“IMMUNOHISTOCHEMICAL CHARACTERIZATION OF A SERIES OF PEDIATRIC MEDULLOBLASTOMAS” - Investigation of JAM molecules expression in pediatric medulloblastoma

Background and aims. Medulloblastoma (MB) is the most frequent embryonal tumor of the CNS and the most common malignat pediatric brain tumor, arising mainly in the fourth ventricle. MB is characterized by four molecularly defined subgroups with distinct clinicopathological features and prognosis: WNT (MBWNT), SHH (MBSHH), group 3 (MBGrp3), and group 4 (MBGrp4). Standard therapy is ineffective in subsets of MB, thus there is a great need to identify novel prognostic markers and treatment targets. Junctional adhesion molecules (JAM)-A, -B, -C may be relevant candidate for this role, based on clinical and preclinical data. The aim of this study was to examine gene and protein expression of JAMs and assess their prognostic value by relying on both open access large datasets and a sample of pediatric MB. Methods. In the gene expression study, founded on two public datasets (n=763 and n=223), combined risk stratification models based on transcript levels of JAM genes (namely F11R, JAM-2 and JAM-3) and clinically relevant features were identified by multivariate Cox proportional hazards analysis. In the experimental study, we sought to examine the immunohistochemical expression of JAM-A,B,C in a sample of pediatric MB (n=65). Nanostring profiling and immunohistochemistry were adopted to obtain molecular classification. Association and survival analyses were conducted to validate trascriptomic findings. Results. Gene expression analyses showed that JAMs were differentially expressed across consensus molecular subgroups of MB (p<0.0001). In particular, JAM-3 was differentially expressed between MBGrp3 and MBGrp4 and its overexpression was significantly associated with large cell/anaplastic (LC/A) pathology (p=0,007) and shorter survival (p=0,002). These findings were only partially supported by immunohistochemical data, since high expression (3+) of JAM-C was strongly associated with a subset of MBGrp3 with LC/A pathology but did not differentiate between MBGrp3 and MBGrp4. Further, survival analyses displayed a trend for a shorter survival in patients with high expression of JAM-C. Thus, JAM-C may be useful to identify a subset of MBGrp3 with LC/A pathology carrying poor clinical outcomes. Conclusions. JAM-C may have a role as subtyping marker and prognostic factor in pediatric MB. Our findings are consistent with recent evidence suggesting that MBGrp3 and MBGrp4 display significant molecular overlap. In particular, JAM-C is useful to identify a subset of MBGrp3 showing LC/A pathology and dismal prognosis which could at least partially correspond to a recently identified high-risk subtype of MBGrp3 showing LC/A pathology and harbouring MYC amplification. Given the lack of reliable immunohistochemical markers for such emerging subgroup, further studies are needed to investigate the role of JAM-C in relation to clinicopathological features, molecular characteristics and overall prognosis.