SEARCH

SEARCH BY CITATION

Keywords:

  • ERG rearrangements;
  • PTEN;
  • fluorescence in-situ hybridization;
  • high-grade prostatic intra-epithelial neoplasia;
  • prostate cancer;
  • Gleason score;
  • disease progression

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

What’s known on the subject? and What does the study add?

So far we know that ERG rearrangements and PTEN deletions interact to induce prostate cancer in transgenic mice. The study confirms that an association also exists between the two genetic aberrations in human prostate cancer, as there is increased incidence of PTEN deletions in cases with ERG rearrangements.

OBJECTIVE

To investigate the interaction between, and significance of, ERG gene rearrangements and PTEN genomic deletions in relation to the development and progression of prostate cancer (PCA).

PATIENTS AND METHODS

We interrogated an initial cohort of 220 men with localized PCA using fluorescence in situ hybridization for ERG rearrangements and PTEN genomic deletions.

RESULTS

The incidences of ERG rearrangements and PTEN deletions in PCA were significantly higher than in high-grade prostatic intra-epithelial neoplasia (HGPIN) and benign prostate tissue (P < 0.001). ERG rearrangements and PTEN deletions were detected in 41.9 and 42.6% of patients’ tumours, respectively. ERG rearrangements were never detected in benign prostate tissue, while PTEN aberrations were present at a basal level of 4.6%. PTEN hemizygous deletions showed higher frequency than homozygous deletions within each diagnostic category from benign prostate tissue to HGPIN and PCA (P ≤ 0.001). Furthermore, in 29 patients where all three tissues were available, PTEN genomic aberrations in PCA were significantly different from those in benign tissue (P = 0.005) and HGPIN (P = 0.02), reflecting the accumulation of genomic aberrations in the early stages of disease progression. Within this cohort, 71.4% of homozygous and 44.2% of hemizygous PTEN deletions occurred simultaneously with ERG rearrangements (P ≈ 0). Stratified according to Gleason score (GS), hemizygous PTEN deletions across various GS groups were observed at a higher frequency than homozygous deletions. However, PTEN homozygous deletions showed positive trends with higher GS, increasing in poorly differentiated PCA (GS 8–10) in comparison to moderately and well differentiated tumours (GS 6 and 7).

CONCLUSION

We show significant association between ERG gene rearrangements and PTEN genomic aberrations in subset of PCA. Our analysis also provides further support for the observation that homozygous PTEN deletions can occur within the subset of HGPIN lesions, and shows accumulating genetic aberrations with disease progression, evidenced by higher detection in PCA than in HGPIN and more PTEN homozygous deletions in GS 8–10 than in 6–7.


Abbreviations
BAC

bacterial artificial chromosome

dUTP

deoxyuridine triphosphate

FISH

fluorescence in situ hybridization

GS

Gleason score

HGPIN

high-grade prostatic intra-epithelial neoplasia

PCA

prostate cancer

TMA

tissue microarray

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

It is well established that prostate cancer (PCA) is one of the most heterogeneous cancers, which has made it difficult to identify and validate significant molecular changes associated with disease progression. However, common molecular changes with established prognostic value include loss of 8p, 6q, 10q, 13q, 16q and 18q, and gain of 8q [1,2].

The recent discovery of the common ERG gene rearrangements in PCA [3] suggests that a molecular classification based on gene fusion status might be possible. Recent expression-profiling studies characterizing genomic events in ERG-rearranged tumours support this hypothesis [4–6].

PTEN, a phosphoinositide 3-phosphatase which negatively regulates the PI3K signalling pathway, is located on chromosome 10 and acts as tumour suppressor gene [7]. Reduction in PTEN function has been associated with disease progression and the androgen-independent state [8–15]. Using in vitro research, it has been shown by several groups that the ERG gene rearrangements by themselves may not be capable of transforming normal cells into cancers and that this requires interaction with other genetic changes [16–18]. Using transgenic mouse models, it is becoming more obvious that PCA initiation requires several genetic hits and might not be fully explained by the presence of TMPRSS2–ERG gene fusion alone [17,18]. Two recent studies showed that PTEN genomic deletions and TMRPSS2–ERG gene fusion (the most common gene fusion resulting from ERG gene rearrangement) interact in driving PCA disease development and progression [16,19,20]. In the study by Carver et al. [16], transgenic mice showing aberrant ERG in conjunction with PTEN heterozygosity showed invasive tumours compared to high-grade prostatic intra-epithelial neoplasia (HGPIN) -like lesions observed in PTEN+/- mice without ERG aberrations. Similar results were obtained by King et al. [20]. Recent reports from our group have identified associations between poor outcome in localized and androgen-independent tumours and PTEN genomic deletion [21,22]. In the study by Yoshimoto et al. [21], localized PCA tumours harbouring TMPRSS2–ERG gene fusion and PTEN genomic deletions in combination were associated with a worse prognosis than tumours without any of those genetic alterations, which had the best prognosis in terms of PSA relapse [23].

However, there is still controversy as to the significance and sequence of events in the human prostate. Whereas all studies agree that ERG gene rearrangements occur early at the HGPIN stage [24–26], only two previous studies reported PTEN losses in HGPIN [27,28], while recent in vivo studies have suggested that the development and progression into invasive cancer would start initially by PTEN deregulation causing higher levels of instability and activation of the AKT pathway. Cells with PTEN haplo-insufficiency have been found to be more unstable [29] and prone to undergo additional genomic alterations, such as ERG gene rearrangements and concomitant aberrant expression of ERG, to fully establish disease development and progression. However, others consider that PTEN genomic deletions probably occur later, at the metastatic stage [16,19,28].

Fluorescence in situ hybridization (FISH) has substantial advantages over other methods for determining clonal presence of genomic aberrations at the cellular level, enabling direct correlation with histologies such as Gleason score (GS) and HGPIN. In the present study, we used FISH to interrogate a large cohort of patients with localized PCA to determine the prevalence of ERG rearrangements and to annotate several PTEN genomic aberrations distinguishing among PTEN gain and monzygous, hemi- and homozygous PTEN genomic deletions at early, intermediate and later steps in the carcinogenic process.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Study population

The patients’ cohort consisted of 220 men treated by radical prostatectomy as a monotherapy for localized disease at the Jewish General Hospital, Montreal, Canada. Complete clinical and pathological variables were available for 196 men as previously published [25,30]. All data were collected with Institutional Review Board approval. To investigate the relationship with disease progression, prostate samples included benign, HGPIN and PCA of each patient when present within the prostate gland. The prostate samples were assembled on three tissue microarrays (TMAs). A mean (range) of 3.3 (1–9) TMA cores, 0.6 mm in diameter, were sampled, including benign, HGPIN and PCA. A total of 1252 cores were available for analysis with 189 benign, 115 HGPIN and 984 cancer cores.

Pathological analysis

All TMA cores were reviewed and assigned a diagnosis (benign, HGPIN or PCA) by one of the study pathologists (T.A.B.) without prior knowledge of any clinical or pathological variables.

Assessment of ERG gene rearrangements and PTEN gene status by FISH

We previously described the dual-colour interphase break-apart FISH assay to assess indirectly the ERG gene rearrangements [25,26]. Two differentially labelled probes were designed to span the telomeric and centromeric neighbouring regions of the ERG locus. As previously described, this break-apart probe system allows differentiation of ERG rearrangement through insertion, ERG rearrangement through an intronic deletion of DNA, and no gene rearrangement [31]. The samples were analysed under a ×60 oil- immersion objective using an Olympus BX-51 fluorescence microscope equipped with appropriate filters, a charge-coupled device camera (Olympus, Center Valley, PA, USA), and the CytoVision FISH imaging and capturing software (Applied Imaging, San Jose, CA, USA). Evaluations of the tests were performed by a pathologist with expertise in analysing interphase FISH experiments. For each sample, we attempted to score at least 100 nuclei. Examples of ERG gene rearrangements by translocation and deletion are shown in Fig. 1.

image

Figure 1. FISH images of ERG rearrangement assay. (A) FISH image of representative PCA nucleus with one yellow and one red signal, demonstrating ERG rearrangement through deletion. (B) FISH image of a representative PCA nucleus displaying ERG rearrangement through insertion. One yellow, one green, and one red signal are present, showing ERG rearrangement through insertion. (C) FISH image of a representative PCA nucleus displaying absence of ERG rearrangement. Two yellow signals per nucleus are present.

Download figure to PowerPoint

Because PTEN genomic deletions are interstitial, and usually restricted to a small region of chromosome 10 [23], a four-colour interphase FISH strategy was used to investigate PTEN gene status, using four bacterial artificial chromosome (BAC) clones spanning both flanking PTEN genomic regions (GDF2 and FAS loci), the PTEN gene locus and a commercially available DNA probe for the centromeric region 10p11.1–q11.1 (SpectrumAqua labelled CEP 10; Vysis Inc., Downers Grove, IL, USA). The following BAC clones were used: (a) BMPR1A gene locus located at 10q23.2: RP11–141D8 (88.28–88.46 Mb), RP11–52G13 (88.44–88.56 Mb) and RP11–420K10 (88.50–88.67 Mb); (b) PTEN gene locus located at 10q23.31: RP11–846G17 (89.66–89.84 Mb); and (c) FAS gene locus located at 10q23.31: RP11–399O19 and RP11–360H20 (90.6–90.8 Mb). The position and name of the BAC clones were taken from the Human March 2006 assembly of the University of California Santa Cruz (UCSC) Genome Browser 1. The presence of PTEN, FAS and BMPR1A sequences and the correct chromosome location of the BAC clones were verified by PCR and by hybridization to metaphase spreads from normal peripheral lymphocytes, respectively. BAC DNA was extracted and labelled with either SpectrumGreen-dUTP (deoxyuridine triphosphate), SpectrumOrange-dUTP or SpectrumRed-dUTP (Abbott Molecular, Inter Medico, Markham, ON, Canada) using the nick-translation kit (Abbott Molecular). Copy number was evaluated for each probe by counting spots in 100 non-overlapped, intact interphase nuclei per tumour tissue core. The establishment of PTEN gene copy number status was defined by considering the adjacent probes (GDF2 and FAS loci) used for the truncation artifacts, aneusomy, nuclear size and chromatin condensation. Based on hybridization in control cores (data not shown), hemizygous deletion of PTEN was defined as >43% (mean + 3 sd in non-neoplastic controls) of tumour nuclei containing one PTEN locus signal and by the presence of CEP 10 signals. Homozygous deletion of PTEN was exhibited by the simultaneous lack of both PTEN locus signals and the presence of control signals in >30% of cells. Where there was only one TMA core available for assessment, we used more conservative 50% threshold levels to rigorously ensure that PTEN genomic losses were correctly evaluated. Both nuclear diameter and shape are important influences on nuclear truncation when setting thresholds for identifying hemi- and homozygous deletions of PTEN by FISH (Squire et al., in preparation). An example of PTEN genomic deletion evaluation is shown in Fig. 2.

image

Figure 2. Representative FISH images are shown for PCA TMA applying the four-colour FISH. The panel shows a pseudo-colour image with the 4′,6-diamidino-2-phenylindole (DAPI) counterstained nuclei. The rectangles show the main FISH image high magnification. (A) Four-colour FISH identifying two signals of red (PTEN locus), pink (BMPR1A locus), green (FAS locus) bacterial artificial chromosome (BAC) probes, as well as paired pale blue (CEP10, Vysis Inc.) signals in most of the nuclei, indicating no deletion of PTEN in tumour cells. (B) FISH analysis for both flanking PTEN genomic regions (BMPR1A and FAS loci); the PTEN gene locus showed low-level copy gain accompanied by blue signals (CEP10) most probably derived from the chromosome 10 aneuploidy (arrows). (C) Representative four-colour PTEN FISH image showing tumour cells with single red signal for PTEN and FAS loci in most of the nuclei and paired pink (BMPR1A locus) and blue signals (CEP10), indicating hemizygous deletion of the PTEN gene region in PCA (arrow). (D) Representative FISH image of homozygous PTEN deletion in PCA showing absence of red signal for PTEN locus in most of the nuclei. The retained single green signal (FAS locus) and pink signal (BMPR1A locus) in most of the nuclei and paired blue signals for CEP10 indicates hemizygous deletion of the BMPR1A and FAS loci, and homozygous deletion of the PTEN locus in PCA (arrow).

Download figure to PowerPoint

Statistical analysis

Chi-squared and Kruskal–Wallis non-parametric tests were used to test for associations between ERG gene rearrangements and PTEN loss and other key variables such as GS. All analyses were performed using S-PLUS (MathSoft, Inc., Seattle, WA, USA). A two-tailed P≤ 0.05 was considered to indicate statistical significance.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Study population

In this cohort, the mean (range) patient age was 64 (42–80) years with mean pre-surgical PSA level of 9 (range 0.5–51.9) ng/mL. Interrogation of ERG gene rearrangements and PTEN genomic aberrations were available for 814 and 874 cores belonging to 167 and 183 patients, respectively. The incidences of ERG gene rearrangements and PTEN genetic aberrations in this cohort were 41.92% (70/167) and 42.62% (78/183) of patients, respectively. Complete patients demographics are shown in Table 1.

Table 1.  Summary of patients’ characteristics
VariableN (%)
  1. GS, Gleason score; pT, pathological stage; SM, surgical margin.

GS 
 3 1 (0.44)
 5 3 (1.33)
 665 (28.76)
 7130 (57.52)
 814 (6.19)
 913 (5.75)
Pathological stage, pT 
 T2142 (65.44)
 T374 (34.1)
 T4 1 (0.46)
Surgical Margin 
 Negative120 (55.3)
 Positive97 (44.7)
-ERG 
 Rearrangement71 (42.5)
 No rearrangements96 (57.5)
 Rearrangement through insertion30 (17.96)
 Rearrangement through deletion41 (24.55)
PTEN 
 No deletion105 (57.38)
 Monozygous 2 (1.09)
 Hemizygous43 (23.5)
 Homozygous25 (13.66)
 Gain 8 (4.37)

ERG gene rearrangements and PTEN genomic aberrations in benign tissue, HGPIN and PCA

Both ERG gene rearrangements and PTEN genomic loss were significantly associated with invasive cancer compared to HGPIN and benign prostatic tissue and both showed positive associations between the different categories (benign, HGPIN and PCA) increasing with disease progression (P ≈ 0). Interestingly, while, ERG gene rearrangements were not detected in benign prostatic tissue, there were PTEN genomic deletions in benign prostatic tissue derived from patients with PCA but at significantly lower frequency than HGPIN and PCA (2.8 vs 9.3 and 21.1%, respectively, for hemizygous deletions; and 0.9 vs 5.3 and 7.6% for homozygous deletions).

To account for tumour heterogeneity, we investigated those two genetic aberrations stratified based on the type of genetic aberration within TMA core. ERG gene rearrangements by deletion were more frequent in each diagnostic category than those occurring by insertion (57.7 vs 42.2% in ERG-rearranged PCA and 3.6 vs 12.5% in rearranged HGPIN). On the other hand, hemizygous PTEN deletions were present at higher frequencies than homozygous deletions within each diagnostic category. However, as background levels of hemizygous losses are more likely to be higher, due to artefacts of nuclear truncation, than those of homozygous losses, these differences could be predominantly technical in nature. Complete results of ERG gene rearrangements and PTEN genomic aberrations are shown in Table 2. We also investigated the association between ERG gene rearrangements and PTEN genomic loss in PCA. There was a significant association between the two genetic aberrations, with PTEN homozygous deletions showing the highest concurrent association with ERG rearrangements, occurring in 71.4% of tumours, and hemizygous deletions occurring in association with ERG rearrangements in 44.2% of the cases (P ≈ 0). These prevalence data are consistent with a model in which hemizygous PTEN loss occurs early in the carcinogenic process; ERG rearrangements may then follow, and subsequent acquisition of homozygous PTEN losses are associated with more advanced disease. Table 3 shows the complete results in relation to ERG rearrangement type and PTEN genomic deletions.

Table 2.  ERG rearrangement and PTEN genomic aberrations status by TMA diagnosis
DiagnosisERG rearrangement*, n (%)
AbsentInsertionDeletion
Benign91 (100)0 (0)0 (0)
HGPIN49 (87.5)5 (8.9)2 (3.6)
Cancer330 (60.8)91 (16.8)122 (22.4)
DiagnosisPTEN**, n (%)
No deletionMonosomyHemizygousHomozygousGain
  • *

    ,**P ≈ 0 (chi-square).

  • †Not all cases were informative. HGPIN, high-grade prostatic intra-epithelial neoplasia.

Benign102 (95.3)0 (0)3 (2.8)1 (0.9)1 (0.9)
HGPIN 64 (85.3)0 (0)7 (9.3)4 (5.3)0 (0)
Cancer459 (68.8)8 (1.2)141 (21.1)51 (7.6)8 (1.2)
Table 3.  Combined PTEN genomic aberrations and ERG rearrangements in PCA foci
ERG rearrangement*PTEN, n (%)
Not deletedMonosomyHemizygousHomozygousGain
  • *

    P ≈ 0 (by chi-square) for the association between ERG rearrangement and PTEN.

Absent303 (80.6)7 (1.9)53 (14.0)10 (2.7)3 (0.8)
Insertion46 (60.5)0 (0)18 (23.7) 11 (14.5)1 (1.3)
Deletion61 (59.8)1 (1.0)24 (23.5)14 (13.7)2 (2.0)

We also investigated the status of ERG rearrangements and PTEN genomic aberrations stratified based on GS of individual TMA cores; only ERG gene rearrangements were significantly associated with GS, with 81.5% in both GS 6 and 7, and 13.9% in GS 8–10 (P = 0.017). Hemizygous PTEN genomic deletions across various GS groups were detected more frequently than homozygous deletions. However, poorly differentiated PCA (GS 8–10) were more likely to show homozygous deletions of PTEN than moderately and well differentiated tumours (GS 6 and 7), in keeping with a model in which complete loss of PTEN is associated with tumour progression. Similar trends were not consistent in hemizygous PTEN deletions. Results of ERG rearrangement and PTEN aberrations relative to GS are shown in Table 4.

Table 4.  ERG rearrangement and PTEN genomic aberrations in TMA GS
GSERG rearrangement*, n (%)
AbsentInsertionDeletion
6248 (58.6)78 (18.4)97 (22.9)
749 (59.8) 11 (13.4)22 (26.8)
8–1031 (86.1)2 (5.6)3 (8.3)
GSPTEN**n (%)
Not deletedMonosomyHemizygousHomozygousGain
  • *

    P = 0.017 (by chi-square);

  • **

    P = 0.56 (by chi-square).

6362 (69.3)5 (1.0)108 (20.7)39 (7.5)8 (1.5)
763 (64.3)3 (3.1)24 (24.5)8 (8.2)0 (0)
8–1029 (69.0)0 (0)9 (21.4)4 (9.5)0 (0)

Notably, where ERG gene rearrangements were homogenous within each cancer focus, PTEN status showed heterogeneity in 46% of the cases associated with ERG gene rearrangement (data not shown).

Combined status of ERG gene rearrangements and PTEN in PCA progression

As there was a positive association between the ERG gene rearrangements and PTEN genomic deletions, we investigated the combined status in cases where benign tissue, HGPIN and PCA were sampled and the FISH test was informative.

There were 29 cases of PTEN that had all three tissue types present and the FISH test was informative. We enumerated PTEN copy number aberrations by interphase FISH in relation to disease progression, giving a score for each genetic aberration (0, no PTEN deletions detected; 1 monosomy for chromsome 10; 2, hemizygous PTEN deletion; 3, homzygous PTEN deletion). The mean basal PTEN score by FISH in benign prostate tissue was 0.31; in HGPIN the score was slightly higher, at 0.38, and in PCA it was 0.83. The difference between PCA and benign prostate tissue was significant (P = 0.005 paired t-test), as was the difference between PCA and HGPIN (P = 0.02), but there was no difference in PTEN score between benign and HGPIN (P > 0.6). However, as noted earlier, measuring subtle differences of interphase FISH enumerations or class can be complicated by nuclear truncation effects. It is notable that all HGPIN lesions with evident homozygous PTEN genomic deletions, both the HGPIN and the associated PCA (where the test was informative), showed ERG gene rearrangements (three of three cases). These results support accumulating PTEN genomic deletions with disease progression from benign to HGPIN and PCA.

For ERG gene rearrangements, there were 19 cases that had all three tissue types present and the test was informative. Five cases had ERG rearrangements present in the tumour. None had ERG rearrangements present in the benign samples (P = 0.021 by paired t-test in comparison with tumour). Three cases showed ERG rearrangements in HGPIN tissues, and this includes one case that is not ERG-rearranged in the tumour (which is not adjacent to the sampled HGPIN) (P > 0.3 by paired t-test in comparison with tumour). Three cases had ERG rearrangements by deletion present in the tumour and two cases in HGPIN tissues, including one case that is negative for ERG rearrangements in the tumour (P > 0.5 by paired t-test in comparison with tumour). These results suggest that ERG rearrangement fusion is present more often in tumour and HGPIN than in benign tissues, but that there may not be much difference between tumour and HGPIN for ERG rearrangements within individual cases (i.e. they are of similar status).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

To date, the ERG gene rearrangement has been investigated in several studies, all of which confirmed its specification to HGPIN and invasive cancer, with the TMPRSS2–ERG gene fusion representing the most common gene fusion resulting from ERG rearrangement in localized PCA [32]. PTEN genomic deletions have also been widely studied as a significant contributor to PCA progression and several studies have confirmed its association with a worse clinical outcome [21,33]. In vitro studies showed that ERG gene rearrangements by themselves are not capable of transforming normal epithelial cell lines (RWPE) but increases its invasion capabilities [17], whereas PTEN genomic deletions caused increased transformation and proliferation [34]. Using in vivo mouse models, it is evident that PCA development requires the interaction between PTEN genomic deletions and ERG gene rearrangements, with PTEN genomic deletions causing AKT activation and HGPIN-like lesions followed by ERG gene rearrangements to allow for the establishment of PCA [16,19]. However, the association and sequence incidence between those two genetic aberrations in human PCA development and progression are not yet fully characterized.

In the present study, we systematically analysed the status of ERG gene rearrangements and PTEN deletions across various stages of disease progression, starting from benign prostatic tissue and moving on to HGPIN and invasive PCA. We confirmed the specificity and significant association of ERG gene rearrangements with PCA, as well as its significant association with PTEN genomic deletions.

We have also confirmed that the two distinct genomic aberrations are detected in a subset of HGPIN, as was recently shown by Han et al. [28]. However, in contrast to these results, we have detected homozygous deletions in ≈5% of HGPIN lesions, compared with 0% in the study by Han et al. [28]. Furthermore, benign prostatic tissue showed evidence of PTEN genomic aberrations as well, albeit at much lower rates than HGPIN and PCA. Although we cannot fully exclude the possibility that some deletions in benign prostatic tissue could represent background nuclear truncations from slide sectioning or early neoplastic changes at molecular levels that are not obvious morphologically at the microscopic level, the observation of PTEN homozygous deletions within a subset of HGPIN (5.3%) coupled with the increased incidence of this aberration in localized cancers with higher GS would argue that PTEN deletions are occurring at the HGPIN stage in co-operation with ERG gene rearrangements and are not always a later event in PCA progression as hypothesized by Han et al. [28]. Although it is possible that some of the HGPIN lesions sampled here could represent intra-ductal spread of cancer, and from these observations we cannot firmly ascertain which of those two genetic aberrations is the initial event in PCA development, the present results are in keeping with the recent in vivo studies of Carver et al. [16] and King et al. [20], which showed that double transgenic mice for PTEN and ERG aberrations develop PCA possibly through AKT activation [16,20].

We hypothesize that a subset of PCA may be driven initially by PTEN genomic hemizygous loss, causing HGPIN lesions. Thereafter, PTEN haplo-insufficiency leads to genomic instability [29,35], which may facilitate the chromosomal rearrangement leading to gene fusion formation and progression to cancer. Subsequent PTEN deregulation by homozygous deletion (or other inactivating mechanisms) could then occur to induce PCA progression (Fig. 3). This hypothetical sequence of events is supported here by the observation that hemizygous deletions were more prevalent in HGPIN and localized PCA than homozygous deletions, with the latter showing significant and increased association with disease progression as evidenced by higher incidence of homozygous deletions in poorly differentiated PCA (GS 8–10) than in well and moderately differentiated tumours (GS 6–7). In the present study, there was significant and higher association between ERG gene rearrangements and homozygous PTEN deletions than with hemizygous deletions in PCA (71.4 vs 44%), suggesting that the interaction between the two genetic aberrations is stronger when there is complete lack of PTEN activity, and this association would signify the highest risk for PCA development and progression. Whether HGPIN lesions with homozygous PTEN deletions are associated with accelerated rate of disease progression is beyond the scope of the present study, and needs to be confirmed in larger cohorts to identify HGPIN lesions at highest risk for disease progression.

image

Figure 3. Schematic model of hypothetical sequence of genomic events in PCA progression. The acquisition of a PTEN haplo-insufficiency in benign prostatic precursors may represent an early event in a subset of PCAs. Reduced PTEN protein levels may facilitate genomic instability, leading to acquisition of ETS (erythroblastosis virus E26 transformation-specific) rearrangements. Synergistic co-operation between PTEN and ETS is associated with early steps of PCA formation. Continuing instability generates genotypic heterogeneity and diversity such that subclones bearing ‘null PTEN’ (PTEN homozygous deletions) have increased selective advantage for tumour progression through activation of the AKT pathway.

Download figure to PowerPoint

It is becoming evident that ERG gene rearrangements signify a distinct molecular subtype of PCA that may have potential clinical implications. However, the prognostic value of these markers are still unclear. While several studies have shown an association with an unfavourable prognosis and aggressive clinical behaviour (such as PSA relapse or cancer-specific mortality) [36–38], others have failed to show such association, or have even pointed towards a better prognosis [25,39,40]. Similarly, only a few reports (including the present study) were able to confirm a significant association with GS [32,41]. These discrepancies could be a reflection of the different methods used in assessing the ERG gene rearrangement status (FISH vs reverse transcription PCR), including detecting variable versions of TMPRSS2–ERG gene fusion. Other potential causes of discrepancy include variable cohort sizes, background and population studied (radical prostatectomy vs watchful waiting) as well as variation in clinical outcome endpoints measured (after radical PSA relapse vs cancer-specific mortality) [25,36,38,40,42–44]. In the present study, there was a significant association between ERG rearrangements and GS, with tumours of GS 6 and 7 more likely to be ERG-rearranged than those of GS > 7. This latter finding would allow us to subtype and differentiate among two tumours with similar GS, should ERG rearrangements be confirmed as a prognostic or therapeutic marker, specifically those with GS ≤ 7, which includes most patients seeking medical advice for PCA diagnosis. A recent report by Attard et al. [45] suggested a significant association between patients with castration-resistant PCA with ERG rearrangements and abiraterone acetate. The prognostic value of PTEN genomic deletions and haplo-insufficieny of the protein in PCA is well documented for patients with homozygous PTEN deletions and this is true for other cancers as well. We and others have confirmed a significant association of homozygous PTEN genomic deletions and ERG gene rearrangements with lymph node metastasis, hormone-refractory PCA and cancer-specific death [21,33,37,46,47].

Recently, our group has shown that tumours with TMPRSS2–ERG gene fusion and PTEN deletions are associated with even more aggressive clinical behaviour than those with any of the genetic aberrations alone [23]. Although the significance of this later finding still needs validation in larger cohorts, the data presented in the present study, showing a strong and significant association between ERG gene rearrangement and PTEN genomic deletions, confirm that those tumours represent distinct molecular subsets of PCA that require more detailed definition and characterization.

These data also provide further evidence signifying the importance of targeting the two genetic aberrations and their pathways as potential markers for disease progression and targets for PCA therapy in the hope of altering or slowing the development of hormone refractory PCA and preventing cancer-specific mortality.

In conclusion, the present study highlights the co-operation and importance of ERG gene rearrangements and PTEN deletions as significant drivers of PCA development and progression. It also confirms that PTEN genomic deletions occur in a subset of HGPIN lesions and show subsequent accumulating genetic aberrations with PCA progression, as evidenced by a greater incidence of homozygous deletions in higher-grade tumours.

ACKNOWLEDGEMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

The authors would like to thank Cheryl Noonan and Maggie Stephenson for technical assistance in preparation of the manuscript.

CONFLICT OF INTEREST

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Jeremy A. Squire is a consultant for CymoGen Dx. Source of Funding: this work was supported in part by Prostate Cancer Canada, The Young Investigator Award of the Prostate Cancer Foundation, USA, The Canadian Institutes of Health Research (CIHR) and the Fonds de la Recherche en Santé du Quebec (T.A.B.), and the Canadian Cancer Society (J.A.S.).

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES