gms | German Medical Science

Objective: Multilevel heterogeneity is known to be characteristic for glioblastoma. With epigenetic tumor features are getting into focus as prognostic markers, it is likely that such heterogeneity is also to be found correspondingly. Whether intra-tumoral heterogeneity in these epigenetic characteristics may influence prognostic value remains unclear.

Methods: Of 154 patients suffering from primary glioblastoma, 120 patients served as the control collective while 34 patients were compiled as test collective. MGMT, p15 and p16 promoter methylation as well as miRNA-expression levels (miRNA-21, -24, -26a and -181d), were measured in each tumor specimen. Serving as a statistical baseline, epigenetic heterogeneity between tumors (inter-tumoral) was estimated within a triplet of three tumor probes from three different control patients. For estimation of epigenetic heterogeneity within a tumor (intra-tumoral), previous results were compared to three tumor specimens within one glioblastoma of a test patient. Resulting levels of heterogeneity were then correlated with survival and validated by an external TCGA data set.

Results: Heterogeneity in MGMT promoter methylation occurred less likely in the test group compared to the control group (p = 0.006). No difference in heterogeneity was observed between test and control group regarding p15 and p16 methylation (p= 0.33 and p= 0.14). Intra-tumoral heterogeneity within the test group regarding miRNA-21, -24, -26a and -181d expression was not distinguishable from inter-tumoral heterogeneity (p = 0.18, p = 0.32, p = 0.48 and p = 0.11). A homogenously increased miRNA-21 expression was associated with reduced overall survival (p = 0.01). The findings could be validated by comparison with TCGA datasets (p = 0.004).

Conclusion: Heterogeneity of distinct epigenetic characteristics in one glioblastoma may be of the same magnitude as heterogeneity between different patients. Not only the extent of epigenetic characteristics but also the extent of their intra-tumoral heterogeneity may influence survival of glioblastoma patients. Resulting, prognostic value of epigenetic markers in glioblastoma should be interpreted carefully.