Can germ cell neoplasia in situ be diagnosed by measuring serum levels of microRNA371a-3p?

As GCNis may present clinically with at least four different features (Berney et al. 2016), we made the following stratifications regarding the patient population:

(Group 1) GCNis in a solitary testis, no contralateral testis is present at the time of examination, (Group 2) GCNis in one testis and a healthy testis is present on the contralateral side, (Group 3) both testes afflicted with GCNis, and (Group 4) one GCNis-afflicted testis is accompanied by a GCT-bearing contralateral one. As any GCT tissue (primary tumour or metastasis) represents a significant source of miR production, we recruited patients exclusively with GCNis only for the present study. Thus, patients with any concurrent GCT tissue at the time of examination (group 4, vide supra) were excluded from this study to avoid confounding of serum levels by invasive malignancy.

Cubital vein serum was obtained from 27 patients aged 35.3 ± 8.8 years (Table 1). Seventeen patients were recruited prospectively from Albertinen-Krankenhaus Hamburg during 2012–2016, 15 of whom had contralateral biopsy in the presence of unilateral GCT, and 2 had bilateral biopsies for evaluation of testicular microlithiasis. Ten patients were enrolled from a serum biobank of the University of Muenster, Department of Andrology, all of the latter had been examined for male infertility during 2005-2016. Six of the patients had been briefly reported earlier (Spiekermann et al. 2015). According to the stratification listed above, groups 1-3 consisted of 11, 14, and 2 patients. All of the 11 patients of group 1 had undergone previous orchiectomy for a contralateral GCT. The interval between orchiectomy for GCT and acquisition of the serum sample ranged from 2 to 31 days (Table 1). According to published experience regarding the decay of miR levels after orchiectomy, clearance of tumour-induced miR expression was expected to be completed after these intervals (Spiekermann et al. 2015). All of the 11 group 1 patients had clinical stage 1 disease, i.e. no metastases were detected and all had normal classical tumour markers. Eleven patients (four of group 1, five of group 2 and two of group 3) had repeat measurements of serum levels of microRNAs after treatment of GCNis. Treatment comprised local radiotherapy to the testis with 13–20 Gy in three cases, chemotherapy with carboplatin in one case, and orchiectomy of the afflicted testis in seven cases. Testicular vein serum was additionally obtained from two patients of group 2 during surgery.

Serum samples of 20 healthy men or individuals with non-malignant testicular disease aged 37.5 ± 10.8 years served as controls, all of whom had been reported previously (Dieckmann et al. 2017).

In all cases, except for the two testicular vein samples, serum samples were derived from cubital vein blood aspiration. Serum was obtained after centrifugation and it was then stored deep frozen at −80 °C until processing.

Four archival orchiectomy specimens of patients with testicular GCTs were analysed by immunohistochemistry and in situ hybridisation (ISH) to look for the presence of miR-371a-3p in GCNis cells neighbouring the invasive tumours. The patients were aged 28.0 ± 0.7 years. Histologically, the GCTs comprised mixed nonseminomatous tumours in two cases, one pure seminoma, and one pure embryonal carcinoma, respectively.

All patients enrolled for the various parts of this study had given informed consent to the examination. Ethical approval of the study was provided by Aerztekammer Bremen (reference No 301, 2011).

For isolation of miRNA, 200 µl of serum was processed with the miRNeasy mini kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. Then 6 µl of RNA was reverse transcribed using the TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems, Darmstadt, Germany) and preamplified using Real-Time Ready cDNA Master (Roche Diagnostics, Mannheim, Germany). TaqMan assays (Applied Biosystems, Darmstadt, Germany) for miR-371a-3p (assay ID 002124), miR-367-3p (assay ID 000555) and the endogenous control miR-93-5p (assay ID 000432) were used in a 1:100 dilution for the preamplification. The qPCR was performed on a 7500 Fast Real-Time PCR System (Applied Biosystems, Damstadt, Germany) using FAST Start Universal Probe Master (Roche Diagnostics, Mannheim, Germany) and the undiluted TaqMan Assays. Finally, the relative quantity (RQ) was calculated according to the 2^−∆∆CT method (Livak and Schmittgen 2001). As cut-off value for miR-371a-3p serum levels, a RQ = 5 was used to differentiate between GCNis and controls corresponding to the method employed previously (Dieckmann et al. 2017).

For morphological identification of GCNis, four orchiectomy specimens with invasive GCTs were randomly selected from the pathology archive. Of the FFPE-blocks, sections of 5 µm were obtained from tumour-free parts of the specimen. Staining with haematoxylin and eosin with standard histological techniques were used to confirm tumour-free areas and to identify tubules containing GCNis cells. Oct4 staining (Diagnostic BioSystems, Pleasanton, CA, USA) was then conducted according to institutional standard operating procedures (Jones et al. 2004; Berney et al. 2015) to safely ascertain GCNis cells.

Sections with clear immunohistochemical evidence of GCNis were then processed with in situ hybridisation with a miRCURY LNA probe (Exiqon, Vedbaek, Denmark; probe ID 38555-15) specific for miR-371a-3p. The protocol was performed according to the manufacturer’s instructions using a proteinase K concentration of 15 µg/ml, a hybridisation temperature of 51 °C and a probe concentration of 80 nM. Microscopic evaluations were performed on an Axioskop 2 plus microscope (Zeiss, Göttingen, Germany) at 100× to 200× magnifications. Histological findings were documented using the AxioCam HRc digital camera (Zeiss, Göttingen, Germany) and then edited with AxioVision Software v.4.8 (Zeiss, Göttingen, Germany). Presence of miR-371a-3p within GCNis cells was defined by distinct blue staining of intratubular cells, and accordingly, only these cells were considered miR-371a-3p positive. Only the presence or absence of the miR-371a-3p in the specimen was evaluated, no quantification was attempted.

The Mann–Whitney U test was used to compare median miRNA expressions among the various subgroups. The Wilcoxon signed-rank test was used for comparison of dependent subgroups. Categorical data was analysed with the Fisher’s exact test. Significance was assumed at p < 0.05. Exact 95% confidence intervals of proportions were calculated.

Table 2 shows the individual results of measurements of miR-371a-3p and miR-367-3p of all of the 27 patients. Serum levels of miR-371a-3p were elevated in 14 of 27 patients (51.9%; exact 95% confidence interval 31.9–71.3%). The relative frequencies of miR-371a-3p expression above the cut-off value of the various GCNis subgroups and controls are presented in Table 3. All subgroups were significantly different from controls (Table 3) but differences among the groups were not significant, statistically (all p > 0.1).

Figure 1a shows the median RQ values of the patient groups. In the entire GCNis group (all patients), the median miR-371a-3p expression was 5.2 (interquartile range, IQR = 35.8). The median serum level of miR-367-3p was 0.0 (IQR = 0.9) in the entire cohort.

The mean expression levels of miR-371a-3p and miR-367-3p of the GCNis subgroups are shown in Table 4. Noteworthily, one of the two patients with GCNis in both testes (group 3) had the highest miR-371a-3p expression of all patients with RQ = 237.18. The second patient with bilateral GCNis also had an expression (RQ = 35.8) above the 3rd quartile of group 2 and 3. Each of the GCNis group had significantly higher median expressions of miR-371a-3p than controls (Table 4), whereas median miR-367-3p expressions of the GCNis groups were not significantly different from controls.

Figures 1b and 2 show the results of repeat measurements after treatment of GCNis in 11 patients. After treatment, the median miR-371a-3p expression decreased to 0.0 (IQR = 6.0) which is significantly lower than the pre-therapeutic median value (p = 0.047). The median miR-367-3p expression was 0.0 (IQR = 0.6) after treatment which is not significantly different from the pre-therapeutic value. Figure 2 shows individual measurements of miR-371a-3p and miR-367-3p levels before and after treatment of GCNis. Apparently, all elevated serum levels of miR-371a-3p dropped to the normal range after treatment.

Levels of miR-371a-3p in testicular vein serum were manifold higher than in the corresponding cubital vein serum: patient 20: RQ = 57.3 vs. 0.0 and patient 23: 295.4 vs. 1.61 (Table 2). The testicular vein miR levels found in GCNis are considerably higher than the median level observed in testicular vein blood of healthy males (RQ = 4.3), as reported previously (Dieckmann et al. 2016).

Two of the four cases examined had morphological evidence of the presence of miR-371a-3p in GCNis cells. Figure 3 shows the result in case 1. The blue-stained intratubular cells represent GCNis expressing miR-371a-3p intracellularly (Fig. 3a, b). Identification of these cells as GCNis was achieved by additional staining with Oct4 (Fig. 3c, d). When identical areas of the sections revealed to be positive for both, Oct4 and the ISH probe, the presence of miR-371a-3p in GCNis was assumed.