The most common pineal region neoplasms are GCT, and they are divided into pure germinoma (55%-65%) and NGGCT (35%- 45%) (Frappaz et al. 2022). Pure germinomas, which often occur in adolescent male patients, account for 40% of pineal region neoplasms (Smirniotopoulos et al. 1992).
Pure germinomas do not secrete appreciable tumor markers and thus a tumor biopsy is typically required for diagnosis (Echevarría et al. 2008). Such procedures carry risks of morbidity due to the difficulties in surgical access to the pineal region (Murray et al. 2016). Finding a blood and / or CSF biomarker for pure germinomas would offer many benefits and circulating microRNAs hold great clinical promise in this area (Murray et al. 2020). It has been previously shown that pure germinomas and NGGCTs (intra & extra-cranial) are characterized by high expression levels of all 8 microRNA members of the miR-371 ~ 373 and miR-302/367 clusters in serum and CSF, regardless of patient age, tumor site, or histological subtype (Palmer et al. 2010a).
This biomarker study presents the results of a miRNA profile expression analysis of pineal region tumors aiming to identify potential biomarkers in plasma that can be used to diagnose pineal body pure germinoma, and differentiate them from other pineal region tumors without the need for a tissue biopsy.
CSF is a suitable repository of clinical biomarkers for CNS tumors, and an increasing number of studies have reported that CSF-derived biomarkers are more abundant than those in the peripheral blood and other sources (Xiao et al. 2020). Qu et al. have also shown that miR-21 level in CSF enabled the identification of glioma patients with higher sensitivity and specificity in comparison with the plasma/serum miR-21 level (Qu et al. 2016). Due to unavailability of CSF samples at diagnosis, analysis of miRNAs in CSF was not included in our study.
Several studies have reported both plasma and serum as acceptable specimens for circulating miRNA analysis (Kroh et al. 2010). Although some studies have shown no or minimal difference between serum and plasma (Wang K et al. 2012), others noted that plasma samples provide a higher recovery of miRNA compared to serum (McDonald et al. 2011). For that reason, we chose plasma over serum for the conduction of our analysis.
Owing to the low reproducibility in the quantification of circulating miRNAs in plasma compared to CSF (Larrea et al. 2016), we used a permissive approach in our analysis using a less stringent p-value (p < 0.5), where we identified eight miRNAs that were differentially expressed when comparing the pure germinoma group with the other group of pineal region tumors. We then applied a more stringent p-value (p < 0.05) in order to discover a more specific expression profile, where only three of the eight miRNAs showed significant differential expression.
While the results presented need to be confirmed and expanded, the finding that the three combined miRNAs demonstrated good discriminatory power (AUC 90.7%, p < 0.001) in a subset of patients (sensitivity of 25%) with a pineal-based pure germinoma (100% specificity with a false discovery rate (FDR) of 0%) suggests that a subset of patients might be able to avoid a costly and potentially morbid surgical procedure.
In our samples, we were unable to detect the previously documented miR-371 ~ 373 and miR-302/367 clusters (Murray et al. 2016). This is in contrast to a recent study by Schönberger et al, where they demonstrated the suitability of miR-371a, miR-372, miR-367 and miR-302d in serum and CSF for diagnosis of CNS GCT, particularly in biomarker-negative pure germinoma (Schönberger et al. 2022). This discrepancy can be attributed to the difference in methodology between the two studies. While we opted to use unselected material as the starting point, they pre-amplified RT-qPCR involving only these specific miRNAs. This makes it difficult to discover if other miRNAs from different clusters would have been differentially expressed in their cohort. Moreover, they used healthy subjects as controls unlike our study where samples from patients with other pineal region tumors were used for comparison. This is an important difference as the primary goal of this work was to investigate if miRNAs can be used to separate patients with pure germinomas from all of the other tumors of the pineal region in pediatric patients as a method to avoid unnecessary intervention.
Another possible explanation of the low expression levels of these miRNA clusters in our samples is the possibility that these clusters were not efficiently labeled, sequenced or amplified, and thus lost to analysis. Several studies have shown that cellular contamination and hemolysis of samples can be a major cause of variation in miRNA levels not related to any biological difference (Blondal et al. 2013).
A review of the miRNAs from the pure germinoma group identified a number of functions associated with signal transduction, cell cycle, development and morphogenesis. miR-143-3p, for example, is known to regulate pathways that mediate cellular processes important in oncogenesis and malignant progression including MAPK, mTOR, and PI3K-Akt signaling pathways in addition to others (Kanehisa et al. 2010). The PI3K/AKT and MAPK pathways have been implicated in the pathogenesis of pure germinoma as they are both present in approximately 83% of tumor cells (Schulte et al. 2016). MAPK pathway alterations are more frequent in pure germinomas than NGGCT, and have a tendency to correlate with a better outcome compared to PI3K pathway mutations. Upregulation of the MAPK pathway by somatic point mutations represents the dominant genetic alteration in pure germinomas (64.3% of cases) (Ichimura et al. 2016). Further investigation on the possible functional significance of the miR-143-3p, and the other miRNAs is needed.
Our study has a number of important limitations. The main limitation is linked to the small number of cases, which is expected with such a rare tumor. To address this issue, multicenter or consortium-based studies will be needed. Since only a blood sample is required, such a study would be feasible and is currently under consideration. Another limitation is the absence of CSF samples, which may have a much better signal to noise ratio, as has been identified in the detection of BRAF and H3K27M mutations (Tan et al. 2021). Finally, due to the rarity of this patient population, a long duration of sample collection was required, which could impact the quality of the starting material from blood.
In conclusion, the results of the present study demonstrated the potential of a three-plasma miRNA signature as a diagnostic biomarker panel for pineal body pure germinomas. While currently only of potential value to a small proportion of pineal-based tumor cases, this approach opens up the opportunity to expand these results and test them prospectively in the hopes that avoidance of surgical biopsy can be achieved for patients with tumors in this critical area of the brain. Coupling plasma/serum with CSF samples and validation with Quantitative real-time PCR (qPCR) for these miRNAs alongside with miR-371 ~ 373 and miR-302/367 clusters are recommended in future studies.