- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
Since its introduction into clinical practice in 1988, serum PSA measurement has had a profound impact on the diagnosis and management of early localized prostate cancer, referred to as ‘stage migration’. This term describes the continuous increasing incidence of prostate cancer with a concomitant shift towards earlier, localized stages at diagnosis [1,2]. Despite lacking specificity, serum PSA level is currently recognized as the most useful tumour marker available in clinical practice for the early detection of prostate cancer. However, the ability of PSA to predict pathological extent and biochemical recurrence, and to identify those at risk of failed treatment, remains inadequate because it lacks cancer specificity [3–5]. Therefore, novel molecular markers to enhance current detection, and staging, as well as prognostic tools are expected to have a major impact on cancer care in the growing population at risk of prostate cancer.
The development of prostate cancer is associated with allelic losses and gains accumulating during tumour growth and disease progression [6,7]. Investigations of prostate tumour tissues showed that loss of heterozygosity (LOH) at specific chromosomal loci might play key roles in cancer progression [8–11]. However, prostate cancer is often multifocal and each focal area might harbour different genetic alterations . The heterogeneity of these tumours might require the examination of several areas to obtain general information on genetic alterations associated with tumour growth. As blood is an easily accessible pool for circulating DNA, it is a source of tumour-derived DNA. We recently detected free tumour DNA with specific allelic losses in the blood circulation of patients with prostate cancer [13–15], as was shown for other cancers previously [16–22].
Thus the aim of the present study was to identify tumour-specific genetic aberrations in cell-free plasma DNA derived from the peripheral blood of 273 patients with prostate cancer, and to investigate their clinical relevance as a diagnostic tool.
PATIENTS AND METHODS
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
Peripheral blood samples were collected prospectively before surgery from 230 patients with prostate cancer and 43 with BPH, between October 2003 and December 2004. Formalin-fixed, paraffin-embedded prostate tumour blocks were available from 101 of the patients with cancer. Informed consent was obtained from all patients, and the study was approved by the local Hamburg research ethics committee. All 230 patients with biopsy-confirmed clinically localized prostate cancer had a radical retropubic prostatectomy (RRP). The RRP specimens were processed according to the Stanford protocol and graded using the Gleason score. Total prostate volume was determined using a 7 MHz endorectal coil and TRUS. In addition, 15 patients with advanced tumour stages were included. The clinical stage was assigned by the attending urologist according to the TNM classification of 2002.
For microscopic evaluation the paraffin-embedded prostate tumours were stained with haematoxylin and eosin (Merck, Darmstadt, Germany). Tumour tissues were microdissected by scraping off tissue from stained slides under a microscope, or by computer-assisted microdissection from unstained slides using a laser pressure catapult.
After blood samples were collected into tubes containing EDTA the samples were immediately stored at 4 °C and processed within 4 h to avoid cell lysis. The blood was fractionated by initial centrifugation of 20 mL EDTA/blood at 2500g for 10 min. The upper phase contained the plasma, from which 3–4 mL was removed for extraction of circulating DNA. To isolate the leukocytes the remaining 16–17 mL of blood was made up to 50 mL with lysis buffer containing 0.3m sucrose, 10 mm Tris-HCl pH 7.5, 5 mm MgCl2 and 1% Triton X100 (Sigma, Taufkirchen, Germany), incubated for 15 min on ice, and centrifuged at 2500g, at 4 °C for 20 min. Genomic DNA was extracted from microdissected tumour tissues, plasma and leukocytes of each patient using the QIAamp DNA Mini Kit and a vacuum chamber (Qiagen QIAvac24) according to the manufacturer instructions (Qiagen, Hilden, Germany). The corresponding leukocyte DNA served as reference DNA.
For the PCR-based fluorescence microsatellite (MS) analysis, 10 ng of DNA samples were amplified in a 10-µL reaction volume containing PCR Gold buffer (150 mm Tris-HCl, pH 8.0 and 500 mm KCl), 2.5 mm MgCl2 (Applied Biosystems, Mannheim, Germany), 200 µm dNTPs (Roche, Mannheim, Germany), 11 pm of primer sets (Sigma, Taufkirchen, Germany) and 1.5 U AmpliTaq Gold DNA-Polymerase (Applied Biosystems). The following MS markers were used: THRB (3p24), D6S474 (6q21–22), D6S1631 (6q16), D7S522 (7q31.1), D8S87 (8p12), D8S137 (8p21.1), D8S286 (8q21.3), D8S360 (8p21), D9S171 (9p21–22), D9S1748 (9p21), D10S1765 (10q23.3), D11S898 (11q22) and D11S1313 (11q11-p11). The sense primer was fluorescence-labelled (HEX, FAM or TAMRA) at the 5′ end. The reaction was started with activation of the DNA polymerase for 5 min at 95 °C, followed by 40 cycles of PCR amplification.
Fluorescence-labelled PCR products were separated by capillary gel electrophoreses and detected by the fluorescence laser of the Genetic Analyser 310 (Applied Biosystems). The peaks were evaluated using the GeneScan Analysis program. The 500-ROX size marker served as an internal standard. LOH was calculated by dividing the quotient of the peak intensity derived from plasma or tumour DNA by the quotient of the intensity derived from corresponding leukocyte DNA. LOH was interpreted if the final quotient was <0.6 or >1.67. MS instability (MSI) was defined by the occurrence of several additional peaks. Uninformative cases were those peaks that were homozygous or not analysable.
The results were analysed statistically using the chi-square test, Fisher’s exact test and binary logistical regression to identify possible associations of patterns of LOH or MSI with the clinical variables, e.g. clinical stage, Gleason score before and after RRP, total PSA (tPSA) and percentage free PSA (%fPSA) levels, prostate volume and margin status, with P < 0.05 considered to indicate statistical significance.
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
Most patients with prostate cancer (80%) had clinical stage T1; the median (range) pre-treatment PSA level was 5.5 (0.2–8950) ng/mL, and the median pre-treatment %fPSA level 13.2 (0.7–53.2)%. The Gleason score before RRP was 3 + 3 in 74% of the men, and in the RRP specimen the Gleason score range was 2 + 3 to 5 + 4, being 3 + 3 in half the men; 85% of the patients had a tumour stage of T2a, b or c, and 15% of 3a or b.
After extracting the DNA the content was measured spectrophotometrically; there was a wide range of DNA yields in the plasma of the 230 patients with prostate cancer and 43 with BPH. In patients with cancer the mean (median, range) DNA concentration was 1395 (735, 11–26 023) ng/mL. Compared with the higher DNA concentrations in the blood of patients with cancer, those with BPH had minor levels of DNA in their plasma; the mean (median, range) DNA concentration was 804 (500, 77–2908) ng/mL.
To evaluate the incidence of LOH and MSI at 13 different microsatellite markers in blood plasma and tumour tissues, we assessed the free plasma DNA from all patients. Comparative PCR-based fluorescence MS analyses with matched tumour tissues from 101 patients with prostate cancer were done. Using the panel of 13 polymorphic MS markers and extracted plasma DNA, we detected LOH and MSI in 8.5% and 2.5% of the informative cases, respectively. In the group of BPH specimens LOH and MSI were less common, at 2.0% and 1.5% of the analysed plasma samples, respectively. Microdissected tumour tissues were a more appropriate source for detecting allelic imbalance (AI; LOH plus MSI). Frequencies of LOH and MSI were 23.5% and 10.5% of all analyses, respectively. For each patient with prostate cancer or BPH we calculated the overall number of AI in his plasma and tumour sample. As shown in Table 1, of the 230 patients with cancer and 43 with BPH, 98 (42.5%) and 13 (30%), respectively, had at least one AI at any of the loci in their plasma sample. In comparison to the 42.5% of patients with cancer and AI in their free plasma DNA, twice as many (87%) had at least one AI in their primary tumour. Three patients had seven, eight or nine genetic alterations in their tumour sample, whereas one patient had seven aberrations in his plasma sample (Table 1). Of LOH recorded in plasma, 63% were also retrieved in the paired tumour samples. The highest concordance in the results (including LOH, MSI and retention of heterozygosity) between tumour and blood samples was detected at the marker D8S360 (83%) and the lowest at D8S137 (47.5%). Figure 1 shows a representative example of LOH on free plasma DNA (Fig. 1A) and a tumour-associated MSI (Fig. 1B) found in the patients with prostate cancer.
Table 1. Overall numbers of LOH and MSI in plasma and tumour samples from each of the patients
|AI (no. of LOH + MSI)||Cancer, n (%)||BPH, n (%) Plasma|
|0||132 (57.5)|| 13 (13)||30 (70)|
|1|| 41 (18)|| 26 (25)||12 (28)|
|2|| 27 (11.5)|| 20 (20)|| 1 (2)|
|3|| 12 (5)|| 17 (17)|| |
|4|| 10 (4.5)|| 11 (11)|| |
|5|| 5 (2)|| 4 (4)|| |
|6|| 2 (1)|| 7 (7)|| |
|7|| 1 (0.5)|| 1 (1)|| |
|8||–|| 1 (1)|| |
|9||–|| 1 (1)|| |
|Total||230 (100)|| 101 (100)||43 (100)|
Figure 1. An example of LOH and MSI in blood plasma from a patient with prostate cancer. The PCR products of leukocytes and blood plasma DNA derived from the patient were separated by capillary gel electrophoresis on a Genetic Analyser and evaluated with the Gene Scan Analysis program. The abscissa indicates the length of the PCR product, while the ordinate gives information on the fluorescence intensity represented as peaks. The upper diagrams show the leukocyte DNA (reference) and the lower diagrams the blood DNA amplified with the primers as indicated in the upper diagrams. The MS alleles are represented by two heterozygote peaks. In the lower diagram the arrow refers to LOH (A) and the additional peaks represent MSI (B).
Download figure to PowerPoint
Table 2 lists the number and frequency of LOH and MSI detected at the different MS markers in plasma and primary tumours. The highest frequencies of LOH were at markers THRB (18.5%) and D8S137 (14.5%), and MSI at D11S898 (9.0%) and D8S87 (7.5%) in the plasma samples, in addition to the highest incidences of LOH at THRB (36.0%), D7S522 (48.0%) and D8S137 (59.5%), and MSI at D8S87 (29.5%), D11S1313 (25.5%) and D11S898 (22.5%) in the tumour samples. A graphical representation of the frequencies of LOH and MSI are shown in Fig. 2.
Table 2. Number and frequency of LOH and MSI detected in plasma and tumour samples at the different MS markers. The highest frequencies of LOH and MSI detected are in bold
|Marker||Plasma, n (%)||Tumour, n (%)|
|THRB||97|| 18 (18.5)|| 0||39|| 14 (36)|| 0|
|D6S474||166|| 13 (8)|| 0||42|| 9 (21.5)|| 0|
|D6S1631||170|| 10 (6)|| 4 (2.5)||60|| 14 (23.5)||10 (16.5)|
|D7S522||141|| 12 (8.5)|| 2 (1.5)||46|| 22 (48)|| 0|
|D8S87||179|| 4 (2)||13 (7.5)||64|| 4 (6.5)||19 (29.5)|
|D8S137||96|| 14 (14.5)|| 2 (2)||37|| 22 (59.5)|| 0|
|D8S286||158|| 15 (9.5)|| 0||58|| 14 (24)|| 0|
|D8S360||122|| 13 (10.5)|| 1 (1)||47|| 9 (19)|| 1 (2)|
|D9S171||121|| 4 (3.5)|| 5 (4)||48|| 10 (21)|| 6 (12.5)|
|D9S1748||192|| 11 (5.5)|| 2 (1)||72|| 8 (11)|| 2 (3)|
|D10S1765||167|| 15 (9)|| 1 (0.5)||65|| 18 (27.5)|| 2 (3)|
|D11S898||182|| 20 (11)||16 (9)||71|| 14 (19.5)||16 (22.5)|
|D11S1313||191|| 15 (8)|| 7 (3.5)||79|| 14 (17.5)||20 (25.5)|
|Total||1982||164 (8.5)||53 (2.5)||728||172 (23.5)||76 (10.5)|
Figure 2. Comparison of the frequency of events (LOH and MSI) detected at 13 different chromosomal markers in tumour tissues (Tu) and blood plasma (Pl) from patients with prostate cancer. The percentage of events (LOH and MSI) was calculated by dividing the number of LOH or MSI at the specific MS marker by the number of informative cases.
Download figure to PowerPoint
Of the 13 MS markers used, eight (D6S474, D7S522, D8S87, D8S286, D8S360, D9S171, D11S898 and D11S1313) had few LOH, with either one or two LOH in free plasma DNA from patients with BPH (Table 1). In this group, two and three MSI were detected at markers D9S171 and D11S1313, respectively. These data suggest that the markers THRB, D6S1631, D8S137, D9S1748 and D10S1765 are unaffected by genetic alteration in patients with BPH and might be more relevant for malignant lesions.
To determine whether the profiles of LOH and MSI were associated with the established risk factors, they were compared statistically in tumour and plasma samples with age, clinical stage, tPSA level, a combination of tPSA and %fPSA level (tPSA*%fPSA), Gleason scores before and after RRP, tumour stage, total prostate volume and surgical margin. High risk tPSA*%fPSA levels were a tPSA of 4–10 ng/mL plus %fPSA < 15%, and tPSA > 10 ng/mL and %fPSA < 10%; low risk tPSA*%fPSA values were a tPSA of 4–10 ng/mL plus %fPSA > 15% and tPSA > 10 ng/mL plus %fPSA > 21%.
Table 3 shows the significance of LOH and MSI and the combination of LOH and MSI with age of the patients with cancer and increasing total prostate volume for the markers D9S1748 and a combination of D6S474 and D7S522 in tumour tissues, respectively. In addition, higher tPSA values were associated with LOH and MSI at D11S1313 in the primary tumours. In plasma samples there was a significant correlation of total prostate volume with LOH and/or MSI at D7S522, and tPSA*%fPSA levels with LOH at THRB (Table 3). The statistical analyses with the other clinical variables, e.g. Gleason score and tumour stage, were not significant (data not shown).
Table 3. Associations of LOH and MSI at the different markers recorded in tumour and plasma samples with established risk factors, and their statistical significance
|LOH + MSI*||LOH or MSI†|
| Total volume||D6S474*D7S522||<0.04||<0.04|
| Total volume||D7S522||<0.05||<0.05|
| tPSA*fPSA%‡||THRB||<0.05||0 MSI§|
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
We assessed blood plasma from 230 patients with prostate cancer for tumour-specific plasma DNA using a panel of 13 different MS polymorphic markers. Comparative analyses were done for paired tumour samples from 101 patients with cancer and plasma from 43 with BPH. There was LOH and MSI in 2.0% and 1.5% of all informative plasma samples from the patients with BPH, respectively, indicating that genetic aberrations might also occur in benign lesions. However, compared with the rate of LOH and MSI in cell-free plasma DNA from patients with cancer (11%) the events are rare in blood from men with BPH. Our preliminary investigations of blood from 20 healthy men showed retention of heterozygosity and no deleted DNA regions, and support the validity of our assay. By contrast with the present data, AI was reported to occur with high frequency even in normal tissue . As BPH is not a pre-neoplastic lesion, our data show that AI can arise in benign growths and increase in tumour tissues. In the present MS marker panel, five markers (THRB, D6S1631, D8137, D9S1748 and D10S1765) were specific for detecting LOH in clinical samples derived from patients with cancer and are therefore more likely to be relevant for discriminating between prostate cancer and BPH. Two of these markers harbour known candidate target genes. CDKN2/p16 (cyclin dependent kinase 2) and PTEN (phosphatase and tensin homologue), which are negative regulators of the cell cycle, map to D9S1748 and D10S1765, respectively, and have been reported to be frequently mutated and deleted in prostate tumours [24,25]. However, the lack of association between the appearance of LOH and both tumour stage and Gleason scores limits these markers to diagnostic purposes.
The lower incidence of AI in the plasma of the present patients with cancer (11%) compared with the paired primary tumours (34%) might be explained by the prevalence of normal DNA in blood, which could mask LOH in cell-free DNA. Cell-free DNA can also be detected in the blood of healthy individuals and patients with benign lesions, albeit at lower concentrations, and is thought to originate from lymphocytes [13,26]. However, the origin of the significantly higher DNA levels in blood from patients with tumour is still unknown. The finding that tumour DNA is diluted by normal DNA in the blood of patients with cancer was discussed by several groups for other tumours [27–29]. Diehl et al. reported that the proportion of mutant adenomatous polyposis coli DNA fragments in plasma derived from patients with colorectal cancer was small, averaging only 11% of the total DNA fragments, even in large, metastatic cancers. They explained the low occurrence of mutant DNA molecules in the plasma by the prevalence of wild-type DNA fragments, that are discharged from necrotic regions, besides mutant DNA. The present results showing an overall LOH rate of 11% in blood from patients with cancer are consistent with these previous results. Furthermore, a low incidence of MS alterations in the plasma DNA of head and neck squamous cell carcinomas was reported by Coulet et al.; they also described sensitive mutation detection methods that allow a more precise interpretation of LOH. Also, Wang et al. suggested that circulating plasma DNA might not accurately reflect the clinical status of breast tumours, because the mean proportion of LOH was much lower than that in primary tumours. Despite these evident limitations, the present findings show that 63% of LOHs identified in blood could also be detected in the paired primary tumours, and that 83% of the LOHs at the marker D8S360 were concordant in both specimens. Further parallels between plasma and tumour DNA were the high frequencies of LOH recorded at markers THRB and D8S137, and MSI at D11S898 and D8S87. These findings suggest that at least part of the free DNA in blood might originate from the primary tumour. The subset of LOH in blood plasma, which was not concordant with the detected tumour alterations, might be due to the known heterogeneity of prostate tumours. To obtain an exact profile of genetic alterations in these tumours, demanding analyses of several areas of the microdissected tumour tissues are required. As cell-free DNA is supposed to be released into the blood by apoptotic or necrotic cells of various areas of the primary tumour, blood might constitute a reservoir of tumour-specific DNA with diverse genetic alterations. To achieve higher LOH rates, improved techniques are required for an advanced extraction method of tumour-specific plasma DNA. That the use of blood might be a more attractive approach for detecting LOH using an exclusive panel of markers than the use of tumour tissues is supported by easy blood sample collection using simple vein puncture and the heterogeneity of the primary tumours.
In the clinical samples of each patient we determined the overall number of affected markers. Although LOH and MSI could also be detected in the men with BPH the number of AI in the plasma of each patient was low, at none to two. The number of AIs in the primary tumour and blood plasma of each patient with cancer was 0–9 and 0–7, respectively. These findings show that the distribution of AI was heterogeneous in the men with prostate cancer, and that each patient might have an individual LOH and MSI pattern. Most men with prostate cancer had tumour stage pT2 and some of them had high frequencies of AI, indicating that simultaneous DNA deletions and amplifications might occur at an early stage of the cancer. These observations are also consistent with the hypothesis that prostate tumours develop multifactorially .
Comparing the data of AI detected in the primary tumours with established risk factors showed a significant association of advancing age of the patients with AI at marker D9S1748 (P = 0.004), which was not apparent in the plasma samples. Whereas in the tumour samples there was a significant correlation of total prostate volume with AI for the combined marker D6S474 and D7S522 (P < 0.04), such an association was detected only for D7S522 in the plasma samples. There were further relationships of LOH at D11S1313 with increasing tPSA levels (P = 0.005) and at THRB with combined tPSA and %fPSA (P < 0.05) in tumour and plasma samples, respectively. These partly diverse significances between DNA derived from primary tumour and plasma might be caused by the heterogeneity of prostate tumours and the prevalence of normal DNA in blood, which could mask genetic aberrations.
In summary, the present results suggest that assessing genetic alterations in peripheral blood might be a promising clinical variable for detecting and staging prostate cancer. It is expected that an optimized extraction of plasma DNA might improve the recovery rate of LOH. A further optimization of the marker panel will improve the sensitivity and specificity of the method.