european urology 49 (2006) 169–175 available at www.sciencedirect.com journal homepage: www.europeanurology.com From Lab to Clinic Relationship of NKX3.1 and MYC Gene Copy Number Ratio and DNA Hypomethylation to Prostate Carcinoma Stage Roland Kindich a, Andrea R. Florl b, Jörn Kamradt a, Jan Lehmann a, Mirko Müller b, Bernd Wullich a, Wolfgang A. Schulz b,c,* a Clinic of Urology and Pediatric Urology, University of the Saarland, Homburg, Germany Department of Urology, Heinrich Heine University, Moorenstr. 5; 40225 Düsseldorf, Germany c Center for Biological and Medical Research, Heinrich Heine University, Düsseldorf, Germany b Article info Abstract Article history: Accepted September 8, 2005 Published online ahead of print on November 15, 2005 Objective: High stage prostate cancers have been reported to frequently harbor chromosome 8 alterations and hypomethylation of LINE-1 retrotransposons. The potential of these parameters for molecular staging of prostate carcinoma was investigated. Methods: High molecular weight DNA was extracted from 63 carcinoma tissues (22 pT2, 38 pT3, 3 pT4). Chromosome 8 alterations were followed by determining the ratio of NKX3.1 (at 8p21) to MYC (at 8q24) gene copy numbers (NKX3.1:MYC ratio) using a new real-time PCR technique. LINE1 hypomethylation was quantified by Southern blot analysis. Results: In 42 carcinomas NKX3.1 copy numbers were altered, with decreases in 32 cases. Copy numbers of MYC were increased in 38 cases and diminished in four. The NKX3.1:MYC ratio was altered in 45 specimens, with a decrease in all but two. NKX3.1 loss was associated with tumor stage ( p < 0.03) and MYC gain with Gleason score ( p < 0.03). The NKX3.1:MYC ratio was highly significantly associated with tumor stage ( p < 0.002), displaying 66% sensitivity and 87% specificity. LINE-1 hypomethylation was related ( p < 0.004) to tumor stage, but exhibited lower sensitivity (59%) and specificity (77%). Conclusion: A straightforward PCR technique detecting chromosome 8 alterations might be useful to predict which prostate cancers are organconfined while determination of hypomethylation appears to be somewhat less well suited. # 2005 Elsevier B.V. All rights reserved. Keywords: Prostate cancer Real-time quantitative PCR Chromosome 8 L1 Retrotransposon DNA Methylation * Corresponding author. Tel. +49 211 81 18966; Fax: +49 211 81 15846. E-mail address: wolfgang.schulz@uni-duesseldorf.de (W.A. Schulz). 0302-2838/$ – see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.eururo.2005.09.012 170 1. european urology 49 (2006) 169–175 Introduction The most frequently altered chromosome in prostate carcinoma may be chromosome 8 [1]. In general, sequences on 8p tend to be lost, whereas sequences on 8q tend to be gained. These changes are however not uniform and in individual carcinomas various regions on 8p and 8q may be affected individually or concomitantly. Among the most consistently gained or lost genes, respectively, are MYC at 8q24.21 and NKX3.1 at 8p21.2. A considerable number of studies employing cytogenetic techniques or loss of heterozygosity analyses have concordantly shown that alterations of chromosome 8 are associated with a more aggressive behavior of prostate cancer, reflected by higher tumor stage, higher grades, or higher rates of metastasis and recurrences found in diverse studies [2–11]. Most studies report an association of such properties with 8q gain, while in others a correlation with 8p loss has been found. In fact, the two changes are in many cases linked mechanistically [12,13]. For instance, formation of an 8q isochromosome or deletion of 8p distally from recurrent breakpoint regions at 8p12 or 8p21 with subsequent duplication of the remaining chromosome segment lead to concurrent gain of 8q and loss of 8p sequences. The individual effects of 8p loss and 8q gain are therefore difficult to discern. For this reason, one might hypothesize that the most robust parameter associated with the overall clinical behavior of prostate carcinoma might be the ratio of 8p to 8q which reflects both changes. Accordingly, we surmised that a simple and rapid measurement of this ratio might be useful for ‘molecular staging’ of this cancer and as a prognostic parameter. As most losses on 8p include NKX3.1 as a target and most gains on 8q include MYC, we have recently developed a relatively simple and straightforward, but highly standardized PCR-based method to determine the relative copy numbers of these two genes as an easily measurable parameter for numerical chromosome 8 alterations [14]. Alterations on chromosome 8 in prostate carcinoma were also found to correlate with diminished methylation of repetitive LINE-1 sequences [15]. LINE-1 sequences are the most frequent retrotransposons in humans constituting approximately 18% of the genome. They are highly methylated in normal somatic tissues including prostate. In cancer cells certain single-copy sequences become hypermethylated, while the methylation of repetitive sequences, especially LINE-1 retrotransposons, is often diminished. In effect, this leads to a decrease in the methylcytosine content of cancer cells termed genome-wide (‘global’) hypomethylation (reviewed in [16]), which can be measured directly or through its parallel effect on LINE-1 methylation. Pronounced global hypomethylation is associated with chromosomal instability in several human cancer types [16]. In prostate carcinoma, hypermethylation of a number of genes including GSTP1 takes place early during the development of most tumors [17,18], while LINE-1 hypomethylation is found only in a subset of cancers, usually those with a higher stage [17]. Particularly pronounced hypomethylation is observed in specimen from metastatic and androgen-independent cases [15]. Therefore, LINE-1 hypomethylation behaves similar to chromosome 8 alterations, begging the question whether the two changes are causally related. We have therefore investigated the relationship of the NKX3.1 and MYC gene copy numbers and their ratio as well as of LINE-1 hypomethylation to prostate carcinoma stage and to Gleason grade. 2. Methods 2.1. Patients and tissues The prostate carcinoma samples used were a subset of a previously described series [17] that was selected by the criteria of availability of greater DNA amounts and a complete standardized patient follow-up. Except for one T4 tumor, specimens were obtained by radical prostatectomy, macroscopically dissected by a specially qualified pathologist with histological control of adjacent sections as described in detail elsewhere [17], rapidly frozen and stored until extraction of DNA and RNA. By the 1997 TNM classification, in the present subseries, 22 carcinomas were staged as pT2, 38 as pT3, and 3 tumors as T4. With respect to Gleason sum, 17 were graded as <7, 27 as 7, and 19 as >7. Lymph node metastases were present in 13 cases. The individual tumors are listed in Table 1. 2.2. DNA extraction and PCR analysis DNA was extracted by a standard procedure using the Blood and Cell Culture DNA Midi Kit (Qiagen, Hilden, Germany) with extended proteinase K digestion. Quantitative real-time PCR for NKX3.1 and MYC was performed as described14 using the LightCyclerTM system (Roche Diagnostics, Mannheim, Germany) with the FastStart DNA Master SYBR Green I LightCycler Kit (Roche Diagnostics). Relative gene copy number was calculated by a modification of the 2 DDCT method from the real-time PCR efficiencies (E) determined for each individual run, and the CP deviations of the target and reference gene in a test sample vs. a control according to the equation: E [(CPtargetCPref)test (CPtarget-CPref)control] . For the present study, the method was recalibrated using a new set of leukocyte DNA samples from 8 healthy individuals. Mean (2  SD) gene copy number values in the normal DNA were 2.10  0.42 for NKX3.1, 1.76  0.27 for MYC, and 1.19  0.21 for the NKX3.1:MYC ratio, 171 european urology 49 (2006) 169–175 Table 1 – Tumor samples studied and overview of the results Tumor No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 Sample No. P8 P10 P14 P26 P28 P36 P42 P43 P44 P47 P49 P51 P52 P65 P74 P80 P85 P87 P91 P97 P99 P107 P109 P117 P119 P123 P133 P135 P137 P139 P141 P143 P145 P147 P153 P155 P161 P163 P165 P167 P169 P171 P173 P175 P177 P183 P185 P193 P197 P201 P203 P205 P209 P211 P213 P219 P223 P225 P230 P236 P238 P243 P247 Patient Age pT 66 68 74 71 68 73 74 58 57 53 62 59 56 62 53 66 68 62 61 72 68 59 66 68 62 57 72 65 73 65 68 66 71 70 75 55 63 66 71 67 72 61 63 73 68 67 56 68 63 66 73 73 71 67 60 64 72 62 67 74 61 46 55 4 3c 2c 3b 3a 3c 3c 3a 2c 3c 3a 2c 3c 3b 3a 3a 3a 3a 3b 3a 2b 3a 3a 3b 3b 2a 2b 2a 2b 3b 2b 3b 4 2b 3a 3a 2b 3a 3a 3a 3a 2b 2b 2b 3c 3a 2b 2b 3b 3a 2a 3a 3a 3b 2a 4 2b 3b 2a 3a 2a 2a 3c pN 2 1 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 1 0 0 0 1 0 1 0 1 0 0 1 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 Gleason Score 10 7 7 7 7 7 8 7 7 7 7 7 9 7 5 8 5 8 9 7 5 7 9 5 9 5 7 5 8 9 4 8 7 5 5 5 5 5 7 9 7 5 7 8 9 6 5 8 8 8 7 7 7 9 7 7 7 6 7 7 6 8 7 % Hypomet. 30 13 0 0 15 1 0 0 0 0 19 0 10 9 13 2 13 6 17 2 0 3 0 11 1 1 0 0 1 1 2 17 16 0 2 5 15 2 49 46 3 8 3 28 3 11 14 0 19 1 0 5 8 6 8 21 4 4 0 4 0 4 3 Gene Copy Number/Cell NKX MYC NKX3.1/MYC 1.05 0.94 1.69 1.38 2.21 0.84 2.06 1.80 2.16 1.70 1.00 1.75 1.01 1.51 1.39 0.79 1.00 2.26 1.91 1.00 1.66 1.01 1.23 3.38 1.19 1.06 1.76 1.26 2.78 1.25 1.65 1.48 1.60 2.82 1.89 1.51 1.51 0.93 1.15 2.58 0.96 1.68 1.98 1.67 1.40 1.64 2.26 2.05 2.32 1.98 2.93 2.63 1.54 2.97 1.24 1.64 2.48 1.04 2.60 2.95 1.87 2.33 2.69 2.35 2.83 1.83 2.98 3.43 2.04 2.19 2.21 1.64 1.92 3.28 2.04 2.33 2.08 3.32 1.49 0.64 1.67 2.04 2.12 2.10 1.76 1.87 2.25 1.72 0.91 2.18 1.14 2.18 1.68 1.87 1.92 1.59 2.57 1.73 1.76 1.65 2.70 2.52 3.35 2.12 1.55 1.62 1.47 2.68 2.26 2.09 2.25 2.25 2.45 2.43 2.12 2.61 2.95 2.08 1.99 2.84 1.51 3.91 2.16 2.01 2.65 2.64 0.45 0.33 0.93 0.46 0.65 0.41 0.94 0.81 1.32 0.88 0.31 0.86 0.43 0.72 0.42 0.53 1.56 1.35 0.94 0.47 0.79 0.57 0.66 1.50 0.69 1.16 0.81 1.11 1.27 0.75 0.88 0.77 1.01 1.10 1.09 0.86 0.92 0.34 0.46 0.77 0.45 1.08 1.22 1.14 0.52 0.72 1.08 0.91 1.03 0.81 1.21 1.24 0.59 1.01 0.60 0.82 0.87 0.69 0.67 1.37 0.93 0.88 1.02 172 european urology 49 (2006) 169–175 respectively. LINE-1 hypomethylation was determined quantitatively by Southern blot analysis as described [15,17]. In keeping with previous reports, normal methylation was defined as <4% hypomethylation, moderate hypomethylation as 4–12% and pronounced hypomethylation as >12%. 2.3. Statistical procedures The non-parametric Mann-Whitney-U test was applied to compare samples with continuous variables (% hypomethylation, NKX3.1:MYC ratio) grouped by nominal data (e.g. pT2 vs. pT3/4, Gleason <7 vs. 7, MYC gain yes/no) as variables did not comply with Gaussian distribution according to the ShapiroWilk method. As a measure of association for nominal data Fisher’s exact test was performed. All p-values were based on two-sided tests and the threshold to accept statistical significance was set at the alpha level 0.05. Analyses were performed with the statistical software package SPSS version 10.0 (SPSS Inc., Chicago, USA). 3. Results Overall, 63 prostate carcinoma tissues were investigated. In 42 (67%) specimens, NKX3.1 copy numbers were significantly altered (see Methods section for cut-offs used), with a decrease indicative of gene loss in 32 (51%) cases. Gene copy numbers of MYC were altered also (incidentally) in 42 (67%) specimens, with a gain in 38 (60%) cases. The ratio of NKX3.1 to MYC gene copy number (NKX3.1:MYC ratio for short) was significantly altered in 45 (71%) specimens, with a decrease in all but two of them. Loss of NKX3.1 correlated significantly with tumor stage ( p < 0.03 for loss). By comparison, MYC gain was rather associated with Gleason grading, showing a highly significant relationship with the prime Gleason grade ( p < 0.003) and a weaker relationship to Gleason sum ( p < 0.03 for Gleason 7). Importantly, the NKX3.1:MYC ratio was strongly correlated to tumor stage. Specifically, a highly significant ( p < 0.002) discrimination between organ-confined cancers (pT2) and cancers extending beyond the prostate (pT3/4) was obtained (Fig. 1). Remarkably, the 15 cancers with the lowest ratios, but also the 2 cancers with the highest ratios, were all staged as pT3 or T4 (Fig. 1). At the optimal cut-off value of 0.85 determined by a receiver operating characteristic (ROC) curve, the NKX3.1:MYC ratio displayed a 66% sensitivity and a 87% specificity in identifying cancers with a stage exceeding pT2. Among the 63 specimens, 32 (51%) did not show significant hypomethylation of LINE-1 sequences compared to normal prostate tissue. Fifteen (24%) carcinomas had moderate hypomethylation and in 16 (25%) specimens hypomethylation was pronounced. Percent LINE-1 hypomethylation was Fig. 1 – Relationship of NKX3.1:MYC gene copy number ratio to prostate carcinoma stage. Box plots of NKX3.1:MYC gene copy number ratio as determined by real-time quantitative PCR for prostate carcinomas staged as pT2 (left) or pT3 or T4 (right). Mean W 2  SD for normal samples are indicated by the horizontal lines. highly significantly related to tumor stage (pT2 vs. pT3/4; p < 0.004), but this value did not discriminate advanced stage cancers quite as well as the NKX3.1:MYC ratio. At a cut-off value of >3% (corresponding to none vs. moderate or pronounced hypomethylation) a sensitivity of 59% and a specificity of 77% was achieved. Interestingly, neither copy numbers of MYC nor NKX3.1 appeared related to hypomethylation. The NKX3.1:MYC ratio was likewise not significantly related to LINE-1 hypomethylation as illustrated by Fig. 2. The figure also illustrates that the combined use of the two markers does not improve the distinction between cancers staged as pT2 vs. pT3 and pT4. In a multivariate analysis using a generalized linear model with a logit link function the predictive value for capsular penetration of the NKX3.1:MYC ratio was compared to the additional clinical parameters Gleason grade, Gleason sum, preoperative PSA value, and the result of digital rectal examination. Only the NKX3.1:MYC ratio remained significant at p = 0.03. LINE-1 hypomethylation did not retain a significant predictive value in multivariate analysis. Neither of the parameters measured (MYC and NKX3.1 copy numbers and their ratio, LINE-1 hypomethylation) was significantly related to biochemical recurrence during the follow-up period so far (average 32 months). 4. Discussion Previous studies have concomitantly demonstrated a relationship between alterations on chromosome 8 and clinical parameters of prostate carcinoma european urology 49 (2006) 169–175 173 Fig. 2 – Relationship of NKX3.1:MYC gene copy number ratio to LINE-1 hypomethylation. (A) The NKX3.1:MYC gene copy number ratio as determined by real-time quantitative PCR is plotted against the extent of LINE-1 hypomethylation as determined by Southern blot analysis. Tumors staged as pT2 are indicated by circles and tumors as pT3 or T4 are indicated by diamonds. The hatched rectangle indicates the area in which normal sample would be located. There is no correlation between gene copy number changes and hypomethylation. The figure illustrates however how higher stage carcinomas tend to present with more pronounced hypomethylation as well as a decreased ratio. Also note that even an optimal combination of hypomethylation and NKX3.1:MYC ratio values (indicated by the dotted line) does not improve sensitivity compared to the NKX3.1:MYC ratio alone: Even though one T2 carcinoma remains in the area below the line, 15 T3 carcinomas would be wrongly classified. (B) An example of LINE-1 hypomethylation analysis: DNA is cut with the methylation-sensitive restriction enzyme HpaII (H) or its insensitive isoschizomer MspI (M), run on an agarose gel, blotted and hybridized with a LINE-1 specific probe. In normal prostate tissue (PN), almost no low molecular weight bands are observed in the HpaII lane, whereas bands appear at the indicated sizes in prostate carcinoma that can be quantified as indicated at the bottom of the figure. such as tumor stage and Gleason score, as well as tumor recurrences after radical prostatectomy [2– 11]. Remarkably, associations were found with losses on 8p in some studies [4,7,8,11] and with gains at 8q in others [5,9,10]. The biological basis of these findings is still not understood. A plausible, but unproven hypothesis is that the progression rate of prostate cancer depends on a balance between several genes located on 8p and 8q. Regarding prostate carcinoma diagnostics the robust association between chromosome 8 alterations and tumor progression raises the interesting prospect that measurements of surrogate parameters that are regularly associated with chromosome 8 alterations could be employed to gain information on an individual prostate carcinoma. We have therefore investigated two parameters related to chromosome 8 alterations. The NKX3.1:MYC copy number ratio is obviously related to chromosome 8 alterations, because the two genes are located on different arms in regions regularly affected by losses and gains, respectively. LINE-1 hypomethylation has been reported to be highly significantly associated with chromosome 8 alterations [15], but this association could well be indirect (see below). In the present study, both parameters showed strong correlations to tumor stage. Interestingly, neither of them was significantly associated with Gleason score, although gains in MYC were, in accord with previous studies [9,10]. Correct prediction of prostate carcinoma stage prior to surgery is currently imperfect, even though it would be desirable for an optimal choice of therapy. According to the results with the present series, the NKX3.1:MYC ratio measured by a relatively simple quantitative PCR assay appeared to be particularly useful for the distinction between carcinomas confined to the organ and those extending beyond it. The ratio remained a significant predictor of capsular penetration in multivariate analysis, suggesting that it may provide additional information to the clinical parameters. If this result can be corroborated in further, ideally prospective studies on distinct and larger tumor series, the determination of this ratio might become a useful addition to nomograms currently in use. An advantage of this 174 european urology 49 (2006) 169–175 assay would be its speed and requirement for minimal amounts of DNA. It could therefore be applied to biopsy samples containing a suitable fraction of carcinoma cells prior to surgery aiding in refinement of the therapeutic strategy. In prostate cancer, most losses of 8p decrease the relative NKX3.1 copy number [19] and most gains of 8q increase the relative MYC gene copy number [10]. Thus, the majority of alterations at 8p or 8q alone or combined ought to result in a change in their ratio. In contrast, it is unclear whether and how LINE-1 hypomethylation and chromosome 8 alterations are mechanistically linked. Both have been found to be common in very advanced cases of prostate cancer, leading to a highly significant statistical association [15]. The present series contained a much lower fraction of androgen-refractory and/or metastatic cases and no significant association between chromosome 8 gene copy number changes and LINE-1 hypomethylation was observed. This finding argues that the two alterations are not mechanistically linked to each other, but instead are both characteristic of very advanced prostate carcinomas. At any rate, LINE-1 hypomethylation did not distinguish tumor stages as well as the NKX3.1:MYC ratio. The use of hypomethylation for molecular staging of prostate cancer is also hampered by the lack of a simple PCR-based assay for its detection. The technique used in the present study is sensitive and quantitative, but requires moderate amounts of high molecular weight DNA, which cannot always be obtained from biopsies. As very recently new techniques for the detection of DNA hypomethylation have been proposed [20,21], it will be interesting to explore their potential in the context of prostate carcinoma. In conclusion, our analysis suggests that both the determination of the NKX3.1:MYC copy number ratio and the detection of LINE-1 DNA hypomethylation might be helpful to predict prostate carcinoma stage. The gene copy number ratio appears to be more useful, being technically more practical and exhibiting a higher sensitivity and specificity for the distinction between organ-confined and advanced stage cancers. We therefore suggest that this technique should be tested more widely to explore and develop its potential in the clinic. Acknowledgements We are most grateful to Ms. Christiane Hader for excellent technical assistance. The study was supported by the Deutsche Krebshilfe (grant numbers 70-2936 Wu I and 70-3193 Schu 1). References [1] Dong JT. Chromosomal deletions and tumor suppressor genes in prostate cancer. Cancer Metast Rev 2001;20:173– 93. [2] Takahashi S, Alcaraz A, Brown JA, et al. Aneusomies of chromosomes 8 and Y detected by fluorescence in situ hybridization are prognostic markers for pathological stage C (pT3N0M0) prostate carcinoma. Clin Cancer Res 1996;2:137–45. [3] Alers JC, Krijtenburg PJ, Rosenberg C, et al. Interphase cytogenetics of prostatic tumor progression: specific chromosomal abnormalities are involved in metastasis to the bone. Lab Invest 1997;77:437–48. [4] Jenkins R, Takahashi S, DeLacey K, et al. Prognostic significance of allelic imbalances of chromosome arms 7q, 8p, 16q, and 18q in stage T3N0M0 prostate cancer. Genes Chromosomes Cancer 1998;21:131–43. [5] Sato K, Qian J, Slezak JM, et al. Clinical significance of alterations of chromosome 8 in high-grade, advanced, nonmetastatic prostate carcinoma. J Natl Cancer Inst 1999;91:1574–80. [6] Alers JC, Rochat J, Krijtenburg PJ, et al. Identification of genetic markers for prostatic cancer progression. Lab Invest 2000;80:931–41. [7] Macoska JA, Trybus TM, Wojno KJ. 8p22 loss concurrent with 8c gain is associated with poor outcome in prostate cancer. Urology 2000;55:776–82. [8] Matsuyama H, Pan Y, Oba K, et al. Deletions on chromosome 8p22 may predict disease progression as well as pathological staging in prostate cancer. Clin Cancer Res 2001;7:3139–43. [9] Steiner T, Junker K, Burkhardt F, et al. Gain in chromosome 8q correlates with early progression in hormonal treated prostate cancer. Eur Urol 2002;41:167–71. [10] Tsuchiya N, Slezak JM, Lieber MM, et al. Clinical significance of alterations of chromosome 8 detected by fluorescence in situ hybridization analysis in pathologic organ-confined prostate cancer. Genes Chromosomes Cancer 2002;34:363–71. [11] Matsuyama H, Pan Y, Yoshihiro S, et al. Clinical significance of chromosome 8p, 10q, and 16q deletions in prostate cancer. Prostate 2003;54:103–11. [12] Macoska JA, Beheshti B, Rhim JS, et al. Genetic characterization of immortalized human prostate epithelial cell cultures. Evidence for structural rearrangements of chromosome 8 and i(8q) chromosome formation in primary tumor-derived cells. Cancer Genet Cytogenet 2000;120:50–7. [13] Birnbaum D, Adélaide J, Popovici C, et al. Chromosome arm 8p and cancer: a fragile hypothesis. Lancet Oncol 2003;4:639–42. [14] Kindich R, Florl AR, Jung V, et al. Application of a modified real-time PCR technique for relative gene copy number quantification to the determination of the relationship between NKX3.1 loss and MYC gain in prostate cancer. Clin Chem 2005;51:649–52. [15] Schulz WA, Elo JP, Florl AR, et al. Genome-wide DNA hypomethylation is associated with alterations on chro- european urology 49 (2006) 169–175 [16] [17] [18] [19] [20] [21] mosome 8 in prostate carcinoma. Genes Chromosomes Cancer 2002;35:58–65. Ehrlich M. DNA methylation in cancer: too much, but also too little. Oncogene 2002;21:5400–13. Florl AR, Steinhoff C, Müller M, et al. Coordinate hypermethylation at specific sites in prostate carcinoma precedes LINE-1 hypomethylation. Brit J Cancer 2004;91: 985–94. Yegnasubramanian S, Kowalski J, Gonzalgo ML, et al. Hypermethylation of CpG islands in primary and metastatic human prostate cancer. Cancer Res 2004;64: 1975–86. Asatiani E, Huang WX, Wang A, et al. Deletion, methylation, and expression of the NKX3.1 suppressor gene in primary human prostate cancer. Cancer Res 2005; 65:1164–73. Chalitchagorn K, Shuangshoti S, Hourpai N, et al. Distinctive pattern of LINE-1 methylation level in normal tissues and the association with carcinogenesis. Oncogene 2004;23:8841–6. Yang AS, Estecio MR, Doshi K, et al. A simple method for estimating global DNA methylation using bisulfite PCR of repetitive DNA elements. Nucleic Acids Res 2004;32:e38. Editorial Comment Henk van der Poel, Amsterdam, The Netherlands h.vd.poel@nki.nl The central role of the MYC oncogene in prostate cancer is proven by several genetic experiments [1,2]. Its combination with other potential oncogenes such as ras and Pim1 already clarified that its role should be seen in the context of other (potential) oncogenes and tumor suppressor genes. The loss of the NKX3.1 gene, a homeobox gene, is an early event in prostate cancer development. Loss of NKX3.1 function was recently shown to be involved in the oxidative stress response [3] and thus the loss of NKX3.1 function fits the hypothesis that prostate cancer is related to inflammatory abnormalities [4]. Both loss of NKX3.1 and overexpression of MYC have been described in presumed premalignant lesions such as prostate intraepithelial neoplasia and low grade cancer. Here, Kindich et al. also suggest that gene copy loss of NKX3.1 was associated with Gleason score whereas MYC gene amplification increased with tumor grade. In this relatively small population the authors claim that the ratio of NKX3.1 and MYC 175 gene copies had additional predictive value to known clinical parameters such as Gleason score, DRE findings, and preoperative PSA with respect to the discrimination between pT2 and pT3 tumors. These data would be interesting for staging if the results are reproducible on prostate biopsies. Until then, these findings will not change urological practice. More importantly, these data do provide ground for new hypotheses on prostate cancer progression. Unfortunately, the authors fail to discuss such hypotheses. It remains to be proven that gene copy number is associated with protein levels but assuming that such a relation is present in these prostate cancer, one can hypothesize that a more aggressive phenotype, i.e. extracapsular extension or nodal metastases, seem dependent on responses to oxidative stress as is suggested by the correlation with a (slightly) lower NKX3.1 copy number (Table 1). If oxidative responses are important in prostate cancer progression, this may provide entries for therapeutic exploitation. Moreover, the higher copy numbers of MYC in pT3 tumors suggests that these tumors may depend on the dedifferentional properties of MYC for extraprostatic growth. Data from analyses like presented by Kindich et al. should be seen in the light of available mechanistic information rather than make authors suggest that these markers may help daily urology, yet. References [1] Ellwood-Yen K, Graeber TG, Wongvipat J, Iruela-Arispe ML, Zhang J, Matusik R, et al. Myc-driven murine prostate cancer shares molecular features with human prostate tumors. Cancer Cell 2003;4:223–38. [2] Williams K, Fernandez S, Stien X, Ishii K, Love HD, Lau YF, et al. Unopposed c-MYC expression in benign prostatic epithelium causes a cancer phenotype. Prostate 2005; 63:369–84. [3] Ouyang X, DeWeese TL, Nelson WG, Abate-Shen C. Loss-offunction of Nkx3.1 promotes increased oxidative damage in prostate carcinogenesis. Cancer Res 2005;65:6773–9. [4] Nelson WG, De Marzo AM, DeWeese TL, Isaacs WB. The role of inflammation in the pathogenesis of prostate cancer. J Urol 2004;172:S6–11.