2021 NPB-14

This case was originally published in 2021. The information provided in this case was accurate and correct at the time of initial program release. Any changes in terminology since the time of initial publication may not be reflected in this case.

The patient is a nine-year-old boy who presented with a two- to three-week history of headache, nausea, and vomiting. Head MR imaging revealed a 5.4 cm solid and cystic right temporal lobe mass.

Tissue Site
Brain, right temporal lobe

The whole slide image provided is an H&E-stained image of the brain from the right temporal lobe resection specimen.

The neuroimaging features of a rim-enhancing cerebral mass with mass effects (Image A) is suggestive, but not diagnostic, of a high-grade glioma. The histopathology featured large, bizarre multinucleated tumor cells, intervening smaller tumor cells with irregular hyperchromatic nuclei (ie, resembling fibrillary astrocytoma cells), pyknotic nuclei, mitotic figures, and foci of necrosis (Image B, Image C, and Image D). Immunostains revealed patchy GFAP positivity (Image E), a lack of IDH1 R132H staining (Image F), loss of ATRX expression (Image G), and extensive p53 immunoreactivity (Image H). Since pediatric H3 G34-mutant high-grade gliomas are typically IDH-wildtype with loss of ATRX expression and strong p53 staining, an immunostain for H3 G34R/V (both R and V mutant protein antibodies combined) was added, but this was also negative (Image I). An immunostain panel for the four mismatch repair proteins was applied and showed retained expression of MLH1, MSH2, and MSH6 (not shown), whereas the PMS2 stain was completely negative in the nuclei of both neoplastic and non-neoplastic cells (Image J); this was repeated to rule out a technical failure, but the results were identical and the positive controls worked in both runs. The tumor was diagnosed as a glioblastoma, giant cell subtype, IDH-wildtype, WHO grade 4, most likely arising in the setting of CMMRD syndrome (also known as bi-allelic mismatch repair deficiency syndrome). Given this suspicion, paired tumor and germline next-generation sequencing (NGS) was obtained. The assay identified bi-allelic germline PMS2 gene inactivation with one heterozygous exon 4 splice donor mutation and one deletion of the 3' coding exons 7-16; this was diagnostic of CMMRD syndrome. The tumor-associated findings included: 1) a microsatellite unstable tumor with instability at 15% of evaluated microsatellites, 2) extremely high somatic mutation burden consistent with "ultrahypermutation", with a predominance of C>T transitions, C>A transversions, and small indels corresponding with a combination of Mutational Signature 6 associated with defective mismatch repair and Mutational Signature 10 associated with altered activity of the DNA polymerase POLE, 3) a pathogenic POLE mutation, and 4) somatic mutations in a number of glioma-associated genes, including ATRX, TP53, EGFR, NF1, and SETD2. In contrast, the IDH and histone 3 genes were wildtype. Following near total resection, the patient was subsequently treated with radiation therapy, bevacizumab (anti-angiogenic agent), and Nivolumab (immune checkpoint inhibitor); at last clinical followup 15 months after surgery, he was in remission. Further examination of the patient revealed a right arm café-au-lait macule, also consistent with CMMRD. The patient was initiated on syndrome-associated surveillance protocols, and the family was referred for genetic counseling. On further testing, one sibling was negative for any germline pathogenic variants and the other showed a heterozygous pathogenic PMS2 variant. The parents have yet to be tested but are presumably both heterozygous carriers.

Giant cell glioblastoma is no longer considered a specific tumor variant but rather a histologic pattern, since it may be encountered within a variety of different molecular glioma subtypes, including IDH-wildtype, IDH-mutant, H3-wildtype, and H3-mutant forms. Nevertheless, it represents a morphologic manifestation of increased genomic instability, with the giant cells thought to arise from several rounds of endoreduplication, a process wherein DNA is replicated, but the cell fails to divide. This giant cell pattern is overrepresented in several hereditary tumor predisposition syndromes, such as Lynch syndrome (due to a heterozygous pathogenic germline variant in a mismatch repair gene), CMMRD (due to bi-allelic germline inactivation of a mismatch repair gene), and Li-Fraumeni syndrome (due to a heterozygous pathogenic germline variant of the TP53 gene). As such, there should be an increased suspicion of these disorders in glioblastomas with prominent giant cell features, especially if the patient is a child or young adult. In such cases, it is useful to screen the tumor with the four MMR protein immunostains. When there is loss of expression of an MMR protein in tumor nuclei with retained expression in non-neoplastic cells, it suggests the possibility of Lynch syndrome (previously also referred to as Turcot syndrome or brain polyposis syndrome type 1 when there is a glioma) with germline inactivation of one allele and somatic inactivation of the second, although it may still be a sporadic tumor if there is bi-allelic somatic inactivation of both alleles; as such, further genetic testing in the patient would be needed to distinguish these possibilities. In contrast, if there is complete loss of expression, including in normal cells, as was seen in the current case, the possibility of CMMRD syndrome must be considered.

As opposed to the autosomal dominant Lynch syndrome resulting from a heterozygous germline pathogenic variant, CMMRD syndrome is an autosomal recessive disorder caused by bi-allelic germline alterations involving one of four mismatch repair genes (MLH1, PMS2, MSH2, and MSH6). Given this bi-allelic requirement, CMMRD is seen most frequently in the setting of consanguinity. Of note, over 90% of CMMRD patients have cutaneous café-au-lait macules and because of this, patients may be misdiagnosed with NF1. In contrast to NF1, however, patients lack axillary and groin freckling, neurofibromas, and Lisch nodules. Instead, they are predisposed to gastrointestinal polyposis and cancers (virtually 100%), T-cell leukemia/lymphoma (up to 30%), and, less commonly, soft tissue sarcomas and genitourinary cancers. In the CNS, high-grade gliomas/glioblastomas are most common, including those with numerous giant cells. Another potential histologic pattern is sheets of small primitive cells mimicking an embryonal neoplasm. Agenesis of the corpus callosum and venous anomalies have also been reported in CMMRD patients. Interestingly, the family history is often uninformative, especially for patients with bi-allelic PMS2 pathogenic variants, given that there are substantially lower risks of cancer in PMS2 heterozygous carriers compared to those with other mismatch repair gene alterations. As such, germline testing is often required to make a firm diagnosis, though it should be noted that the PMS2 gene has multiple pseudogenes that can sometimes make the interpretation of sequencing assays challenging. The ultra-hypermutation genotype (>100 mutations per megabase, mostly involving single nucleotides) found in this patient’s glioblastoma is unique to CMMRD, given the frequently superimposed POLE or POLD DNA polymerase mutation resulting in a complete DNA replication deficiency. For reasons that are poorly understood however, these tumors have much less microsatellite instability than typically encountered in colorectal and endometrial carcinomas with MMR deficiencies. CMMRD syndrome families benefit from genetic counseling and surveillance protocols, which can lead to early detection and enhanced patient survival. Since the glioma cells are inherently resistant to the most common chemotherapy agents, such as temozolomide, other therapeutic strategies are often utilized, including immune checkpoint blockade, given the remarkably high neoantigen burden in these ultrahypermutated neoplasms.

• Glioblastomas with prominent giant cells in young patients may arise in the setting of tumor predisposition syndromes, including CMMRD, Lynch, and Li-Fraumeni syndromes.
• IHC for MMR proteins is a particularly useful screening tool whenever CMMRD is suspected clinically or pathologically.
• CMMRD is most commonly seen in the setting of consanguinity, with related parents each contributing one of the recessive alleles.
• The high frequency of café-au-lait macules in CMMRD patients sometimes leads to a clinical misdiagnosis of NF1, so this possibility should be considered if one encounters a giant cell glioblastoma in a patient thought to have NF1 (even though this is also possible, albeit less common in NF1).
• CMMRD-associated high-grade gliomas are often treated differently than their sporadic counterparts. Specifically, therapy with alkylating agents such as temozolomide is less likely to be effective, whereas therapy with immune checkpoint inhibition is more likely to be effective.

1. Bakry D, Aronson M, Durno C, et al. Genetic and clinical determinants of constitutional mismatch repair deficiency syndrome: report from the constitutional mismatch repair deficiency consortium. Eur J Cancer. 2014;50(5):987-996.
2. Bouffet E, Larouche V, Campbell BB, et al. Immune checkpoint inhibition for hypermutant glioblastoma multiforme resulting from germline biallelic mismatch repair deficiency. J Clin Oncol. 2016;34(19):2206-2211.
3. Wimmer K, Rosenbaum T, Messiaen L. Connections between constitutional mismatch repair deficiency syndrome and neurofibromatosis type 1. Clin Genet. 2017;91(4):507-519.
4. Guerrini-Rousseau L, Varlet P, Colas C, et al. Constitutional mismatch repair deficiency-associated brain tumors: report from the European C4CMMRD consortium. Neurooncol Adv. 2019;1(1):vdz033.
5. Farah RA, Maalouf F, Chahine NA, et al. Ongoing issues with the management of children with Constitutional Mismatch Repair Deficiency syndrome. Eur J Med Genet. 2019;62(8):103706.
6. Kim B, Tabori U, Hawkins C. An update on the CNS manifestations of brain tumor polyposis syndromes. Acta Neuropathol. 2020;139(4):703-715.