The occurrence of a mutation in a cancer predisposition gene has been estimated to account for more than 40% of the medulloblastoma (MB), SHH-activated [3, 4]. Fourteen percent of them have been reported with bi-allelic alterations of ELP1, a tumor-suppressor gene being currently the most frequent to predispose to MB [4]. The ELP1 gene encodes for the protein ELP1 which is a component of the elongator complex, a six-subunit protein complex (ELP1-6) implicated in neurogenesis [1, 2]. The bi-allelic inactivation of ELP1 results from the combination of a germline alteration and a loss of chromosome 9q [4]. The aim of our study was to evaluate the sensitivity and specificity of ELP1 immunostaining to detect ELP1-associated MB.

Our study included a total of 132 DNA-methylation profiled MB: 57 SHH-activated (aged from 0- to 18-year-old), 15 WNT-activated, 30 group 3, and 30 group 4. We performed immunohistochemistry (IHC) for the ELP1 antibody (clone 6G9; 1:50 dilution; Sigma-Aldrich; Bromma, Sweden) on 3 µm-thick sections of formalin-fixed, paraffin-embedded tissue samples, performed on an Omnis automate. Tumoral molecular analysis of ELP1 was conducted with a custom Next-Generation Sequencing (NGS) panel (Supplementary table S1). The library was prepared with the SureSelect XT-HS according to the manufacturer’s protocol (Agilent) and sequenced on an Illumina NovaSeq 6000. The sequence of all coding exons of ELP1 (NM_003640.4) and PTCH1 (NM_000264.3) and the loss of the heterozygosity (LOH) status of chromosome 9q were analyzed afterward. Proteomic has been quantified by a data-independent acquisition method following the same protocol as in [4]. We selected a proteome dataset composed of 16 MB, SHH-activated with five samples showing the ELP1 pathogenic variation identified by genome and proteome techniques and previously reported in [4]. Finally, we tested by immunohistochemistry other pediatric tumor types of the posterior fossa (47 ependymomas, group A, 15 ependymomas, group B, 10 embryonal tumors with multilayered rosettes, 10 atypical teratoid and rhabdoid tumors, 10 central nervous system tumors with BCOR internal tandem duplication, and 10 pilocytic astrocytomas).

A complete loss of cytoplasmic ELP1 staining in all tumor cells (with intra-tumoral vessels as a positive internal control) was observed in 12/57 (21%) of MB, SHH-activated (Fig. 1a–c), and was preserved in all other MB subgroups (Fig. 1d) and in other tumor types (Supplementary Fig. 1). Molecular analyses revealed the presence of bi-allelic ELP1 alterations (Table 1 for details) in 11/12 MB, SHH-activated, where ELP1 stained negatively. Thus, the sensitivity and the specificity of the IHC were evaluated as 99% (121/122) and 100% (11/11), respectively, in MB. Interestingly, for the unique discordant case, proteomic analyses revealed concordant downregulated levels of ELP1 (Supplementary Fig. 2). From a molecular perspective, this MB harbored a chromosome 9q copy-neutral LOH (confirmed by FISH analysis of chromosome 9) but the sequencing analysis failed to reveal any additional nucleotidic or copy number alteration at the ELP1 locus.

Fig. 1
figure 1

ELP1 expression in medulloblastomas SHH-activated. Case 1: A distinct ELP1 loss in one case of medulloblastoma, SHH-activated, with a bi-allelic alteration of ELP1 with cerebellum parenchyma and vessels as positive internal controls (a, magnification, ×150). Case 1: A distinct ELP1 loss in one case of medulloblastoma, SHH-activated, with a bi-allelic alteration of ELP1 (b, magnification, ×400, with endothelial cells as positive internal control). Case 7: A distinct ELP1 loss in one case of medulloblastoma, SHH-activated, with a bi-allelic alteration of ELP1 (c, magnification, ×400). A preserved expression of ELP1 in one case of medulloblastoma, SHH-activated, without alteration of ELP1 (d, magnification, ×400). Black scale bars represent 100 µm (a) and 50 μm (b–d)

Full size image
Table 1 Histomolecular data of our cases with ELP1 loss of immuno-expression
Full size table

Altogether, ELP1 IHC constitutes a fast, low-cost and conservative tissue-consuming method to detect ELP1-associated MB. Only one case presented a loss of expression without a bi-allelic alteration of ELP1 identified, suggesting the presence of a cryptic alteration (no deep intronic pathogenic variant, complex structural variant, promoter genomic alteration or hyper-methylation was detected with our NGS analysis). The higher proportion (19 vs. 14% in the literature) is explained by the large number of children in our cohort [4]. Here, none of the group 4 MB (n = 30) harbored an ELP1 mutation, confirming that this is a rare event, as already suggested by the previous literature [4].

To conclude, we demonstrated that ELP1 IHC is a highly specific and sensitive biomarker for identifying ELP1-associated MB and should be part of the neuropathologist’s routine panel of antibodies to possibly screen a related predisposition syndrome in these children.