Int. J. Cancer: 98, 889 – 894 (2002) © 2002 Wiley-Liss, Inc. DOI 10.1002/ijc.10263
Publication of the International Union Against Cancer
HIGH FREQUENCY OF SERUM DNA ALTERATIONS IN RENAL CELL CARCINOMA DETECTED BY FLUORESCENT MICROSATELLITE ANALYSIS Rolf VON KNOBLOCH1*, Axel HEGELE1, Heidrun BRANDT1, Zoltan VARGA1, Sebastian WILLE1, Tilman KA¨ LBLE2, Axel HEIDENREICH1 and Rainer HOFMANN1 1 Department of Urology, Philipps-University Marburg Medical School, Marburg, Germany 2 Department of Urology, Sta¨dtisches Klinikum Fulda, Fulda, Germany To date there are no reliable serological markers for renal cell carcinoma (RCC). We applied fluorescent microsatellite analysis (MSA) to detect serum DNA alterations in patients with RCC. Fresh tumour, peripheral blood and serum specimens from 60 consecutive patients treated for malignant renal tumours (nⴝ 53 RCC and nⴝ 7 non-RCC) were prospectively collected. After DNA extraction, we performed MSA with a total of 9 markers from the chromosomal regions 3p, 5q, 7p, 7q, 9p, 13q, 17p and 17q to identify tumour specific serum DNA alterations in Group I (nⴝ 53 RCC); 11 additional markers were used in the first 23 RCCs (Group II) in order to increase sensitivity; and 20 healthy controls were investigated with 10 markers. Besides the histomorphological diagnosis the RCCs were genetically stratified according to the “Heidelberg Classification” of renal tumours. Detection of allelic imbalance and loss of heterozygosity (LOH) was carried out on an automated laser sequencer. In Group I we identified serum DNA alterations in 74% (39/53) of cases. When applying 20 markers, the sensitivity was elevated to 87% (20/23) in Group II. Investigating 20 healthy controls with 10 markers, the method rendered 85% specificity. The highest incidence of alterations was detected for chromosomal regions 3p and 5q. The presence of serum DNA alteration was not associated with tumour nuclear grade but exhibited a trend towards advanced stages (p ⴝ 0,044). In RCC, the microsatellite analysis has a high sensitivity in the detection of serum DNA alterations when a sufficient number of markers from various chromosomal regions are used. Advanced tumours tend to express serum DNA alterations more frequently. © 2002 Wiley-Liss, Inc. Key words: serum DNA alterations; renal cell carcinoma; genetics, microsatellite analysis; follow-up
Renal cell carcinomas (RCC) represent approximately 2% of all newly diagnosed malignancies. At present, there are no reliable serological tumour markers available for renal cell carcinoma. Therefore, diagnosis and follow-up of renal malignancies is dependent on imaging studies to morphologically detect solid tumour growth. Primary RCCs are commonly diagnosed in routine abdominal sonography. After surgical therapy, again ultrasound and also CT or MRI scanning are imaging tools to detect local or distant tumour recurrence. As RCC is accepted to be chemo- and radio-resistant, the only curative therapy of recurrent or metastatic tumour growth is total surgical excision with the option of adjuvant immunomodulatory treatment. An ideal tumour marker with a high sensitivity and specificity would offer the opportunity to early detect even occult micrometastatic tumour spread long before imaging studies would be capable to identify solid tumour growth. The low tumour burden could possibly increase the response to immunotherapy. Recent studies have proved molecular techniques applicable for the detection of smallest amounts of circulating tumour DNA in serum and plasma of cancer patients.1–3 Employing the microsatellite analysis (MSA), Goessl and co-workers identified plasmaDNA alterations in 63% of patients with RCC applying only 4 markers for chromosome 3p.2,3 With renal cell tumours being one of the best genetically characterised tumour entities having already lead to a genetic classification system (“Heidelberg Classification of renal cell tumours”4), we are aware that only the conventional
or non-papillary renal cell tumours harbour 3p deletions in up to 98% of cases.4 Therefore, we extended the approach to diagnose serum DNA alterations in RCC by applying up to a total of 20 microsatellite markers from the chromosomal regions 3p, 5q, 7p, 7q, 8p, 9p, 13q, 17p and 17q in order to raise the sensitivity of the method and to identify highly sensitive markers. Furthermore, the application of a higher number of markers from numerous chromosomal regions also allowed the stratification of the primary RCCs according to the genetical Heidelberg Classification.4 MATERIAL AND METHODS
Tumour and blood sampling From 1998 –1999, we prospectively collected preoperative peripheral blood samples (10 ml EDTA blood and 10 ml full blood samples) in 53 consecutive RCC cases. A representative fresh tumour biopsy was collected intraoperatively by resection from the nephrectomy specimen. Hereby, we paid careful attention in resecting a macroscopically “clean” tumour sample not contaminated by healthy kidney tissue. All samples were immediately shock frozen in liquid nitrogen and stored at ⫺80°C prior further processing. Only the full-blood specimen was centrifuged for 10 min at 3,000g to obtain the serum from the supernatant and the isolated serum was then stored as mentioned above. Additionally, 10 ml EDTA blood and 10 ml full blood were collected from 20 healthy individuals (10 men and 10 women) for a control group and for comparison reasons to demonstrate that the choice of markers is more specific for RCC blood and tumour samples from 7 consecutive non-RCC renal tumours (1 renal B cell lymphoma and 6 urothelial pelvis tumours) were also collected and analysed. The pathohistological diagnosis of the tumour was obtained from the Department of Pathology at the Philipps-University Marburg Medical School for the main tumour specimen, which was performed according to the UICC classification of 1997.5 Furthermore, the primary tumours were genetically stratified according to the “Heidelberg Classification of renal cell tumours”.4 Prior to collecting blood and tissue samples an informed consent was obtained from the patients. The Heidelberg classification of renal cell tumours (brief)4 The applied classification subdivides renal tumours of parenchymal origin into benign and malignant. Every subcategory is limited to the most commonly documented genetic alterations. Benign tumours are stratified into metanephric adenomas and Grant sponsor: Stiftung P.E. Kempkes, Marburg; grant number: 12/1998 *Correspondence to: Department of Urology, Philipps-University Marburg Medical School Baldingerstrasse, D-35043 Marburg/Lahn, Germany. Fax: ⫹49-6421-286-5590. E-mail:
[email protected] Received 19 July 2001; Revised 22 November 2001; Accepted 30 November 2001 Published online 9 February 2002
890
VON KNOBLOCH ET AL.
adenofibromas, papillary renal cell adenomas and renal oncocytomas. The malignant tumours are subclassified into non-papillary or conventional renal cell carcinomas characterised by loss of chromosome 3p as well as chromosomal alterations ⫹5q, ⫺6q, ⫺8p, ⫺9p and ⫺14q; papillary renal cell carcinomas characterised by chromosomal alterations ⫹3q, ⫹7, ⫹8, ⫹12, ⫹16, ⫹17, ⫹20 and ⫺y; chromophobe renal cell carcinomas characterised by chromosomal alterations ⫺1, ⫺2, ⫺6, ⫺10, ⫺13, ⫺17 and ⫺21; collecting duct carcinomas, with medullary carcinomas of the kidney displaying no consistent chromosomal alterations and unclassifiable renal cell carcinomas. The genetic subcategories correlate with distinct histological findings. This genetic classification system was applied to our cohort of renal cell tumours, as it does not rely on intervariable histomorphological findings. The classification system further was the basis for choosing the series of microsatellite makers for our study. DNA isolation A 5–10 mm piece of frozen tumour tissue was allowed to thaw in TE9 buffer (0.5 M Tris-Cl, pH 9, 0.1 M EDTA) in a Petri dish. The tumour cells then were carefully scraped and pressed from the connective tissue. The tumour cells afterwards were resuspended in TE9 buffer and incubated in 2% sodiumdodecylsulfate (SDS) plus 0.5 mg/ml proteinase-K for at least 3 hr at 56°C. After digestion, the DNA was isolated by the phenol chloroform method with final ethanol precipitation. Extracted DNA was resuspended in TE1 buffer (10 mM Tris-Cl, pH 7.5, 1 mM EDTA). By this technique the contamination of the tumour DNA with healthy tissue DNA was reduced to a minimum. To obtain the corresponding normal DNA the same method was applied with peripheral blood lymphocytes from the 10 ml EDTA blood samples. DNA extraction from the serum probes was carried out with the Qiamp Midi-Kit (Qiagen, Hilden, Germany) according to the protocols for blood and body fluids supplied by the manufacturer. Serum DNA extraction was performed with 2 to 4 ml of clear serum supernatant obtained after centrifugation of 10 ml of full-blood samples at 3,000g for 10 min. Serum DNA concentrations were measured photometrically. Microsatellite analysis and PCR conditions For the identification of tissue and serum alterations in Group I 9, highly polymorphic markers for the chromosomal regions 3p, 5q, 7p, 7q, 9p, 13q, 17p and 17q (D3S1560, D3S2450, D5S1720, D7S1796, D7S1807, D9S925, D13S153, D17S799 and D17S1306) were used in the PCR-based MSA in all of the 53 RCCs and the 7 non-RCC renal tumours. In Group II, the first
consecutive 23 RCCs were analysed with additional 11 markers (a total of 20 markers), hereby also incorporating chromosome 8p (D3S3666, D3S2408, D3S1259, D5S1480, D5S476, D5S818, D8S261, D8S560, D17S783, D17S1298 and D17S807). The rationale for increasing the number of markers was to evaluate whether the higher number of markers would significantly raise the sensitivity of the method. Moreover, testing more markers from chromosomal regions known to be involved in RCC tumourigenesis offers a better chance to identify highly sensitive markers for the method of serum DNA analysis in RCC. DNA sequences for the microsatellite markers were obtained from the genome database either under http://lpg.nci.nih.gov/html-chlc/ChlcMarkers.html or http://carbon.wi.mit.edu:8000/cgi-bin/contig/sts_info? and are shown in Table I; 50 –100 ng normal, tumour and serum DNA were used as templates in 10 l PCR reactions with 50 mM KCL, 10 mM Tris-Cl pH 8.3, 1.5 mM MgCl2, 200 moles each dNTP (deoxynucleotide triphosphate), 10 pmol forward primer (labelled with Cy-5), 10 pmol reverse primer, 0.01% BSA and 0.5 U Taq DNA polymerase (Promega Madison, WI). Amplifications were performed using the same conditions for all microsatellites in a PTC100 thermocycler (MJ Research, Watertown, MA) with initial denaturation at 95°C for 2 min, followed by 28 cycles each with 40 sec at 95°C, 30 sec at 55°C followed by 30 sec at 72°C. After a final extension at 72°C for 10 min, 30 l of stop solution (50 mM EDTA and 5 mg/ml Dextraneblue 2000 in 100% deionized formamide) were added to each reaction. Four microlitres of the PCR products were separated on 8% polyacrylamide gels (ReproGel “long read”, Amersham Pharmacia Biotech, Freiburg, Germany) at constant 1,500 V, 60 mA, 30 W and 55°C gel temperature on the 200 mm gel cassettes in 0.5⫻ TBE buffer for 5 hr. Fragment analysis was performed on an automated DNA laser sequencer (ALFexpressII威, Amersham Pharmacia Biotech, Freiburg, Germany). Results were computed by using the Fragment Manager (FM 1.02, Amersham Pharmacia Biotech) software. Allelic imbalance (AI) in heterozygous PCR products was described as LOH or deletion, when loss of genetic information was known to occur for this chromosomal region, but AI could also describe differences in allele intensity caused by genetic amplification from, i.e., duplication, when known for the specific site. AI caused by either loss or gain was summarised as alteration of the genetic locus. AI identification was qualitative and not quantitative using measured cut offs. The sensitivity of each microsatellite marker in detecting tumour specific alterations in serum was evaluated independently and in combination with the other markers. The sensitivity was
TABLE I – MICROSATELLITE MARKERS USED FOR PCRS Chromosome
Marker name
Nucleotide repeat1
3p
D3S1560 D3S2450 D3S3666 D3S2408 D3S1259 D5S1480 D5S476 D5S1720 D5S818 D7S1797 D7S1807 D8S261 D8S560 D9S925 D13S153 D17S783 D17S1298 D17S799 D17S1306 D17S807
CA GATA CA TTA CA AAT CA GATA GATA GATA GGAA CA CA GATA CA CA GAAT CA GGAA CA
5q
7p 7q 8p 9p 13q 17p 17q 1
Product size2
242 258 135 188 187 227 176 220 157 229 236 143 143 167 221 236 252 185 179 115
Sequences3 (5⬘—3⬘)
F: F: F: F: F: F: F: F: F: F: F: F: F: F: F: F: F: F: F: F:
GCATCTACAGGGGGTGTCT, R: AGGCTGATTTTCAGCACAA ACCTTCCTCAGGGGAATAGA, R: AATTTGCAAACAATCTCCCA GGGACTGGNCAAAGTT, R: CAAATACCCTGTAAGGCACT ACCTGCTAACACTGACAGTGC, R: ACACAGATACCAATGGGTGG GCTGGACTATATTTGAAACTCATT, R: TTTCAGTGAGCCAAGATCGT TTGGGAAGAATAGCTTTCCC, R: TTCTAGCTTCCCCCTATGCT CCTCCTGACAGCAGATATGA, R: CTGAGGTCTCTCCTGGGTGT TCCCAACAAAGTCATTTGGT, R: TAGCTCCCACAATCATGTGA GGGTGATTTTCCTCTTTGGT, R: TGATTCCAATCATAGCCACA TTCAAGAGCTAATCCATGCC, R: AAATTGAGATCGCAGCTGAC TCCTTTTCCTTTTCCCTTTC, R: ATTAATAGGTTTGTCACGATTAACC TGCCACTGTCTTGAAAATCC, R: TATGGCCCAGCAATGTGTAT GGCATTTCAGAGGACC, R: TGCAAAGATGGGCTCAG TGTGAGCCAAGGCCTTATAG, R: GTCTGGGTTCTCCAAAGAAA AGCATTGTTTCATGTTGGTG, R: CAGCAGTGAAGGTCTAAGCC AGGACTCGAAATGCTTTCAT, R: TAACAGAAAACTTGGAGCCG TTCACACTTTTAGTGTTGGTGG, R: TGACCTGAGTTTGACTGGGT ATTGCCAGCCGTCAGTT, R: GACCAGCATATCATTATAGACAAGC TTATGCACAAAGGAGTTGCA, R: ATGGTTGTATTCTTACTCCTCTCC TCCACCTGTAGACCTGGTAAA, R: AGTGCTGCGTCTTACAACCT
F, forward primer; R, reverse primer sequence;–2Product size in base pairs (bp);–3A, Adenosine; T, Thymidine; C, Cytidine; G, Guanosine.
891
SERUM DNA ALTERATIONS IN RENAL CELL TUMOURS TABLE II – DISTRIBUTION OF THE 53 RCCS ACCORDING TO THE TNM-SYSTEM AND THE “HEIDELBERG CLASSIFICATION” Tumour type
pT1/pT2
pT3a/b
G1
G2
G3
Total
Conventional RCC Papillary RCC Chromophobe RCC Bellini-Duct-RCC Non-classifiable RCC
28 4 1 1 4 38
15 — — — — 15
11 — 1 — 2 14
29 4 — 1 2 36
3 — — — — 3
43 4 1 1 4 53
FIGURE 1 – Frequencies of alterations in primary tumours and serum specimens detected at various chromosomal sites investigated with 9 markers in n⫽ 53 RCC (Group I).
determined by measuring the overall rate of identified alterations in the serum as well as measuring the proportion of identified alterations as a fraction of the rate of alterations observed in the tumour. Serum DNA alterations were only accepted if they were also detectable in the tumour DNA. The specificity of the method for detecting serum DNA alterations was investigated by applying the analysis on 20 healthy controls with 10 markers (D3S1560, D3S2450, D5S1720, D5S476, D8S261, D8S560, D9S925, D13S153, D17S799 and D17S1306). Statistical analysis To identify the possible association between serum DNA alterations and tumour stage and/or grade as well as with RCC subtypes the Student’s t test and the Mann Whitney test were used. A p value ⱕ 0.05 was regarded as significant. RESULTS
Histopathological staging and grading of tumours The distribution of the 53 primary renal cell carcinomas investigated according to the TNM system (UICC 19975) and the nuclear grading is shown in Table II. The tumourtypes were stratified according to the “Heidelberg Classification of renal cell tumours” by genetic analysis as possible with the here investigated chromosomal regions.4 At the time of surgery 8 primary tumours already demonstrated distant metastasis and 2 had regional lymph node disease. Of the 4 non-classifiable RCCs, 2 were histomorphologically described as clear cell, 1 as clear-cell eosinophile and 1 as papillary chromophile. Additionally 6 urothelial renal pelvis tumours (1 pTa,G1, 1 pT1,G1, 1 pT2,G2 and 3 pT3,G3) and 1 renal lymphoma were investigated for comparison reasons. Microsatellite analysis of tumour and serum DNA The frequencies of tumour and serum DNA alterations at the 8 chromosomal regions applying 9 microsatellite markers (Group I, n⫽ 53 RCC) are shown in Figure 1 and the frequency of alterations applying a total of 20 markers from 9 chromosomal regions (Group II, n⫽ 23 RCC) is shown in Figure 2. The highest incidence of alterations in the tumour DNA was ob-
served for chromosomes 3p and 5q with 78% and 49.1% in Group I as well as in Group II with 87% and 78.3%, respectively. Hereby, all conventional RCC displayed a chromosome 3p-LOH (see also Fig. 3). When applying 9 markers in Group I (n⫽ 53), the sensitivity in diagnosing tumour-specific serum DNA–alterations was 74% (39/ 53). By raising the number of markers to 20 in Group II (n⫽ 23), the method achieved a sensitivity of 87% (20/23). Regarding the 9 chromosomal regions individually, the highest incidence of alterations in the serum DNA was identified at chromosomes 3p in 24.5% and at 5q in 20.8% (Table III). Raising the number of markers in Group II (n⫽ 23 RCC) resulted in a markedly higher sensitivity in diagnosing serum DNA alterations. This rise in sensitivity, however, is statistically not significant when comparing the results of the 23 RCCs analysed with 9 and with 20 markers. The sensitivity for diagnosing serum-DNA alterations with the 9 markers of Group I accomplished a sensitivity of 80.4% (Fig. 2). Although the rise in sensitivity for the 23 RCCs was not statistically significant, adding more markers from chromosomes 3 and 5, as expected, elevated the sensitivity for diagnosing alterations at these loci. When applying 3 additional markers for chromosomes 3p and 5q serum-DNA alterations at chromosome 3p were diagnosed in 47.8% (vs. 24.5% in Group I) and at chromosome 5q in 65.2% (vs. 20.8% Group I). The frequencies of alterations at the other chromosomal regions, ranging between 4.3% at 9p and 26.1% at 8p and 17p and the matched incidences of alterations of Group II analysed with 9 and 20 markers, are demonstrated in Figures 1 and 2. Statistically, a weak association of tumours, displaying a serum alteration with advanced stage, was evaluated (p ⫽ 0.044; n⫽ 53 RCC), but there was no association with tumour nuclear grading. Furthermore, none of the genetical subgroups of RCC demonstrated an overrepresentation of serum DNA alterations (Fig. 4). Restricting the analysis to the 6 markers with the highest incidence of alterations in Group II (n⫽ 23; markers D3S2450, D3S1259, D5S476, D8S261, D13S153 and D17S799), the method still yielded a sensitivity of 69.6%.
892
VON KNOBLOCH ET AL.
FIGURE 2 – Frequencies of alterations detected in primary tumours and serum specimens at various chromosomal sites investigated with 9 markers in n⫽ 23 RCC (Group II). Arrows indicate rise in sensitivity when analysing with additional markers (D3S3666, D3S2408, D3S1259, D5S1480, D5S476, D5S818, D8S261, D8S560, D17S783, D17S1298 and D17S807).
FIGURE 3 – Results of fluorescent microsatellite analysis for RCC 9 (conventional RCC, pT3apN0M1,G3) and RCC 18 (chromophobe RCC, pT1,G1). Curves for normal DNA (N) from lymphocytes, serum DNA (S) and tumour DNA (T) are shown. RCC 9 demonstrates clear loss of allele-1 in tumour for markers D3S2450, D3S1269, D8S261 and D8S560 with a clear allelic loss in serum DNA for marker D3S2450 and allelic imbalance at loci D3S1269 and D8S261. Retention of constitutional heterozygosity was observed for loci D5S476 and D13S153 in tumour as well as in serum DNA. Compare RCC 18 with clear LOH of tumour DNA at loci D5S1480, D5S476 and D17S807. Serum DNA demonstrates clear specific allelic loss at D5S476 and specific allelic imbalance at loci D5S1480 and D17S807. Loci D3S1259, D8S560 and D13S153 are retained in tumour and in serum DNA.
Serum DNA allelic imbalance (AI) was only regarded as a tumour-specific alteration if the identical alteration was also present in the tumour DNA (Fig. 3, Table III). AI in the serum DNA only was therefore neglected and interpreted as an artefact without diagnostic relevance. By applying the 9 markers of Group I, the 7 non-RCCs displayed alterations in 5 of 7 cases (71%). As 6 of the 7 non-RCC tumours were transitional cell carcinomas (TCC) of the renal pelvis, the pattern of alterations resembled that of TCCs of the urinary bladder mainly displaying chromosome 5qand 9p-LOH. A loss of chromosome 3p was only observed in 2 cases. In the control group of 20 healthy individuals, serum DNA was amplifiable in every case. The average extracted DNA concentration in the probes of the healthy controls with 35.8 ng/ml was markedly lower than in the cancer patients though (see below). In 3 of 20 controls (15%), we identified PCR artefacts resembling an allelic imbalance. In relation to the total number of polymerase chain reactions performed (n⫽ 200), the rate of artefacts is 1.5%. Nonetheless, the specificity of the method remains to be 85%. This specificity is in accordance to the observation of discordant tumour unspecific allelic imbalance identified at a rate of 0 to 22.6%, depending on the microsatellite marker used (Table III). The average DNA concentration of all serum probes of the 53 RCCs was 94 ng/ml. Those probes displaying an alteration had a slightly higher average concentration of 96 ng/ml compared with those probes without an alteration identified (89 ng/ml).
DISCUSSION
With no clinically approved serological tumour markers available for RCC, which could ideally identify micrometastatic spread, the clinician has to rely on imaging diagnostics to morphologically detect tumour recurrence during follow-up after surgical therapy. We applied the microsatellite analysis with fluorescent-labelled markers to discover tumour-specific serum DNA alterations and to evaluate the applicability of the method for serum bound RCC diagnosis. By using 9 markers from the chromosomal regions 3p, 5q, 7p, 7q, 9p, 13q, 17p and 17q in 53 patients with RCC, the method achieved a sensitivity of 74% (39/53) in detecting tumour-specific serum DNA alterations. In a subgroup of n⫽ 23 RCC patients, we performed the analysis with additional 11 markers from the chromosomal regions 3p, 5q, 8p, 17p and 17q and achieved a sensitivity of 87% (20/23). This elevation in sensitivity was insignificant though, as the method rendered a sensitivity of 80.4% for the same 23 RCC patients with 9 markers of group I (Fig. 2). Out of the 20 markers tested, we were able to identify 6 highly sensitive markers. Restricting MSA to these 6 markers, the method still yielded a sensitivity of 69.6%. With 10 markers tested in 20 healthy controls, the method showed 3 false positive alterations, resulting in 85% specificity. Statistically, the identification of serum DNA alterations showed a weak association with advanced tumour stages (p ⫽ 0.044) but not with nuclear grading (p ⫽ 0.124).
893
SERUM DNA ALTERATIONS IN RENAL CELL TUMOURS TABLE III – INCIDENCE OF ALTERATIONS AT VARIOUS MARKER SITES FOR TUMOUR AND SERUM IN TOTAL COHORT WITH 9 MICROSATELLITE MARKERS Chromosome
Marker name
Product size1 (bp)
Repeat2
Alterations in: tumour3 (%)
Serum4 (%)
DNA amplified5 (%)
3p
D3S1560 D3S2450 D5S1720 D7S1797 D7S1807 D9S925 D13S153 D17S799 D17S1306
242 258 220 234 236 167 221 185 179
CA GATA GATA GATA GGAA GATA CA CA GGAA
67.9 56.6 49.1 22.6 26.4 26.4 32.1 24.5 26.4
5.7 (5.7) 24.5 (0) 20.8 (22.6) 18.9 (5.7) 17.0 (11.3) 5.7 (0) 17.0 (9.4) 11.3 (11.3) 17.0 (17.0)
64.2 79.2 90.6 96.2 92.5 39.6 90.6 90.6 94.3
5q 7p 7q 9p 13q 17p 17q
1 Microsatellite product size in base pairs.–2A, adenosine; T, thymidine; C, cytidine; G, guanosine.–3n ⫽ 53 RCCs (100%).–4n ⫽ 53 (100%) serum specimens; values given in parentheses ⫽ percentage of additional unspecific allelic imbalance (AI) identified (100% ⫽ n ⫽ 53).–5Percentage of serum DNA amplified with each marker.
For a variety of primary tumours (e.g., small- and non-small-cell lung cancers, head and neck cancers, breast cancer, renal cell carcinoma and bladder carcinoma), recent studies have proved molecular techniques capable of detecting even the smallest amounts of free-circulating genomic plasma DNA.1–3,6 –11 Applying only 4 markers for chromosome 3p, Goessl and co-workers2,3 were able to identify plasma-DNA alterations with a sensitivity of 63% in 40 patients with clear-cell RCC. As only nonpapillary or conventional RCC exhibit 3p-LOH, we were able to markedly raise the sensitivity of the method by first applying a larger number of markers and also extending the investigation to other chromosomal regions known to be involved in RCC pathogenesis.4,12,13 By investigating the chromosomal regions mentioned above, we were also capable of stratifying our cohort of renal cell tumours according to the “Heidelberg Classification”.4 Within this cohort of consecutive patients receiving surgical therapy for RCC at our institution from 1998 to 2000, “conventional” renal cell tumours represent the majority with 81% (43/53). The conventional or non-papillary RCC are characterised by chromosome 3p loss. The other 19% of tumours within our cohort therefore demand the use of markers from other chromosomal regions to allow serum diagnosis. Genetically stratifying the renal cell tumours did not lead to the identification of subtypes with markedly elevated rates of serum DNA alterations. Simultaneously performing MSA on primary tumours reduces the chance of regarding PCR artefacts as serum DNA allelic imbalance, thus as tumour-specific alterations. Owing to the low serum DNA concentrations of approximately 100 ng/ml, the highly sensitive MSA with laser product detection can mimic an AI. Recently, Coulet and co-workers9 warned from misinterpreting MSA results in plasma-DNA diagnosis. We ruled out artefacts with the simultaneous assay of the primary tumours and only regarded an AI as tumour specific if the serum DNA alteration matched the alteration in the tumour DNA. Applying the analysis on 20 healthy controls rendered a specificity of 85%. This specificity is in agreement with our observation of discordant serum DNA allelic imbalance seen in 0 to 22.6% of cases, depending on the microsatellite marker used (Table III). These discordant alterations were interpreted as PCR artefacts. As in our recent publication applying the microsatellite analysis on serum DNA alterations of bladder cancer patients, we strongly recommend the simultaneous investigation of the primary tumours to rule out unspecific allelic imbalance not being correlated to a tumour DNA alteration (Fig. 3). Goessl and co-workers2,3 demonstrated 2 discordant serum DNA-LOH findings with retention of heterozygosity in the tumour–DNA. They interpreted these findings as either being related to unidentified metastatic spread or an insufficient microdissection technique in preparation of the tumour DNA. In comparison with our results, the rate of discordant serum DNA-AI or-LOH seems low but may be explained by the fact that in the majority of cases Goessl et al.2,3 did not perform MSA of corresponding tumour DNA. Besides pointing out the possibility of PCR artefacts, Coulet on the other hand confirmed MSA of being highly sensitive allowing
FIGURE 4 – Incidences of serum DNA alterations stratified for various primary-tumour nuclear grades and stages. We did not identify an association with grade (p ⫽ 0.124) but a trend toward advanced stages (p ⫽ 0.044).
the identification of DNA alterations down to a ratio of tumour- / normal DNA of 0.5%.9 As unaltered normal genomic DNA is always present in the serum and plasma samples, complete loss of one allele in MSA results is almost impossible, which is why we refer to alterations in serum DNA as allelic imbalance (Fig. 3). Our results showed a weak statistical association (p ⫽ 0.044) between MSA and tumour stage but not with tumour nuclear grade for serum-RCC diagnosis (Fig. 4). In the majority of other investigations dealing with MSA of serum DNA in cancer patients as well as in our previous analysis of bladder cancer patients no association between detection of alterations in the serum DNA and tumour stage or grade was observed.1–3,6,7,9 –11 Only few investigators were able to identify an association between plasma-DNA detection and advanced disease and high grade in cancer patients.8,14,15 The association is plausible, considering that advanced stage tumours are usually accompanied by a high tumour load, elevating the ratio of tumour to normal cells and thereby raising the chance of DNA excretion into the circulation. Additionally, advanced tumour stages are commonly associated with a poorly differentiated tumour cell population and a high proliferation rate. High-grade cells are characterised by a high mitotic index with a high genetic instability again potentially elevating free tumour DNA levels in the serum. Serum samples with alterations were also accompanied by higher DNA concentrations (96 vs. 88 ng/ml). Investigating free circulating DNA in the serum of healthy controls and cancer patients, Leon and Shapiro16,17 identified significantly elevated amounts of free DNA in cancer patients compared with healthy patients or patients with benign diseases. They described the amount of free circulating DNA to be independent of tumour size or stage but to be markedly elevated in patients with metastatic disease.16,17 Looking at the chromosomal regions investigated, as expected 3p-LOH was observed with the highest frequency of 78 – 87%
894
VON KNOBLOCH ET AL.
owing to the fact that 81% (43/53) of RCCs within the cohort were classified as conventional RCC, all displaying chromosome 3p loss (Figs. 1, 2). The markers used to investigate chromosomal region 3p only identified serum DNA-LOH in 24.5– 47.8%, meaning 45 to 68% of tumour alterations were missed on serum DNA investigation. Another chromosomal region with a high incidence of alterations in the tumour DNA was 5q in 49 –78%, which is in accordance with rates reported in the literature.18 Alterations at 5q represent duplications and are suggested to be associated with progression of conventional RCC.18 For this chromosomal region, serum DNA-AI was observed in 20.8 – 65.2%. By increasing the number of microsatellite markers employed for analysis (Group I vs. Group II), we could raise the sensitivity more than for chromosomal region 3p. As previously suggested, we believe the marker characteristics to be responsible for differing sensitivity levels in detecting specific serum DNA alterations.11 Especially the PCR-product size may be of importance since serum DNA is certainly fragmented to a high extent rendering markers with smaller product size more sensitive (Table I). An exception seems to be tetranucleotide marker D9S925 with a product size of only 167 base pairs identifying serum DNA alterations in merely 5.7% of cases (n⫽ 53 RCC), although the primary tumours displayed chromosome-9p-AI in 26.4% of cases (Table I, Figs. 1, 2). A good example of a marker with a small product size (143 bp) and a high
sensitivity level for serum DNA alterations is D8S261 (8p22), detecting 8p-LOH in primary tumours in 34.8% and in 26.1% in serum. Out of the cohort of 20 markers tested, we were able to identify markers with a high sensitivity for microsatellite analysis in serum-bound cancer diagnosis. With only 6 of the most sensitive markers of Group II, we already achieved a sensitivity of 69.6% in detecting tumour-specific serum DNA alterations (markers D3S2450, D3S1259, D5S476, D8S261, D13S153 and D17S799). Recently, new molecular and non-molecular serological markers tested in combination with RCC as for example determination of telomerase activity, tumour growth receptor 1 (TGF 1) and tumour M2 pyruvate kinase have failed to prove acceptable specificity levels, therefore not allowing their use in routine clinical diagnosis.18 –20 With serological tumour detection rates of 74 – 87%, depending on the number of markers used, the MSA offers a promising method of early detecting micrometastatic tumour spread in RCC patients during follow-up after initial radical surgery. Owing to unspecific discordant serum DNA allelic imbalance in up to 20%, we recommend the simultaneous investigation of primary tumour DNA, which additionally allows a genetic characterisation with potential prognostic relevance for the underlying RCC disease.
REFERENCES
1.
Chen XQ, Stroun M, Magnenat JL, et al. Microsatellite alterations in plasma DNA of small cell lung cancer patients. Nat Med 1996;2: 1033–5. 2. Goessl C, Heicappell R, Munker R, et al. Microsatellite analysis of plasma DNA from patients with clear cell renal carcinoma. Cancer Res 1998;58:4728 –32. 3. Goessl C. Laser-fluorescence microsatellite analysis and new results in microsatellite analysis of plasma/serum DNA of cancer patients. Ann N Y Acad Sci 2000;906:63– 6. 4. Kovacs G, Akhtar M, Beckwith BJ, et al. The Heidelberg classification of renal cell tumours. J Pathol 1997;183:131–3. 5. Sobin LH, Wittekind C. TNM classification of malignant tumors, 5th ed. New York: Wiley-Liss, 1997. 6. Nawroz H, Koch W, Anker P, Stroun M, Sidransky D. Microsatellite alterations in serum DNA of head and neck cancer patients. Nat med 1996;2(9):1035–1037. 7. Sozzi G, Musso K, Ratcliffe C, et al. Detection of microsatellite alterations in plasma DNA of non-small cell lung cancer patients: a prospect for early diagnosis. Clin Cancer Res 1999;5:2689 –92. 8. Silva JM, Dominguez G, Garcia JM, et al. Presence of tumor DNA in plasma of breast cancer patients: clinicopathological correlations. Cancer Res 1999;59:3251– 6. 9. Coulet F, Blons H, Cabelguenne A, et al. Detection of plasma tumor DNA in head and neck squamous cell carcinoma by microsatellite typing and p53 mutation analysis. Cancer Res 2000;60:707–11. 10. Shaw JA, Smith BM, Walsh T, et al. Microsatellite alterations plasma DNA of primary breast cancer patients. Clin Cancer Res 2000;6: 1119 –24. 11. von Knobloch R, Hegele A, Brandt H, et al. Serum DNA and urine DNA alterations of urinary transitional cell bladder carcinoma de-
12. 13. 14. 15. 16. 17. 18. 19. 20. 21.
tected by fluorescent microsatellite analysis. Int J Cancer 2001;94: 67–72. Bugert P, Kovacs G. Molecular differential diagnosis of renal cell carcinomas by microsatellite analysis. Am J Pathol 1996;149:2081– 8. Steiner G, Sidransky D. Molecular differential diagnosis of renal carcinoma: from microscopes to microsatellites. Am J Pathol 1996; 149:1791–5. Lo YM, Chan LY, Lo KW, et al. Quantitative analysis of cell-free Epstein-Barr virus DNA in plasma of patients with nasopharyngeal carcinoma. Cancer Res 1999;59:1188 –91. Kawakami K, Brabender J, Lord RV, et al. Hypermethylated APC DNA in plasma and prognosis of patients with esophageal adenocarcinoma. J Natl Cancer Inst 2000;92:1805–11. Leon SA, Shapiro B, Sklaroff DM, et al. Free DNA in the serum of cancer patients and the effect of therapy. Cancer Res 1977;37: 646 –50. Shapiro B, Chakrabarty M, Cohn EM, et al. Determination of circulating DNA levels in patients with benign or malignant gastrointestinal disease. Cancer 1983;51:2116 –20. Bugert P, von Knobloch R, Kovacs G. Duplication of two distinct regions on chromosome 5q in non-papillary renal-cell carcinomas. Int J Cancer 1998;76:337– 40. M¨uller M, Heicappell R, Krause H, et al. Telomerase activity in malignant and benign renal tumors. Eur Urol 1999;35:249 –55. Wunderlich H, Steiner T, Kosmehl H, et al. Increased transforming growth factor beta1 plasma level in patients with renal cell carcinoma: a tumor-specific marker? Urol Int 1998;60:205–7. Roigas J, Schulze G, Raytarowski S, et al. Tumor M2 pyruvate kinase in renal cell carcinoma: studies of plasma in patients. Urologe A 2000;39:554 – 6.