Circulating mitochondrial DNA in the serum of patients with testicular germ cell cancer as a novel noninvasive diagnostic biomarker


Jörg Ellinger, Klinik und Poliklinik für Urologie, Universitätsklinikum Bonn, Sigmund–Freud–Strasse 25, 53105 Bonn, Germany.



To analyse the diagnostic and prognostic value of cell-free mitochondrial (mt)DNA in patients with testicular cancer, as increased levels of cell-free circulating mtDNA have been reported in patients with cancer.


In all, 74 patients with testicular cancer (seminoma in 39, nonseminoma in 35) and 35 healthy individuals were included in the study. Circulating DNA was isolated from 1 mL of serum. A quantitative real-time polymerase chain reaction was used to analyse the levels of a 79-bp (mtDNA-79) and 220 bp (mtDNA-220) fragment of the mitochondrial specific 16S-RNA. The mtDNA integrity was expressed as the ratio of mtDNA-220 to mtDNA-79.


mtDNA-79 and mtDNA-220 levels were significantly (P < 0.001) greater in patients with testicular cancer than in healthy individuals. The mtDNA integrity was similar in patients and healthy controls (P = 0.435). Receiver operator characteristic curve analysis showed that cell-free mtDNA (mtDNA-79) levels distinguished, with a sensitivity of 59.5% and a specificity of 94.3%, between patients and healthy individuals (area under curve, 0.787). Also, mtDNA-79 levels could be used to distinguish between patients (31) with conventional markers (α-fetoprotein, human chorionic gonadotrophin, placental alkaline phosphatase and lactate dehydrogenase) within normal ranges and healthy individuals, with a sensitivity of 64.5% and specificity of 91.4% (area under curve 0.797). Cell-free mtDNA levels were not correlated with any clinicopathological variable (pT stage, lymph node invasion, vascular invasion, clinical stage, International Germ Cell Cancer Collaborative Group classification, tumour markers; all P > 0.05).


Cell-free mtDNA levels are greater in patients with testicular cancer and might provide valuable information for managing patients with testicular anomalies, especially those with normal levels of established tumour markers.




area under curve


genomic DNA


lactate dehydrogenase


mitochondrial DNA


placental alkaline phosphatase


receiver operating characteristic


International Germ Cell Cancer Collaborative Group.


Testicular germ cell cancer represents ≈1% of male neoplasms [1] and 5% of all urological tumours [1]. The incidence of testicular cancer has almost doubled in Europe during the last 40 years [2]. Most often, testicular cancer appears as a painless intrascrotal mass. The analysis of α-fetoprotein (AFP), hCG and lactate dehydrogenase (LDH) is mandatory in patients with suspected testicular cancer [3]. However, normal levels of these markers do not exclude testicular cancer; increased marker levels are only detected in ≈60% of patients with testicular cancer [4], and therefore each patient with a scrotal mass must undergo inguinal exploration [3,5,6]. Tumour markers are also measured during follow-up visits, but likewise the sensitivity is limited and tumour markers are not abnormal in ≈40% of patients with cancer recurrence [4]. Therefore, the development of additional markers would facilitate the management of patients with testicular cancer.

Circulating DNA is a noninvasive tumour marker in patients with malignancy. In 1977, increased levels of cell-free DNA were reported in the plasma of patients with cancer [7]. Despite this long history, the origin of circulating DNA remains unknown; only a small proportion seems to originate from cancerous cells. Several studies have reported greater levels of circulating genomic (g)DNA in patients with various urological malignancies than in healthy individuals and in patients with benign disease [8–10]. In addition to quantitative differences, the fragmentation pattern of circulating DNA is changed in patients with cancer; large DNA fragments (i.e. >300 bp) were increased in patients with breast [11,12], colon [13] and ovarian cancer [12,13], whereas in patients with prostate [8] and bladder cancer [10] there was an increase of short DNA fragments. An increase in short DNA fragments indicates an apoptotic origin, whereas the increase in large DNA fragments indicates a predominantly necrotic breakdown in patients with cancer [14].

Most recent studies focused on the analysis of circulating gDNA. The precise quantification of a few nanograms of circulating DNA might be difficult even using a sensitive method like real-time PCR. A single cell contains several hundred copies of mitochondrial (mt)DNA, as opposed to two copies of gDNA. Therefore, the analysis of mtDNA enables a more precise analysis than is possible with gDNA [15]. A recent study compared mtDNA and gDNA levels in patients with prostate cancer. Interestingly, mtDNA levels were not correlated with gDNA levels, indicating different a pathophysiological mechanism leading to the increase in these DNA types [16]. Furthermore, it was shown that high mtDNA levels were associated with a poor prognosis in patients with advanced prostate cancer [16]. Also, mtDNA levels were associated with biochemical recurrence in patients undergoing radical prostatectomy for clinically localized prostate cancer [17].

To date, circulating mtDNA has not been analysed in patients with testicular cancer. The aim of the present study was to analyse circulating mtDNA levels and the mtDNA fragmentation pattern in patients with testicular cancer and in healthy individuals.


We analysed 74 consecutive patients with suspected testicular cancer undergoing inguinal exploration at the Department of Urology, University of Bonn, Germany, between 1996 and 2004, and who agreed to tissue and serum banking. Blood samples were withdrawn before therapy (i.e. radical orchidectomy). The histology was seminoma germ cell cancer in 39 patients and nonseminoma in 35 (embryonal carcinoma in three, yolk sac tumour in one, immature teratoma in two, and nonseminoma with mixed components in 29). We also analysed 35 healthy male controls; these volunteers were recruited from employees at the University Hospital Bonn and had no malignancy or other severe disease before or at the time of blood withdrawal. The clinicopathological variables are shown in Table 1 for. Blood samples were collected in serum S-Monovette Gel tubes (Sarstedt, Nürnbrecht, Germany). Clotting of serum samples was allowed for ≥60 min before centrifugation (1800g, 10 min) and supernatants were stored at −80 °C. All participants had given written informed consent for the collection and analysis of serum samples. Cell-free DNA was isolated from 1 mL of serum using the ChargeSwitch gDNA Kit (Invitrogen, Paisley, Scotland).

Table 1.  Clinicopathological variables in the three groups
VariableSeminomaNonseminomaHealthy controls
  1. na, not available.

No. of participants393535
Age, years
Elevated serum tumour markers, n (%)
 AFP 3 (8)15 (43)na
 hCG 8 (21)24 (69)na
 PLAP10 (26) 3 (9)na
 LDH 4 (10)12 (34)na
Markers normal range22 (56) 9 (26) 
Pathological tumour stage
 pT131 (80)26 (74) 
 pT2 6 (15) 7 (20) 
 pT3 2 (5) 1 (3) 
Clinical stage
 I28 (72)18 (51) 
 IIA-C 7 (18)13 (37) 
 IIIA-C 3 (8) 4 (11) 
 Not available 1 (3) 0 
IGCCCG classification
 Not classified36 (90)27 (77) 
 Good prognosis 3 (8) 6 (17) 
 Intermediate prognosis 1 (3) 1 (3) 
 Poor prognosis 0 1 (3) 

DNA from cells undergoing apoptosis is truncated to internucleosomal fragments of ≈200 bp; by contrast, DNA derived from necrotic cells is distinctly larger (>10 kbp). Two primer sets specific for mitochondrial 16 s-rRNA were designed to quantify and characterize the fragmentation pattern of circulating mtDNA. The mtDNA-79 primer set amplified a 79-bp fragment that corresponds to total mtDNA and includes DNA truncated by apoptosis. The mtDNA- 230 primer set amplified a 230-bp fragment that indicates mtDNA derived from necrotic origin. The primer sequences were published previously [17]. The fragmentation pattern was expressed as mtDNA integrity (the ratio of mtDNA-230 to mtDNA-79).

Quantitative real-time PCR (qPCR) experiments were conducted essentially as described previously [17]. Briefly, reactions were run in triplicate on an ABI Prism 7900HT (Applied Biosystems, Foster City, CA, USA). Each 10 µL of PCR reaction consisted of 1 × SYBRGreenER Mix (Invitrogen), 400 nm forward/reverse primer and 0.7 µL of DNA sample. Cycle conditions were 90 °C for 10 min, followed by 40 cycles at 95 °C for 15 s and 60 °C for 60 s. The specificity of the PCR products was confirmed by melting-curve analysis. Each run included serial dilutions of an external standard and water blanks.

A 490-bp DNA fragment that covered both mtDNA fragments (primer sequences: forward 5′-GGG-ATA-ACA-GCG-CAA-TCC-TA-3′ and reverse 5′-ATG-TTG-GGA-CCT- TTG-CGT-AG-3′) was amplified to create a standard curve. The PCR product was purified using the Qiagen PCR purification Kit (Hilden, Germany); the DNA concentrations were determined using the PicoGreen dsDNA Quantification Kit (Molecular Probes, Eugene, OR, USA). The copy number of the 490 bp mtDNA fragment was calculated according to the protocol ‘Creating Standard Curves with Genomic DNA or Plasmid DNA Templates for Use in Quantitative PCR’ (Applied Biosystems). Samples were analysed without prior knowledge of the specimen identity.

Differences in mtDNA levels in the serum of patients with testicular cancer and healthy individuals were analysed using the Mann–Whitney test. Correlations between clinicopathological variables were assessed using the Mann–Whitney and Kruskal–Wallis tests, as appropriate. Follow-up information was available for 53 patients; the mean (median, range) follow-up was 82.5 (87, 2–130) months. Disease recurrence was defined as relapse after radical orchidectomy and adjuvant chemotherapy, and/or retroperitoneal lymph node dissection. Among these patients, only one had a disease recurrence and one died from pneumonia; thus we thus did analyse survival.


The median level of large and short mtDNA fragments were greater in patients than in healthy individuals (P < 0.001). The median (10–90th percentile) mtDNA-79 levels were 2.04 (0.25–5.67) × 106 copies/mL in patients and 0.47 (0.16–1.59) × 106 copies/mL in healthy individuals. The median levels of mtDNA-220 were lower in patients, at 1.54 (0–4.77) × 106 copies/mL, and in the control group, at 0.32 (0.01–1.76) × 106 copies/mL, indicating truncation of circulating mtDNA. The fragmentation pattern was similar (P = 0.435) in healthy individuals (median mtDNA integrity 0.562, 0.089–1.596) and patients (0.760, 0–1.437). The median mtDNA-79 and mtDNA-220 levels were somewhat higher in nonseminoma patients, but the differences were not significant (P = 0.944 and 0.905, respectively; Fig. 1 and Table 2).

Figure 1.

Short (A) and large (B) mtDNA levels were greater in patients with testicular cancer and allowed an accurate discrimination from healthy individuals. mtDNA integrity (C) was similar in patients and healthy controls.

Table 2.  Diagnostic information
Group comparisonThreshold, copies/mLSensitivity, %Specificity, %AUC (95% CI)
Seminoma vs controls1.9 × 10648.797.10.769 (0.661–0.877)
Nonseminoma vs controls1.6 × 10668.694.30.807 (0.695–0.918)
Patients vs controls1.6 × 10659.594.30.787 (0.702–0.872)
Marker-negative patients vs controls1.6 × 10664.591.40.797 (0.683–0.912)
Seminoma vs controls7.6 × 10564.177.10.707 (0.585–0.830)
Nonseminoma vs controls7.0 × 10577.177.10.787 (0.675–0.899)
Patients vs controls7.0 × 10570.277.10.745 (0.654–0.836)
Marker-negative patients vs controls6.1 × 10577.471.40.757 (0.629–0.884)
mtDNA integrity
Seminoma vs controls0.78148.767.70.505 (0.369–0.642)
Nonseminoma vs controls0.47980.044.10.593 (0.457–0.729)
Cancer patients vs controls0.77150.067.60.547 (0.426–0.668)
Marker-negative patients vs controls0.78361.367.60.584 (0.441–0.728)

Receiver operator characteristic (ROC) curve analysis showed that circulating mtDNA-79 fragment levels distinguished accurately between patients and healthy controls; the sensitivity was 59.5% and the specificity was 94.3% (area under curve, AUC, 0.787). The diagnostic information was similar in patients with seminoma and nonseminoma germ- cell cancer (AUC 0.769 and 0.807). Most interestingly, patients with serum tumour markers within the normal ranges (22 with seminoma and nine with nonseminoma) were also discriminated, with a sensitivity of 64.5% and a specificity of 91.4%, from healthy individuals. The diagnostic information of mtDNA-220 fragments was somewhat lower (AUC 0.745). mtDNA integrity was not useful for diagnostic purposes (AUC 0.584; Fig. 1 and Table 2.

AFP, hCG, placental alkaline phosphatase (PLAP) and LDH levels were not available for healthy individuals. Assuming normal levels of these conventional markers, we combined the analysis of mtDNA levels and conventional markers. The combination diagnosed testicular cancer in the patients with a diagnostic sensitivity of 85.1% and a specificity of 91.4% (AUC 0.883, 95% CI 0.811–0.955).

There was no significant correlation of cell-free mtDNA fragment levels (P > 0.05) with pT stage, lymph node metastasis, vascular invasion, clinical stage, the International Germ Cell Cancer Collaborative Group (IGCCCG) classification or serum tumour markers (AFP, hCG, LDH, PLAP; Table 3). Follow-up information was available for 51 patients; among these, one had a recurrent relapse with the teratoma component 2, 11, 21 and 45 months after the initial diagnosis; his circulating mtDNA levels were inconspicuous (mtDNA-79 1.22 × 106 copies/mL; mtDNA-220 2.31 × 105 copies/mL) before orchidectomy. Another patient died from pneumonia 45 months after diagnosis with no signs of recurrence.

Table 3.  Correlation with clinicopathological variable
mtDNA 70 bpmtDNA 220 bpmtDNA integrity
  • *


  • Mann–Whitney-test.

Lymph node involvement*0.2040.2890.533
Distant metastasis0.7640.8000.593
Clinical stage*0.5440.5890.770
IGCCG classification*0.5730.3030.537
AFP elevated0.7490.8330.706
hCG elevated0.6080.4100.509
PLAP elevated0.1380.5970.215
LDH elevated0.1380.5970.720


The analysis of AFP, hCG, PLAP and LDH is used routinely in the diagnosis and surveillance of patients with testicular germ cell cancer. However, these markers are within normal ranges in ≈40% of patients with testicular cancer at the time of diagnosis and/or at the time of cancer recurrence [4]. The development of novel biomarkers would facilitate the management of these patients. Circulating cell-free gDNA was shown to be a universal tumour marker in the diagnosis and prognosis of patients with various malignancies [8–10,12,18,19]. Recent studies showed that mtDNA fragments might also be useful in the diagnosis [16] and prognosis [16,17] of patients with prostate cancer. To date, mtDNA levels have not been analysed in patients with testicular cancer.

In the present study there was a significant (P < 0.001) increase in short (79 bp) and large (220 bp) mtDNA fragments in patients with seminoma and nonseminoma testicular germ cell cancer. The median mtDNA-79 levels were four times, and mtDNA-220 levels were five times higher in patients than in controls. The analysis of short mtDNA fragments was highly specific (94%) and moderately sensitive (60%). Importantly, 65% of patients with normal conventional tumour markers had increased mtDNA levels. The analysis of a single conventional tumour marker was far less sensitive (AFP 24%, hCG 41%, PLAP 13%, LDH 16%) than mtDNA levels. Interestingly, the combined analysis of mtDNA levels and conventional tumour markers had a sensitivity of 85%. We therefore assume that the analysis of cell-free DNA levels is especially helpful for managing patients with marker-negative testicular cancer.

The analysis of circulating mtDNA levels might also provide prognostic information. The disease-specific survival was shorter in patients with advanced prostate cancer and high levels of mtDNA [16]. The likelihood of biochemical recurrence after radical prostatectomy for localized prostate cancer was greater in patients with high preoperative mtDNA levels [17]. In the present study, mtDNA levels were not correlated with any clinicopathological variable (i.e. pT stage, lymph node metastasis, vascular invasion, clinical stage, the IGCCCG classification or serum tumour markers). Unfortunately, survival analysis is not feasible because only one patient had a disease recurrence.

A limitation to the present study was the lack of patients with inflammatory disease (i.e. orchitis) or stromal tumours (i.e. Leydig cell tumour and Sertoli cell tumour). It is possible that these conditions are also associated with increased levels of mtDNA. However, a recent study reported distinctly lower levels of circulating DNA in patients with various non-malignant conditions (i.e. infectious disease, cardiovascular disease, diabetes, autoimmune disease) than in patients with different malignancies (i.e. ovarian cancer, cervical cancer or lung cancer) [19]. It is therefore possible that mtDNA could also be useful to distinguish testicular cancer from other nonmalignant testicular diseases.


Our work was supported by research grants from the North–Rhine Westphalian Association of Urology to Jörg Ellinger and the Reinhard-Nagel foundation to Patrick J. Bastian.


None declared.