostate
Footnotes
Additional Supporting Information may be found in the online version of this article.
References
- 1.De Ferrari L, Aitken S. Mining housekeeping genes with a Naive Bayes classifier. BMC Genomics. 2006;7:277. doi: 10.1186/1471-2164-7-277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.de Jonge HJ, Fehrmann RS, de Bont ES, Hofstra RM, Gerbens F, Kamps WA, de Vries EG, van der Zee AG, te Meerman GJ, ter Elst A. Evidence based selection of housekeeping genes. PLoS ONE. 2007;2:e898. doi: 10.1371/journal.pone.0000898. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Haverty PM, Weng Z, Best NL, Auerbach KR, Hsiao LL, Jensen RV, Gullans SR. HugeIndex: a database with visualization tools for high-density oligonucleotide array data from normal human tissues. Nucleic Acids Res. 2002;30:214–7. doi: 10.1093/nar/30.1.214. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Eisenberg E, Levanon EY. Human housekeeping genes are compact. Trends Genet. 2003;19:362–5. doi: 10.1016/S0168-9525(03)00140-9. [DOI] [PubMed] [Google Scholar]
- 5.Hsiao LL, Dangond F, Yoshida T, Hong R, Jensen RV, Misra J, Dillon W, Lee KF, Clark KE, Haverty P, Weng Z, Mutter GL, et al. A compendium of gene expression in normal human tissues. Physiol Genomics. 2001;7:97–104. doi: 10.1152/physiolgenomics.00040.2001. [DOI] [PubMed] [Google Scholar]
- 6.Suzuki H, Morris JS, Li Y, Doll MA, Hein DW, Liu J, Jiao L, Hassan MM, Day RS, Bondy ML, Abbruzzese JL, Li D. Interaction of the cytochrome P4501A2, SULT1A1 and NAT gene polymorphisms with smoking and dietary mutagen intake in modification of the risk of pancreatic cancer. Carcinogenesis. 2008;29:1184–91. doi: 10.1093/carcin/bgn085. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Luby TM. Targeting cytochrome P450 CYP1B1 with a therapeutic cancer vaccine. Expert Rev Vaccines. 2008;7:995–1003. doi: 10.1586/14760584.7.7.995. [DOI] [PubMed] [Google Scholar]
- 8.Taylor CA, Sun Z, Cliche DO, Ming H, Eshaque B, Jin S, Hopkins MT, Thai B, Thompson JE. Eukaryotic translation initiation factor 5A induces apoptosis in colon cancer cells and associates with the nucleus in response to tumour necrosis factor alpha signalling. Exp Cell Res. 2007;313:437–49. doi: 10.1016/j.yexcr.2006.09.030. [DOI] [PubMed] [Google Scholar]
- 9.Salehi Z, Mashayekhi F. Expression of the eukaryotic translation initiation factor 4E (eIF4E) and 4E-BP1 in esophageal cancer. Clin Biochem. 2006;39:404–9. doi: 10.1016/j.clinbiochem.2005.11.007. [DOI] [PubMed] [Google Scholar]
- 10.Lai MD, Xu J. Ribosomal proteins and colorectal cancer. Curr Genomics. 2007;8:43–9. doi: 10.2174/138920207780076938. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Rhodes DR, Kalyana-Sundaram S, Mahavisno V, Varambally R, Yu J, Briggs BB, Barrette TR, Anstet MJ, Kincead-Beal C, Kulkarni P, Varambally S, Ghosh D, et al. Oncomine 3.0: genes, pathways, and networks in a collection of 18,000 cancer gene expression profiles. Neoplasia. 2007;9:166–80. doi: 10.1593/neo.07112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Rosenthal R. The file drawer problem and tolerance for null results. Psychological Bulletin. 1979;86:638–41. [Google Scholar]
- 13.Ochsner SA, Steffen DL, Hilsenbeck SG, Chen ES, Watkins C, McKenna NJ. GEMS (Gene Expression MetaSignatures), a Web resource for querying meta-analysis of expression microarray datasets: 17beta-estradiol in MCF-7 cells. Cancer Res. 2009;69:23–6. doi: 10.1158/0008-5472.CAN-08-3492. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Dennis G, Jr., Sherman BT, Hosack DA, Yang J, Gao W, Lane HC, Lempicki RA. DAVID: Database for Annotation, Visualization, and Integrated Discovery. Genome Biol. 2003;4:P3. [PubMed] [Google Scholar]
- 15.Hashimoto K, Goto S, Kawano S, Aoki-Kinoshita KF, Ueda N, Hamajima M, Kawasaki T, Kanehisa M. KEGG as a glycome informatics resource. Glycobiology. 2006;16:63R–70R. doi: 10.1093/glycob/cwj010. [DOI] [PubMed] [Google Scholar]
- 16.Kanehisa M. The KEGG database. Novartis Found Symp. 2002;247:91–101. discussion -3, 19-28, 244-52. [PubMed] [Google Scholar]
- 17.Luo JH. Gene expression alterations in human prostate cancer. Drugs Today (Barc) 2002;38:713–9. doi: 10.1358/dot.2002.38.10.704653. [DOI] [PubMed] [Google Scholar]
- 18.Singh D, Febbo PG, Ross K, Jackson DG, Manola J, Ladd C, Tamayo P, Renshaw AA, D'Amico AV, Richie JP, Lander ES, Loda M, et al. Gene expression correlates of clinical prostate cancer behavior. Cancer Cell. 2002;1:203–9. doi: 10.1016/s1535-6108(02)00030-2. [DOI] [PubMed] [Google Scholar]
- 19.Welsh JB, Sapinoso LM, Su AI, Kern SG, Wang-Rodriguez J, Moskaluk CA, Frierson HF, Jr., Hampton GM. Analysis of gene expression identifies candidate markers and pharmacological targets in prostate cancer. Cancer Res. 2001;61:5974–8. [PubMed] [Google Scholar]
- 20.Yuan J, Horlin A, Hock B, Stutte HJ, Rubsamen-Waigmann H, Strebhardt K. Polo-like kinase, a novel marker for cellular proliferation. Am J Pathol. 1997;150:1165–72. [PMC free article] [PubMed] [Google Scholar]
- 21.Warbrick E, Coates PJ, Hall PA. Fen1 expression: a novel marker for cell proliferation. J Pathol. 1998;186:319–24. doi: 10.1002/(SICI)1096-9896(1998110)186:3<319::AID-PATH184>3.0.CO;2-2. [DOI] [PubMed] [Google Scholar]
- 22.Tsimaratou K, Kletsas D, Kastrinakis NG, Tsantoulis PK, Evangelou K, Sideridou M, Liontos M, Poulias I, Venere M, Salmas M, Kittas C, Halazonetis TD, et al. Evaluation of claspin as a proliferation marker in human cancer and normal tissues. J Pathol. 2007;211:331–9. doi: 10.1002/path.2095. [DOI] [PubMed] [Google Scholar]
- 23.Todorov IT, Werness BA, Wang HQ, Buddharaju LN, Todorova PD, Slocum HK, Brooks JS, Huberman JA. HsMCM2/BM28: a novel proliferation marker for human tumors and normal tissues. Lab Invest. 1998;78:73–8. [PubMed] [Google Scholar]
- 24.Tanaka Y, Kanai F, Tada M, Tateishi R, Sanada M, Nannya Y, Ohta M, Asaoka Y, Seto M, Shiina S, Yoshida H, Kawabe T, et al. Gain of GRHL2 is associated with early recurrence of hepatocellular carcinoma. J Hepatol. 2008;49:746–57. doi: 10.1016/j.jhep.2008.06.019. [DOI] [PubMed] [Google Scholar]
- 25.Schrader C, Janssen D, Klapper W, Siebmann JU, Meusers P, Brittinger G, Kneba M, Tiemann M, Parwaresch R. Minichromosome maintenance protein 6, a proliferation marker superior to Ki-67 and independent predictor of survival in patients with mantle cell lymphoma. Br J Cancer. 2005;93:939–45. doi: 10.1038/sj.bjc.6602795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Olszewski M, Huang W, Chou PM, Duerst R, Kletzel M. Wilms' tumor 1 (WT1) gene in hematopoiesis: a surrogate marker of cell proliferation as a possible mechanism of action? Cytotherapy. 2005;7:57–61. doi: 10.1080/14653240510018046. [DOI] [PubMed] [Google Scholar]
- 27.Li SS, Xue WC, Khoo US, Ngan HY, Chan KY, Tam IY, Chiu PM, Ip PP, Tam KF, Cheung AN. Replicative MCM7 protein as a proliferation marker in endometrial carcinoma: a tissue microarray and clinicopathological analysis. Histopathology. 2005;46:307–13. doi: 10.1111/j.1365-2559.2005.02069.x. [DOI] [PubMed] [Google Scholar]
- 28.Konstantinidou AE, Korkolopoulou P, Kavantzas N, Mahera H, Thymara I, Kotsiakis X, Perdiki M, Patsouris E, Davaris P. Mitosin, a novel marker of cell proliferation and early recurrence in intracranial meningiomas. Histol Histopathol. 2003;18:67–74. doi: 10.14670/HH-18.67. [DOI] [PubMed] [Google Scholar]
- 29.Jalava P, Kuopio T, Juntti-Patinen L, Kotkansalo T, Kronqvist P, Collan Y. Ki67 immunohistochemistry: a valuable marker in prognostication but with a risk of misclassification: proliferation subgroups formed based on Ki67 immunoreactivity and standardized mitotic index. Histopathology. 2006;48:674–82. doi: 10.1111/j.1365-2559.2006.02402.x. [DOI] [PubMed] [Google Scholar]
- 30.Heck MM, Earnshaw WC. Topoisomerase II: A specific marker for cell proliferation. J Cell Biol. 1986;103:2569–81. doi: 10.1083/jcb.103.6.2569. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Diop S, Letestu R, Orsolani D, Leboeuf Y, Le Tutour P, Thiam D, Diakhate L, Valensi F. [Expression of proliferation marker Ki 67 in chronic lymphocytic leukemia]. Dakar Med. 2005;50:65–8. [PubMed] [Google Scholar]
- 32.Dietrich DR. Toxicological and pathological applications of proliferating cell nuclear antigen (PCNA), a novel endogenous marker for cell proliferation. Crit Rev Toxicol. 1993;23:77–109. doi: 10.3109/10408449309104075. [DOI] [PubMed] [Google Scholar]
- 33.Karssen AM, Her S, Li JZ, Patel PD, Meng F, Bunney WE, Jr., Jones EG, Watson SJ, Akil H, Myers RM, Schatzberg AF, Lyons DM. Stress-induced changes in primate prefrontal profiles of gene expression. Mol Psychiatry. 2007;12:1089–102. doi: 10.1038/sj.mp.4002095. [DOI] [PubMed] [Google Scholar]
- 34.Vina J, Borras C, Gomez-Cabrera MC, Orr WC. Part of the series: from dietary antioxidants to regulators in cellular signalling and gene expression. Role of reactive oxygen species and (phyto)oestrogens in the modulation of adaptive response to stress. Free Radic Res. 2006;40:111–9. doi: 10.1080/10715760500405778. [DOI] [PubMed] [Google Scholar]
- 35.Sherman M, Multhoff G. Heat shock proteins in cancer. Ann N Y Acad Sci. 2007;1113:192–201. doi: 10.1196/annals.1391.030. [DOI] [PubMed] [Google Scholar]
- 36.Li Z, Srivastava P. Heat-shock proteins. Curr Protoc Immunol. 2004 doi: 10.1002/0471142735.ima01ts58. Appendix 1:Appendix 1T. [DOI] [PubMed] [Google Scholar]
- 37.Calderwood SK, Ciocca DR. Heat shock proteins: stress proteins with Janus-like properties in cancer. Int J Hyperthermia. 2008;24:31–9. doi: 10.1080/02656730701858305. [DOI] [PubMed] [Google Scholar]
- 38.Murphy C, McGurk M, Pettigrew J, Santinelli A, Mazzucchelli R, Johnston PG, Montironi R, Waugh DJ. Nonapical and cytoplasmic expression of interleukin-8, CXCR1, and CXCR2 correlates with cell proliferation and microvessel density in prostate cancer. Clin Cancer Res. 2005;11:4117–27. doi: 10.1158/1078-0432.CCR-04-1518. [DOI] [PubMed] [Google Scholar]
- 39.Lu S, Lee J, Revelo M, Wang X, Lu S, Dong Z. Smad3 is overexpressed in advanced human prostate cancer and necessary for progressive growth of prostate cancer cells in nude mice. Clin Cancer Res. 2007;13:5692–702. doi: 10.1158/1078-0432.CCR-07-1078. [DOI] [PubMed] [Google Scholar]
- 40.Thomas G. An encore for ribosome biogenesis in the control of cell proliferation. Nat Cell Biol. 2000;2:E71–2. doi: 10.1038/35010581. [DOI] [PubMed] [Google Scholar]
- 41.Montanaro L, Mazzini G, Barbieri S, Vici M, Nardi-Pantoli A, Govoni M, Donati G, Trere D, Derenzini M. Different effects of ribosome biogenesis inhibition on cell proliferation in retinoblastoma protein- and p53-deficient and proficient human osteosarcoma cell lines. Cell Prolif. 2007;40:532–49. doi: 10.1111/j.1365-2184.2007.00448.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Ayrault O, Andrique L, Larsen CJ, Seite P. [The negative regulation of ribosome biogenesis: a new Arf-dependent pathway controlling cell proliferation?] Med Sci (Paris) 2006;22:519–24. doi: 10.1051/medsci/2006225519. [DOI] [PubMed] [Google Scholar]
- 43.Duncan R, Hershey JW. Heat shock-induced translational alterations in HeLa cells. Initiation factor modifications and the inhibition of translation. J Biol Chem. 1984;259:11882–9. [PubMed] [Google Scholar]
- 44.Brostrom CO, Brostrom MA. Regulation of translational initiation during cellular responses to stress. Prog Nucleic Acid Res Mol Biol. 1998;58:79–125. doi: 10.1016/s0079-6603(08)60034-3. [DOI] [PubMed] [Google Scholar]
- 45.Connolly E, Braunstein S, Formenti S, Schneider RJ. Hypoxia inhibits protein synthesis through a 4E-BP1 and elongation factor 2 kinase pathway controlled by mTOR and uncoupled in breast cancer cells. Mol Cell Biol. 2006;26:3955–65. doi: 10.1128/MCB.26.10.3955-3965.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Dustin ML. Cell adhesion molecules and actin cytoskeleton at immune synapses and kinapses. Curr Opin Cell Biol. 2007;19:529–33. doi: 10.1016/j.ceb.2007.08.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Le Clainche C, Carlier MF. Regulation of actin assembly associated with protrusion and adhesion in cell migration. Physiol Rev. 2008;88:489–513. doi: 10.1152/physrev.00021.2007. [DOI] [PubMed] [Google Scholar]
- 48.Margadant C, van Opstal A, Boonstra J. Focal adhesion signaling and actin stress fibers are dispensable for progression through the ongoing cell cycle. J Cell Sci. 2007;120:66–76. doi: 10.1242/jcs.03301. [DOI] [PubMed] [Google Scholar]
- 49.Jennbacken K, Gustavsson H, Welen K, Vallbo C, Damber JE. Prostate cancer progression into androgen independency is associated with alterations in cell adhesion and invasivity. Prostate. 2006;66:1631–40. doi: 10.1002/pros.20469. [DOI] [PubMed] [Google Scholar]
- 50.Mason MD, Davies G, Jiang WG. Cell adhesion molecules and adhesion abnormalities in prostate cancer. Crit Rev Oncol Hematol. 2002;41:11–28. doi: 10.1016/s1040-8428(01)00171-8. [DOI] [PubMed] [Google Scholar]
- 51.Shah GV, Thomas S, Muralidharan A, Liu Y, Hermonat PL, Williams J, Chaudhary J. Calcitonin promotes in vivo metastasis of prostate cancer cells by altering cell signaling, adhesion, and inflammatory pathways. Endocr Relat Cancer. 2008;15:953–64. doi: 10.1677/ERC-08-0136. [DOI] [PubMed] [Google Scholar]
- 52.Palakurthi SS, Fluckiger R, Aktas H, Changolkar AK, Shahsafaei A, Harneit S, Kilic E, Halperin JA. Inhibition of translation initiation mediates the anticancer effect of the n-3 polyunsaturated fatty acid eicosapentaenoic acid. Cancer Res. 2000;60:2919–25. [PubMed] [Google Scholar]
- 53.LaRonde-LeBlanc N, Santhanam AN, Baker AR, Wlodawer A, Colburn NH. Structural basis for inhibition of translation by the tumor suppressor Pdcd4. Mol Cell Biol. 2007;27:147–56. doi: 10.1128/MCB.00867-06. [DOI] [PMC free article] [PubMed] [Google Scholar]