Logo of amjpatholAmerican Journal of Pathology For AuthorsAmerican Journal of Pathology SubscribeAmerican Journal of Pathology SearchAmerican Journal of Pathology Current IssueAmerican Journal of Pathology About the JournalAmerican Journal of Pathology
Am J Pathol. 2002 Jun; 160(6): 2169–2180.
PMCID: PMC1850818

Decrease and Gain of Gene Expression Are Equally Discriminatory Markers for Prostate Carcinoma

A Gene Expression Analysis on Total and Microdissected Prostate Tissue

Abstract

Information on over- and underexpressed genes in prostate cancer in comparison to adjacent normal tissue was sought by DNA microarray analysis. Approximately 12,600 mRNA sequences were analyzed from a total of 26 tissue samples (17 untreated prostate cancers, 9 normal adjacent to prostate cancer tissues) obtained by prostatectomy. Hierarchical clustering was performed. Expression levels of 63 genes were found significantly (at least 2.5-fold) increased, whereas expression of 153 genes was decreased (at least 2.5-fold) in prostate cancer versus adjacent normal tissue. In addition to previously described genes such as hepsin, overexpression of several genes was found that has not drawn attention before, such as the genes encoding the specific granule protein (SGP28), α-methyl-acyl-CoA racemase, low density lipoprotein (LDL)-phospholipase A2, and the anti-apoptotic gene PYCR1. The radiosensitivity gene ATDC and the genes encoding the DNA-binding protein inhibitor ID1 and the phospholipase inhibitor uteroglobin were significantly down-regulated in the cancer samples. DNA microarray data for eight genes were confirmed quantitatively in five normal and five cancer tissues by real-time reverse transcriptase-polymerase chain reaction with a high correlation between the two methods. Laser capture microdissection of epithelial and stromal compartments from cancer and histological normal specimens followed by an amplification protocol for low levels of RNA (<0.1 μg) allowed us to distinguish between gene expression profiles characteristic of epithelial cells and those typical of stroma. Most of the genes identified in the nonmicrodissected tumor material as up-regulated were indeed overexpressed in cancerous epithelium rather than in the stromal compartment. We conclude that development of prostate cancer is associated with down-regulation as well as up-regulation of genes that show complex differential regulation in epithelia and stroma. Some of the gene expression alterations identified in this study may prove useful in the development of novel diagnostic and therapeutic strategies.

Adenocarcinoma of the prostate is the most frequent cancer in males in western countries. Of major public health importance is the development of reliable discriminating strategies for early detection of neoplastic disease. 1

Levels of serum prostate-specific antigen, a serine protease, frequently serve as a diagnostic marker for prostate cancer, although elevated concentrations can also be found in benign prostatic hyperplasia and acute and chronic inflammation. 2 Histopathological diagnosis of prostate carcinoma is still regarded as the decisive standard in clinical practice. Tumors are graded as proposed by Gleason. 3 This grading system relies on histological patterns of glandular differentiation. Patient group survival can be determined quite reliably when grading is used in combination with tumor stage. 4 Morphologically similar tumor types can show different biological behavior, however.

Precancerous lesions are referred to as prostatic intraepithelial neoplasia. Although prostatic intraepithelial neoplasias constitute highly predictive markers for adenocarcinoma, prostatic intraepithelial neoplasias are based on diagnostic criteria that are subject to a certain degree of subjectivity as is diagnosis of different degrees of epithelial dysplasias in general. 5 Postatrophic hyperplasia, which can be characterized by small densely packed glands with an increased nuclear/cytoplasm ratio, sometimes can be difficult to distinguish from prostatic adenocarcinoma. 6 Postatrophic hyperplasia or proliferative inflammatory atrophy have been implicated in prostatic carcinogenesis. 6,7

Gene expression profiling is now being considered as an objective supplementary approach to the histopathological work-up of precancerous or cancerous lesions of the prostate. Using high-density microarrays with a large collection of cDNAs or gene-specific oligonucleotides one can identify marker genes or clusters of genes the altered expression of which is characteristic of specific stages of tumor disease. 8-11

Laser-assisted microdissection of atypical glandular structures and subsequent analysis of multiple genes with DNA arrays or of single marker genes by quantitative real-time reverse transcriptase-polymerase chain reaction (RT-PCR) is a powerful refinement to gene expression profiling protocols and is likely to enhance the diagnostic value of gene expression data. 12 This approach excludes contribution of RNA from fibromuscular tissue and tumor-infiltrating mononuclear cells to the gene expression profile. An additional advantage is that it may be potentially applicable to prostate biopsies obtained in preoperative diagnostic procedures.

In this study gene expression profiles were generated from adenocarcinoma of the prostate and from adjacent normal tissue resected from patients not previously treated by chemotherapy or radiotherapy. Profiling of microdissected glands and stroma, both normal and cancerous, was also performed. The results show that carcinoma can be differentiated from histological nontumorous prostate by both significant increases as well as decreases in expression of specific genes, some of which have not been identified previously in conjunction with gene expression patterns in prostate cancer.

An association between marker gene expression and carcinoma may be used to enhance the diagnostic value of the pattern-oriented histological grading system that is currently in use. Analysis of gene expression profiles can also reveal metabolic or signal transduction pathways that might be targeted by new therapeutic strategies.

Materials and Methods

Prostate Tissue

Prostate cancer tissue samples were obtained from patients who had undergone radical prostatectomy for prostate cancer. None of the patients included in this study had received preoperative hormonal therapy, chemotherapy, or radiation therapy. Seventeen primary cancers and 9 normal adjacent to cancer tissues were examined. Collection of tissue and use for this study were approved according to standard guidelines by the ethics committee of the Medical Faculty of the University of Heidelberg. Table 1 [triangle] shows ages, pre- and postoperative serum prostate-specific antigen concentrations, Gleason scores, and staging of all patients from whom prostate tissue was obtained for this study. After radical prostatectomy, tissues were flash-frozen in liquid nitrogen and stored at −80°C. Seven-μm sections were cut with a standard cryostat and stained with hematoxylin and eosin to identify tumor-free (N1 to N9) and tumorous tissue parts (T1 to T17); cancerous tissue was graded according to the Gleason scoring system 3 by a pathologist. The nonmalignant samples contained predominantly epithelial cells and relatively low amounts of fibromuscular stroma cells; nevertheless the ratio of epithelial cells to stroma was higher in cancer than in tumor-free parts of prostate cancers.

Table 1.
Clinicopathological Information

RNA Isolation and Oligonucleotide Array Hybridization

Total RNA was extracted by the method of Chomczynski and Sacchi. 13 RNA quality was monitored by agarose gel electrophoresis; 20 μg of total RNA were reverse-transcribed using Superscript II reverse transcriptase (Life Technologies Inc., Gaithersburg, MD), then converted into double-stranded cDNA, and biotin-labeled during in vitro transcription from the T7 promoter using the ENZO RNA Transcript labeling kit; all reactions were performed essentially according to the Affymetrix protocol (Affymetrix, Sunnydale, CA). Each sample was tested for RNA integrity by hybridization to Affymetrix Test2 Chips. Only cRNA samples that passed this test were used on Human Genome U95A chips (HG-U95A, ~12,600 sequences; the list of genes is available at www.affymetrix.com). The default average intensity of all mRNAs on a chip was uniformly set at 1000, signals below a signal intensity of 200 were disregarded. Under such conditions, the reproducibility of two identical samples (T7 and T7R) from one tumor resulted in a correlation coefficient of r = 0.98 (not shown). Usually, ~60% of the sequences on a HG-U95A chip gave a present call.

Quantitative Real-Time RT-PCR for Confirmation of Microarray Data

RT-PCR products from five cases of the normal group (N1, N3, N4, N5, N6) and five cases of the cancer group (T10, T14, T15, T16, T17) were used to confirm the microarray data by quantitative real-time RT-PCR. The PCR reactions were performed in the LightCycler apparatus using the LC-FastStart DNA Master SYBR Green I kit (Roche Diagnostics, Mannheim, Germany).

Two μg of the same total RNA used for microarray assay were used for the first-strand cDNA synthesis with Superscript II reverse transcriptase and oligo d(T)12–18 primer according to the manufacturer’s protocol (Life Technologies).

The primer sequences used in this study are given in Table 2 [triangle] . We used eight genes (five genes found to be increased in microarray assay and three that were decreased) for confirmation by the LightCycler. After optimizing of all PCR reactions at the same annealing temperature of 60°C, thermocycling for each reaction was performed in a final volume of 20 μl containing 2 μl of cDNA sample, 4 mmol/L MgCl2, 0.5 μmol/L of each primer, and 2 μl of LC-FastStart DNA Master SYBR Green I. After 480 seconds of initial denaturation at 95°C, the cycling conditions of 45 cycles for each gene consisted of denaturation at 95°C for 15 seconds, annealing at 60°C (for all genes) for 5 seconds, elongation at 72°C for 10 seconds, and a short temperature increase to 82°C for 3 seconds (for fluorescence measurement). For preparing the standard curve, we used GAPDH as the reference gene because it showed similar expression levels in normal and cancer samples (data from microarray assay) and it was amplified with an efficiency similar to seven of eight genes that were confirmed by RT-PCR. Serial dilutions (1:10, 1:100, 1:1000) were prepared from each cDNA sample and GAPDH was amplified. Expression levels of all other genes are given relative to the expression levels of GAPDH by evaluation of their crossing-over points of product accumulation curves relative to the standard curve of GAPDH. All PCR products were checked by melting point analysis and by gel electrophoresis to verify that products were of the correct lengths. Several of the PCR products were cycle-sequenced to confirm their identity.

Table 2.
Sequences of Primers Used in This Study

Laser-Assisted Microdissection for Oligonucleotide Array

Microdissection was performed using P.A.L.M. microlaser technology (P.A.L.M. GmbH, Bernried, Germany) on frozen sections stained by hematoxylin to obtain ~20,000 pooled epithelial and stromal cells each from five histological normal (N2, N4, N5, N6, N7) and five cancerous samples (T8, T9, T10, T14, T17). For the normal epithelium the procedure was performed twice to hybridize the resulting labeled cRNA to two different chips and determine reproducibility. Total RNA was extracted by the Chomczynski and Sacchi 13 method as above; however, in view of the small amounts of total RNA expected, 2 μl of Pellet Paint (Novagen, Darmstadt, Germany) were added as a co-precipitant before RNA precipitation. RNA quality was checked on RNA lab chips by the Bioanalyzer 2100 Lab-On-A-Chip system (Agilent Technologies, Palo Alto, CA) and found to be excellent, with a 2:1 ratio of 28S to 18S RNA. From peak heights obtained in a separate pilot study with 200, 100, and 20 ng of total mouse RNA, the RNA amounts were estimated at between 20 ng and 100 ng in four samples (two normal epithelium, one tumor epithelium, one tumor stroma), whereas it was ~20 ng in the normal stroma sample. These RNA amounts were amplified by the protocol of Baugh and colleagues, 14 with the following modification: a temperature of 50°C was used in the RT step (total volume 10 μl) because this resulted in a higher yield of cDNAs. This modification had been verified previously by using 8 μg (standard Affymetrix protocol) versus 200 or 20 ng of mouse total RNA that were amplified and hybridized to Affymetrix MG-U74A chips (M. Kenzelmann and colleagues, manuscript in preparation). The yields of biotinylated cRNAs from the microdissected samples were between 7.2 (from the lowest RNA amount) and 11.5 μg. For hybridization, 7.2 μg of each of the five samples were hybridized to Affymetrix Test3 chips, which demonstrated that the amplified cRNAs were of good quality; the samples were then hybridized to HG-U95A chips as described above.

Laser-Assisted Microdissection for Quantitative Real-Time RT-PCR

Approximately 4000 epithelial and stromal cells each were microdissected as above for RT-PCR quantification of two selected messages, the GRO2 oncogene and fractalkine transcripts. These were quantified separately for three normal (N2, N4, N7) and three cancer samples (T8, T10, T17). The primer sequences used are given in Table 2 [triangle] . Reverse transcription was performed as described above; real-time quantitative RT-PCR was performed on a TaqMan ABI 7700 Sequence Detection System (Applied Biosystems, Weiterstadt, Germany) using heat-activated TaqDNA polymerase (AmpliTaq Gold, Applied Biosystems). After an initial hold of 2 minutes at 50°C and 10 minutes at 95°C the samples were cycled 40 times at 95°C for 15 seconds and 60°C for 60 seconds. The cDNA content of each sample was compared with another sample following the δ-δ-Ct technique. 15,16 Similar amplification efficiencies for targets and housekeeping genes were demonstrated by analyzing serial cDNA dilutions showing an absolute value of the slope of log input cDNA amount versus δ-CT (=Ct > housekeeping gene − Ct target) of <0.1.

Statistics

We selected genes for clustering according to Welsh and colleagues. 11 Data from all microarrays [28 samples including 2 repetitions for testing the reproducibility of the method (N2R and T7R)] were first analyzed by Affymetrix software (Data Mining Tool) for genes with the highest standard deviation (SD) (SD > 2500). This list of genes was used in Gene Spring software (Silicon Genetics, Redwood, CA) to perform a hierarchical clustering analysis 17 without any information given on histopathology of the prostate samples. A second statistical analysis was also performed on the whole data set of the normal adjacent to cancer samples (N1 to N9) and for the 10 cancer samples assigned by clustering as most distinct from normal (T8 to T17) because both of these two groups appeared as relatively homogeneous in their gene expression patterns internally. Genes with moderate to high expression levels and with a fold change >2.5 between normal and cancer gene groups and a P < 0.05 by Student’s t-test were identified. The t-test used the two-tailed distribution and the heteroscedastic type (two samples, unequal variance).

Results

Expression Profiles of mRNA Species from Prostate Cancer and Normal Adjacent to Cancer Samples

Tumor tissues from prostate cancer patients (n = 17) were compared to the histological normal tissues (n = 9) of prostatectomized patients. Tissues from the peripheral region were used so that comparisons between the peripheral area and central regions (where benign prostatic hyperplasia is preferentially located) would be avoided.

The reproducibility of our procedures for expression profiling on DNA microarrays was tested in two ways: first, two neighboring areas (N2 and N2R) of normal tissue from the same patient were extracted for total RNA; after conversion of both RNA samples to biotin-labeled cRNA according to the Affymetrix protocol and hybridization to two chips, the correlation coefficient was r = 0.95. Second, one-and-the-same sample from a tumor area (T7 and T7R) was also hybridized to two chips yielding a correlation coefficient of r = 0.98.

When the expression data from the 28 tissues (a tumor and a normal sample in duplicate, respectively) were analyzed statistically, a list of genes could be defined that showed very large standard deviations (>2500). Unsupervised clustering with this discriminatory set of sequences could distinguish between tumor and nontumor samples on the basis of gene expression patterns alone (Figure 1) [triangle] with the exception of two tumors that were grouped with normal adjacent to cancer samples. In a second analysis, 10 tumor samples appearing most homogeneous in their gene expression pattern were compared to the nine normal samples; genes with a greater than 2.5-fold difference in expression were grouped; 63 genes were found with significantly increased RNA levels, and 153 genes were detected with a significant decrease in RNA levels; surprisingly the down-regulated genes were 2.4-fold more numerous than the elevated genes (Tables 3 and 4) [triangle] [triangle] .

Figure 1.
Hierarchical clustering analysis of the 129 genes with the highest SD. Rows represent individual genes; columns represent individual prostate samples (normal adjacent to cancer samples, N1 to N9; cancer samples, T1 to T17). Each cell in the matrix represents ...
Table 3.
Genes Overexpressed in Prostate Cancer
Table 4.
Genes Underexpressed in Prostate Cancer (50 of 153 Found Genes Are Shown)

Tables 3 and 4 [triangle] [triangle] include very recently described putative marker mRNAs as well as new ones described here as significantly elevated, eg, SGP28, α-methylacyl-CoA racemase, PYCR1, and significantly down-regulated, eg, ID1, ATDC, and uteroglobin.

Confirmation of Array Data by Quantitative Real-Time RT-PCR

Eight different gene transcript species including hepsin, which has been reported previously by several investigators to be elevated in prostate cancer, were selected for confirmation of the array data by real-time RT-PCR (Figure 2) [triangle] . Of the eight genes chosen, five were overexpressed in prostate cancer: α-methylacyl-CoA racemase, LDL-phospholipase A2, hepsin, pyrroline 5-carboxylate reductase 1, transcriptional regulator ERG. The remaining three genes were underexpressed: uteroglobin (inhibitor of phospholipase), ataxia telangiectasia group D-associated protein (ATDC), and DNA-binding protein inhibitor ID1. For all increased RNAs, corroboration by real-time PCR was obtained (correlation coefficients of r = 0.72 to 0.96) indicating quantitative agreement between oligonucleotide array data and RT-PCR data. Decreased RNA results also were confirmed for uteroglobin and ATDC, whereas ID1 was decreased in three of five cancer samples with an overall correlation coefficient of r = 0.52.

Figure 2.
Confirmation of microarray data by quantitative real-time RT-PCR. RT-PCR products from five cases of the normal adjacent to cancer group (N1, N3, N4, N5, N6) and five cases of the cancer group (T10, T14, T15, T16, T17) were used for confirmation. These ...

Expression Profiles from Microdissected Samples

To achieve an analysis of up- or down-regulated genes not only in bulk tissue but also in epithelial or stromal cells of tumors and of normal tissue adjacent to cancer, laser-assisted microdissection was performed on cancerous glands and stromal areas, respectively, from five advanced tumors (T8, T9, T10, T14, T17) and glands and stroma from five normal adjacent to cancer tissues (N2, N4, N5, N6, N7). The excised areas were pooled and extracted for RNA as described before; RNAs were tested for their quality on RNA electrophoresis chips and showed no evidence of RNA degradation. Between 20 and 100 ng of RNA were amplified by the protocol of Baugh and colleagues 14 with minor modifications. When the two pools of normal epithelia (NE1, NE2) were compared on oligonucleotide microarrays the correlation coefficient was r = 0.95 indicating excellent reproducibility of the method. Evaluation of the three pools of normal [two normal epithelial cells (NE1, NE2), one normal stroma] and separately of the two pools of tumor tissue [tumor epithelial cells (TE) and tumor stromal cells] indicated that many of the genes were differentially expressed (by at least fourfold) in epithelial versus stromal cells (see Tables 3 and 4 [triangle] [triangle] ). A comparison of the tumor epithelium microarray (TE) with the two normal epithelium microarrays (NE1, NE2) is given in Table 5 [triangle] for genes with a greater than 15-fold difference in gene expression.

Table 5.
Comparison of Microdissected Tumor Epithelial Areas with Microdissected Normal Epithelial Areas

To corroborate the microarray data from microdissected epithelial and stromal cells real-time PCR of cDNA products from epithelial and stromal cells from three tumor (T8, T10, T17) and three normal samples (N2, N4, N7) was performed for the messages of two chemokines, GRO2 and fractalkine. In addition to their chemotactic properties on mononuclear cells, GRO2 and fractalkine are thought to possess mitogenic/oncogenic (GRO2) and anti-apoptotic activities (fractalkine). 18,19 GRO2 and fractalkine mRNAs were down-regulated in tumor epithelium and surrounding stroma (Figure 3) [triangle] .

Figure 3.
Quantitative real-time RT-PCR of microdissected samples for two down-regulated genes. Approximately 4,000 epithelial and stromal cells each were microdissected and RT-PCR products were quantified separately for three normal adjacent to cancer (N2, N4, ...

Cytogenetic Position of Overexpressed or Underexpressed Genes

Several genes over- or underexpressed in prostate cancer were found to be in close genomic proximity. Four exemplary bands are shown for chromosomes 6, 7, and 21 (Table 6) [triangle] .

Table 6.
Cytogenetic Positions of Genes with Significantly (P <0.05) Increased or Decreased Gene Expression in Prostate Cancer

Discussion

The gene expression analysis on 17 prostate carcinomas revealed several genes that showed increases in expression of up to 21-fold (ie, specific granule protein SGP28) and decreases of up to 12-fold (ie, complement factor B) in comparison to adjacent nontumorous tissue (Tables 3 and 4) [triangle] [triangle] . These genes, some of which have already been described, may serve as markers of prostate carcinoma. Our data on total prostate tissue confirm and extend the findings of recently published microarray studies using oligonucleotide array technology. 8-11,20 Our study confirms a statistically significant overexpression of the gene for hepsin (HPN, X07732, a transmembrane serine protease) in the majority of tumors (Table 3) [triangle] . Overexpression of the message for pyrroline-5-carboxylate reductase 1 (PYCR1, M77836) was reported before as a p53-dependent protein in a DNA microarray study on cultured tumor cells, 21 and we now have found this message overexpressed in prostate cancer. It is of note that the compound pyrroline-5-carboxylate 1 induced apoptosis in cultured cells overexpressing PYCR1. 21 Prostate-specific membrane antigen, a novel folate hydrolase associated with prostatic carcinogenesis and metastasis, 22 also was up-regulated in tumor epithelium in our study. Of particular note in the present study is that down-regulated genes may be at least as informative as up-regulated genes in characterizing prostate cancer. Loss or a decrease of gene expression may be relevant in early prostate cancers, that were studied. This corresponds to results in cytogenetic studies that have described preferentially losses of genes in early stages of prostate cancer. 23,24 Down-regulation of certain genes may prove to be directly involved in fostering tumor development or metastasis. Bullous pemphigoid antigen 1 is a component of hemidesmosome plaque 1; its reduction may loosen epithelial adherence. The mRNAs of cytoskeletal proteins such as keratins 5, 15, and 7 as components of basal cells, which are not part of atypical prostatic epithelium, were significantly down-regulated.

In addition, a novel aspect of this report is the microarray analysis on microdissected tissues and the usefulness of this approach in revealing compartment-specific expression profiles. When the new protocol for amplification of ng amounts of total RNA was tested for consistency using 10 μg versus 20 ng of mouse thymus RNA, only 230 of 12,600 (~2%) of the sequences behaved as true outliers (at least fourfold change). Several of the overrepresented sequences were short and contained repetitions of A-rich segments (M. Kenzelmann and colleagues, in preparation). Microdissection gene expression analysis can potentially be done on prostate biopsies as the described RNA amplification procedure can be performed on a few hundred cells. This approach would divulge the gene expression pattern of tumor cells without fibromuscular tissue and without inflammatory mononuclear cells. It thus might be feasible to define differences in marker gene expression also between prostate intraepithelial neoplasia and prostate carcinoma; this may provide supplementary information on the pathogenesis of precancerous prostate lesions.

Gene expression analysis alone cannot provide an overall integrative molecular understanding of the genesis and growth of prostate carcinoma because chromosomal aberrations, translational control of messages, and posttranslational modifications, to name just a few molecular events, certainly play a major role in the biological behavior of prostate carcinoma. Table 6 [triangle] shows chromosomal locations of some genes that were found to be significantly up- or down-regulated in our study. At four chromosomal locations 6p21, 7p14, 7q21, and 21q22, aberrantly regulated genes were densely concentrated. The dense location of genes with altered expression in prostate cancer at one site of chromosome 21 is interesting because this chromosome has the lowest gene density of all chromosomes. These gene expression hot spots seem to overlap for chromosome 7 with the chromosomal positions found to be amplified or deleted by comparative genomic hybridization; there are no similarities in location with comparative genomic hybridization for the other two chromosomes 6 and 21. Molecular cytogenetic analysis has identified common sites of chromosomal alterations in prostate cancer: gains at 7p, 7q, 8q, Xq, and losses at 5q, 6q, 8p, 13q, 16q, and 18q. 23-26

New hypotheses on prostatic carcinogenesis may be entertained by inspection of the gene expression changes identified in the current study, in particular by the array data from microdissected epithelial and stromal cells. Apart from messages for putative new markers mentioned before, several themes emerged from the compilation of up- and down-regulated genes: first, the up-regulated expression of several members of the histone family (Tables 3 and 6) [triangle] [triangle] we observed may indicate a dysregulation of chromatin structure. 27 Second, transcription factors and growth factors were not uniformly up-regulated. Several transcription factors/activators and growth factors actually were clearly down-regulated (Table 4) [triangle] . Also the down-regulation of suppressor and growth arrest genes can be seen as particularly important in the pathogenesis of early prostate cancers that were studied. As varying numbers of mononuclear infiltrates can be found in prostate carcinoma, it was expected that stromal and/or epithelial cells might express chemokines. For example, surprisingly the mRNA of chemokine GRO2 (growth-related oncogene) and of the cell-bound CX3C chemokine fractalkine were found decreased as well as the messages for chemokine receptor 2, receptor for the β-chemokine MCP-1. Besides their role in the induction and propagation of inflammatory reactions the two tested chemokines can exhibit growth factor-like (GRO2) and anti-apoptotic (fractalkine) activities. 18,19 The decrease of GRO2 and fractalkine was observed in tumor epithelium and tumor stroma, as revealed by microdissection and real-time PCR. This also shows that the observed down-regulation of genes in bulk tumor tissue cannot simply be explained by the nonpresentation of stromal genes in tumors that have a relative low stroma content. Third, several mRNAs coding for proteins in fat and steroid metabolism were dysregulated in prostatic carcinoma in this study, including up-regulation of α-methylacyl-CoA racemase 10 that catalyzes the degradation of branched fatty acids and C27 steroids 28 and of LDL phospholipase A2 that generates short fatty acids (up-regulated sixfold). Fatty acid metabolism and steroids have been implicated in prostate carcinogenesis. Enzymes in fatty acid synthesis have been proposed as targets for anti-neoplastic therapy. 29,30 Uteroglobin (a polychlorinated biphenyl-binding protein) was found down-regulated in the prostate carcinoma samples. This protein can inhibit phospholipase A2 activity. 31 It is able to disrupt the generation of platelet-activating factor and has been reported to reduce the growth of an adenocarcinoma cell line. 32 In accordance with another report 33 the message of cyclooxygenase-2, a catalytic enzyme for the synthesis of inflammatory and carcinogenic prostaglandin derivatives, was not found up-regulated in cancerous tissue. Cyclooxygenase-2 apparently is important in the development of proliferative inflammatory atrophy.

In summary we showed here that mRNA expression analysis with microarray identified a set of genes that characterize prostate cancer and normal tissue adjacent to cancer. Down-regulated genes were found to be more numerous than up-regulated genes and are equally suited for differentiation of bulk normal from cancerous tissues. Array analysis of microdissected epithelia and stromal cells localized the majority of up-regulated genes to cancerous epithelia. These data can be used to design an array with a restricted number of cDNA or oligonucleotides for the study of larger sample sets and eventually supplementary diagnostic purposes. The current study demonstrated that even small amounts of prostate mRNA can be used for array expression profiling. Microarray studies can be performed on microdissected tissue separating tumor epithelium and stroma and may give more insight into cellular oncogenesis of the prostate.

Acknowledgments

We thank Claudia Schmidt for expert technical assistance and Karl-Rudolf Mühlbauer for sequencing of the PCR products.

Footnotes

Address reprint requests to H.-J. Gröne, M.D., Dept. of Cellular and Molecular Pathology, German Cancer Research Center Heidelberg, Im Neuenheimer Feld 280, D-69120, Heidelberg, Germany. E-mail: .ed.zfkd@eneorg.j-.h

Supported by a grant from the Deutsche Forschungsgemeinschaft (SFB 405, B10 to H.-J. G.).

T. E. and M. H. both contributed equally to the study.

References

1. Foster CS, Cornford P, Forsyth L, Djamgoz MBA, Ke Y: The cellular and molecular basis of prostate cancer. Br J Urol 1999, 83:171-194 [PubMed]
2. Catalona WJ, Ritchie JP, Ahmann FR: Comparison of digital rectal examination and serum prostate specific antigen in the early detection of prostate cancer: results of a multicenter clinical trial of 6,630 men. J Urol 1994, 151:1283-1290 [PubMed]
3. Gleason DF: Tannenbaum M. eds. The prostate. Histological grading and clinical staging of prostatic carcinoma. Urologic Pathology. 1977:pp 171-197 Lea & Febinger, Philadelphia
4. Stamney TA, McNeal JE, Yemoto CM: Biological determinants of cancer progression in men with prostate cancer. J Am Med Assoc 1999, 281:1395-1400 [PubMed]
5. Keetch DW, Humphrey P, Stahl D, Smith DS, Catalona WJ: Morphometric analysis and clinical follow up of isolated prostatic intraepithelial neoplasia in needle biopsy of the prostate. J Urol 1995, 154:347-351 [PubMed]
6. Shah R, Mucci NR, Amin A, Macoska JA, Rubin MA: Postatrophic hyperplasia of the prostate gland. Am J Pathol 2001, 158:1767-1773 [PMC free article] [PubMed]
7. De Marzo AM, Marchi VL, Epstein JI, Nelson WG: Proliferative inflammatory atrophy of the prostate. Implifications for prostatic carcinogenesis. Am J Pathol 1999, 155:1985-1992 [PMC free article] [PubMed]
8. Bull JH, Ellison G, Patel A, Muir G, Walker M, Underwood M, Khan F, Paskins L: Identification of potential diagnostic markers of prostate cancer and prostatic intraepithelial neoplasia using cDNA microarray. Br J Cancer 2001, 84:1512-1519 [PMC free article] [PubMed]
9. Dhanasekaran SM, Barrette RR, Ghosh D, Shah R, Varambally S, Kurachi K, Pienta KJ, Rubin MA, Chinnaiyan AM: Delineation of prognostic biomarkers in prostate cancer. Nature 2001, 412:822-826 [PubMed]
10. Luo J, Duggan DJ, Chen Y, Sauvageot J, Ewing CM, Bittner ML, Trent JM, Isaacs WB: Human prostate cancer and benign prostatic hyperplasia: molecular dissection by gene expression profiling. Cancer Res 2001, 61:4683-4688 [PubMed]
11. Welsh JB, Sapinoso LM, Su Al, Kern SG, Wang-Rodriguez J, Moskaluk CA, Frierson HF, Hampton GM: Analysis of gene expression identifies candidate markers and pharmacological targets in prostate cancer. Cancer Res 2001, 61:5974-5978 [PubMed]
12. Luzzi V, Holtschlag V, Watson MA: Expression profiling of ductal carcinoma in situ by laser capture microdissection and high-density oligonucleotide arrays. Am J Pathol 2001, 158:2005-2010 [PMC free article] [PubMed]
13. Chomczynski P, Sacchi N: Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 1987, 162:156-159 [PubMed]
14. Baugh LR, Hill AA, Brown EL, Hunter CP: Quantitative analysis of mRNA amplification by in vitro transcription. Nucleic Acids Res 2001, 29:5-29 [PMC free article] [PubMed]
15. Fink L, Seeger W, Ermert L, Hänze J, Stahl U, Grimminger F, Kummer W, Bohle RM: Real-time quantitative RT-PCR after laser-assisted cell picking. Nat Med 1998, 4:1329-1333 [PubMed]
16. Specht K, Richter R, Müller U, Walch A, Werner M, Höfler H: Quantitative gene expression analysis in microdissection archival formalin-fixed and paraffin-embedded tumor tissue. Am J Pathol 2001, 158:419-429 [PMC free article] [PubMed]
17. Eisen MB, Spellmann PT, Brown PO, Botstein D: Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA 1998, 95:14863-14868 [PMC free article] [PubMed]
18. Segerer S, Nelson PJ, Schlöndorff D: Chemokines, chemokine receptors and renal disease: from basic science to pathophysiologic and therapeutic studies. J Am Soc Nephrol 2000, 11:152-176 [PubMed]
19. Meucci O, Fatatis A, Simen AA, Miller RJ: Expression of CX3Cr1 chemokine receptors on neurons and their role in neuronal survival. Proc Natl Acad Sci USA 2000, 97:8075-8080 [PMC free article] [PubMed]
20. Magee JA, Araki T, Patil S, Ehrig T, True L, Humphrey PA, Catalona WJ, Watson MA, Milbrandt J: Expression profiling reveals hepsin overexpression in prostate cancer. Cancer Res 2001, 61:5692-5696 [PubMed]
21. Maxwell SA, Davis GE: Differential gene expression in p53-mediated apoptosis-resistant vs. apoptosis-sensitive tumor cell lines. Proc Natl Acad Sci USA 2000, 97:13009-13014 [PMC free article] [PubMed]
22. Heston WD: Characterization and glutamyl preferring carboxypeptidase function of prostate specific membrane antigen: a novel folate hydrolase. Urology 1997, 49:104-112 [PubMed]
23. Alers JC, Krijtenburg P-J, Vis AN, Hoedemaeker RF, Wildhagen MF, Hop WCJ, van der Kwast TH, Schröder FH, Tanke HJ, van Dekken H: Molecular cytogenetic analysis of prostatic adenocarcinomas from screening studies. Am J Pathol 2001, 158:399-406 [PMC free article] [PubMed]
24. El Gedaily A, Bubendorf L, Willi N, Fu W, Richter J, Moch H, Mihatsch MJ, Sauter G, Gasser TC: Discovery of new DNA amplification loci in prostate cancer by comparative genomic hybridization. Prostate 2001, 46:184-190 [PubMed]
25. Visakorpi T, Kallioniemi AH, Syvanen AC, Hyytinen ER, Karhu R, Tammela T, Isola JJ, Kallioniemi OP: Genetic changes in primary and recurrent prostate cancer by comparative genomic hybridization. Cancer Res 1995, 55:342-347 [PubMed]
26. Fu W, Bubendorf L, Willi N, Moch H, Mihatsch MJ, Sauter G, Gasser TC: Genetic changes in clinically organ-confined prostate cancer by comparative genomic hybridization. Urology 2000, 56:880-885 [PubMed]
27. Cairns BR: Emerging roles for chromatin remodeling in cancer biology. Trends Cell Biology 2001, 11:15-21 [PubMed]
28. Ferdinandusse S, Denis S, Clayton PT, Graham A, Rees JE, Allen JT, McLean BN, Brown AY, Vreken P, Waterham HR, Wanders RJ: Mutations in the gene encoding peroxisomal alpha-methylacyl-CoA racemase cause adult-onset sensory motor neuropathy. Nat Genet 2000, 24:188-191 [PubMed]
29. Kuhajda FP, Jenner K, Wood FD, Hennigar RA, Jacobs LB, Dick JD, Pasternack GR: Fatty acid synthesis: a potential selective target for antineoplastic therapy. Proc Natl Acad Sci USA 1994, 91:6379-6383 [PMC free article] [PubMed]
30. Hughes-Fulford M, Chen Y, Tjandrawinata RR: Fatty acid regulates gene expression and growth of human prostate cancer PC-3 cells. Carcinogenesis 2001, 22:701-707 [PubMed]
31. Andersson O, Nordlund-Moller L, Barnes HJ: Heterologous expression of human uteroglobin/polychlorinated biphenyl-binding protein. Determination of ligand binding parameters and mechanisms of phospholipase A2 inhibition in vitro. J Biol Chem 1994, 269:19081-19087 [PubMed]
32. Peri A, Bonaccorsi L, Muratori M, Luconi M, Baldi E, Granchi S, Pesciullesi A, Mimi E, Gioppi F, Forti G, Serio M, Miele L, Maggi M: Uteroglobulin reverts the transformed phenotype in the endometrial adenocarcinoma cell line HEC-1A by disrupting the metabolic pathways generating platelet-activating factor. Int J Cancer 2000, 88:525-534 [PubMed]
33. Zha S, Gage WR, Sauvageot J, Saria EA, Putzi MJ, Ewing CM, De Marzo AM, Isaacs WB: Cyclooxygenase-2 is upregulated in proliferative inflammatory atrophy of the prostate, but not in prostate carcinoma. Cancer Res 2001, 61:8617-8623 [PubMed]

Articles from The American Journal of Pathology are provided here courtesy of American Society for Investigative Pathology
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

  • Cited in Books
    Cited in Books
    PubMed Central articles cited in books
  • Gene (nucleotide)
    Gene (nucleotide)
    Records in Gene identified from shared sequence links
  • MedGen
    MedGen
    Related information in MedGen
  • Nucleotide
    Nucleotide
    Published Nucleotide sequences
  • PubMed
    PubMed
    PubMed citations for these articles
  • Substance
    Substance
    PubChem Substance links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...