Advances in Genomics Explain Medulloblastoma Behavior at the Bedside

ATRT: atypical teratoid/rhabdoid tumors CNS: central nervous system MB: medulloblastoma PARP: poly (ADP-ribose) polymerase SHH: Sonic hedgehog SMO: smoothened WHO: World Health Organization WNT: Wingless Central nervous system (CNS) tumors are the most prevalent pediatric solid tumors and a major cause of cancer-related mortality in children and adolescents. Medulloblastomas (MBs) are clinically aggressive, fast-growing embryonal tumors that arise in the posterior fossa due to aberrations of cerebellar development.1,2 They make up about 20% of all malignant childhood CNS tumors.3-5 Since Bailey and Cushing first introduced the term “medulloblastoma” to describe all small round blue cell tumors of the cerebellum,6 our understanding of these tumors has greatly evolved. During the 1990s, the discovery of alterations in the tumor suppressor gene hSNF5/INI1 in atypical teratoid/rhabdoid tumors (ATRT) allowed pathologists to distinguish MBs from ATRTs for the first time.7,8 The explosion of genomic studies that followed further led to the identification of distinct MB molecular subgroups—Wingless (WNT), Sonic hedgehog (SHH), group 3, and group 4—and advances in molecular profiling techniques set the stage for biology-based risk stratification of patients, and a new generation of clinical trials that explore targeting specific molecular alterations.9-13 Current therapy for MB patients includes surgical resection, craniospinal irradiation (for children older than 3 yr of age), and chemotherapy, which does cure the majority of patients diagnosed with MB.14,15 However, these intensive therapies are associated with significant long-term toxicities, eg, neurocognitive deficits, deficits in neuroendocrine functions, deafness, diminished fertility, and radiation induced cancers.4,16 Furthermore, metastatic disease responds poorly to conventional therapy, and recurrent MB remains a nearly universally fatal disease.17-21 PROGNOSTIC VALUE OF MOLECULAR CLASSIFICATION In 2006, Thompson et al12 reported for the first time on different subgroups of MB based on genome-wide expression profiles. Over time, several laboratories from around the world validated and expanded these findings.9-11 At a conference in Boston in 2010, experts agreed on the existence of 4 main MB subgroups, WNT, SHH, group 3, and group 4 MB; each with their own specific demographic, transcriptional, and genetic differences with important clinical implications.13 Both histological and molecular information is now being used clinically to treat medulloblastoma patients, as reflected in the recent 2016 World Health Organization (WHO) classification of CNS tumors.22 The molecular classification of MBs is rapidly becoming part of routine diagnostics, and molecular subgrouping testing should be implemented in pediatric oncology centers caring for these patients. At our institution, a Clinical Laboratory Improvement Amendments-approved nanoString nCounter assay is routinely used at diagnosis for all patients with MB with a turnaround time of less than 2 wk. WNT MBs originate from lower rhombic lip progenitors of the dorsal brainstem.23 The hallmark of these tumors is activation of the WNT signaling pathway, usually secondary to somatic mutations of the CTNNB1 gene that encodes β-catenin.13 Monosomy of chromosome 6 is present in about 80% of these tumors,24 but it is not pathognomonic as it is present at lower frequency across all non-WNT subgroups. Germline mutations in the APC tumor suppressor gene, which encodes the negative regulator of β-catenin APC protein, predispose to Turcot syndrome, and is associated with an increased risk of developing MB.13 WNT MBs are the least common subgroup, accounting for about 10% of all MBs, and are typically diagnosed in older children and adolescents.25 These tumors rarely metastasize and multiple prospective and retrospective studies confirmed the excellent survival rates consistently above 90% with current standard therapy.26-29 Nowadays, clinical trials focus on de-escalating chemo-radiotherapy for these patients with the goal of maintaining excellent survival while reducing long-term side effects. SHH MBs derive from granule neural progenitors of the cerebellar external granule layer23 and are defined by activation of the SHH pathway. Somatic mutations of PTCH1, SMO, and SUFU, and amplifications of GLI2 and MYCN drive the development of SHH MB.30 Deletion of chromosome 9q seems to be exclusive in SHH MB and supports the important role of the PTCH1 gene, located on chromosome arm 9q22. Furthermore, germline mutations of the PTCH1 gene in Gorlin syndrome predispose to basal-cell carcinomas and MB.13 Twenty percent of SHH tumors harbor TP53 mutations, which define a very high risk group of patients, whereas in WNT MB, TP53 mutations have no correlation with outcome.31 Chromosome 14q loss, chromosome shredding (chromotripsis) as well as focal GLI2 and/or MYCN amplification are other features associated with poor outcome in SHH MB patients.28 SHH tumors account for one-third of all MBs with age peaks in infancy and adulthood. Whereas PTCH1 mutations are seen across all ages, SUFU mutations are typically seen in infants, TP53 mutations in school-aged children, and SMO and TERT promoter mutations in adults.30 SHH MBs were initially described as intermediate risk tumors; however, there is a significant heterogeneity in outcomes related to the underlying genetic alterations, with p53-mutated SHH tumors having a dismal survival outcome.13,26 The recent WHO classification has defined p53-mutated SHH MB as a separate entity and these patients should be prioritized for the introduction of new therapies.22 Group 3 MBs remain of unclear cell origin. Group 3 MBs have a GABAergic gene expression signature,24 and it has been proposed that these tumors might arise from multipotent neural stem cells in the ventricular zone of the embryonic cerebellar anlage that generate cerebellar GABAergic neurons. MYC amplifications appear to be almost exclusive to group 3 MBs and are associated with a poor outcome.28 Other characteristics of this subgroup are gain of chromosome 1q, loss of chromosome 5q and chromosome 10q as well as isochromosome 17q.13 Isochromosome 17q was found to correlate with poorer outcome only in group 3 patients,28 however not across all studies.27 Accounting for 25% to 30% of all MBs, group 3 tumors affect exclusively infants and children and are frequently metastatic at diagnosis. Group 3 MBs have the worst prognosis compared with the other subgroups,26,29 particularly in infants32 and children with metastatic disease and/or MYC amplification28 (5-yr progression-free survival less than 50%). Group 4 MB cellular origin has not yet been defined.23 Although group 4 MBs represent the most common subgroup, the molecular pathogenesis is the least understood, among other reasons, due to the lack of mouse models of group 4 tumors.13 Group 4 MBs have a glutamatergic molecular signature, and recurrent cytogenetic alterations include MYCN amplification, CDK6 amplification, loss of chromosome 11, and duplication of the region in chromosome 5 where the SNCAIP gene is located.24 Isochromosome 17q is the most prevalent cytogenetic change in group 4 MB but is not predictive for survival.11,28 Interestingly, retrospective analysis of a large cohort of MB showed that group 4 tumors with chromosome 11 loss and/or whole chromosome 17 gain had a good outcome regardless of metastatic disease and a recent expert consensus defined patients with localized group 4 tumors with chromosome 11 loss or 17 gain as a low-risk group.33 RISK STRATIFICATION BASED ON CLINICAL AND MOLECULAR CRITERIA Since the 1969 Chang risk classification (Table), patient stratification for postoperative treatment has been based solely on clinical criteria; patients older than 3 yr of age at diagnosis are classified as high-risk patients if they show incomplete surgical resection with a residual tumor mass ≥1.5 cm2 on postsurgical imaging and/or the presence of leptomeningeal dissemination (based on craniospinal magnetic resonance imaging and cerebrospinal fluid analysis following lumbar puncture). Patients without these features are stratified as average risk, usually showing a better disease control and therefore a higher 5-yr survival.13,33,34 Children under 3 yr of age generally have a poor prognosis and are considered high risk, except those historically classified as nodular desmoplastic MB or extensive nodularity MB, which belong exclusively to the SHH subgroup.35 Due to their young age, craniospinal irradiation is contraindicated and the patients are treated using radiation-sparing strategies.TABLE: Comparison between Risk Classification based on Clinical Criteria and New Patient Risk Groups based on the 2016 Expert Consensus Statement33Stratification of pediatric MB patients into 1 of the 2 risk groups does not reflect the disease heterogeneity and the spectrum of patient outcome and response to treatment. Furthermore, MB survivors continue to pay a high price for cure with debilitating long-term sequelae.3,4 A better understanding of MB biology and response to treatment are crucial in order to properly assign a patient to a more or less intensive treatment. During a consensus meeting in 2015, a new approach of risk stratification in noninfant patients (ages 3-17) with MB was defined, based on survival outcomes of different patient groups according to clinical and molecular biomarkers.33 Patients with completely resected nonmetastatic WNT MB as well as patients with nonmetastatic group 4 tumors with chromosome 11 loss should be considered low risk, with a predicted overall survival above 90%. Even WNT patients with incomplete resections are likely low risk and de-escalation of therapy is likely appropriate.36 Nonmetastatic SHH patients without TP53 mutations or MYCN amplifications as well as nonmetastatic, non-MYC amplified group 3 and group 4 MB without chromosome 11 loss should be considered standard risk (75%-90% survival). Metastatic dissemination is generally a marker of poor prognosis. Therefore, patients with metastatic noninfant TP53 wild-type SHH or metastatic group 4 MB should be considered high-risk patients with a predicted overall survival between 50% and 75%. Nonmetastatic, MYCN-amplified SHH patients should also be stratified as high risk.33 p53-mutant SHH and MYC-amplified, metastatic group 3 patients have a dismal prognosis and should be considered very high risk. These recommendations should aid the design of the next generation of clinical trials, in order to address questions specific to distinct MB subgroups and expedite the introduction of novel therapies.4 SUBGROUP-BASED TREATMENT AND TARGETED THERAPIES Multimodal treatment consisting of maximal safe tumor resection and radiotherapy for patients older than 3 yr, including irradiation of the tumor bed and the craniospinal axis, followed by chemotherapy improved overall survival of patients with childhood MB.14,15,26 Although surgery plays a key role in the first-line treatment and local tumor control, a recent study showed that there is no statistically significant difference of overall survival between gross total (no residual tumor) and near-total resection (<1.5 cm2 tumor remaining) in MB patients. Only in patients with metastatic group 4 MB, gross total resection increases progression-free survival compared to subtotal resection (≥1.5 cm2 tumor remaining), but there is no concomitant improvement in overall survival. Therefore, although maximum safe surgical resection remains standard of care, aggressive surgical resection should not be performed at the risk of postoperative neurological morbidity.36 Molecular profiling analyses of primary tumor samples have allowed the identification of key signaling pathways, sonic hedgehog and WNT, involved in tumorigenesis of MB; inhibitors of these pathways have shown antitumor activity in preclinical studies. In several cases, subdivisions of patients, who may benefit from a specific targeted therapy, can be preselected and inhibitors have been investigated in clinical trials.1,37,38 The WNT pathway plays a key role in cell proliferation and differentiation, among other essential cell processes.39 β-catenin, as the main effector of the WNT pathway, as well as the poly (ADP-ribose) polymerase (PARP) enzyme are overexpressed in WNT MB. Most WNT pathway inhibitors target PARP causing destruction of β-catenin.40 Although several drugs have been tested in preclinical studies, current clinical trials in Europe and North America focus on de-escalation of treatment, as patients with WNT MB are at low risk for relapse with an excellent prognosis. SHH tumors are driven by an overactivated SHH pathway, and these patients have been the first to benefit from targeted treatment. Vismodegib, an inhibitor of the transmembrane Shh coreceptor smoothened (SMO), was the first to show antitumor activity in a patient with MB and extraneural metastases,41 and is the most extensively studied SMO inhibitor so far. It has been well-tolerated in phase I/II clinical trials.42 SMO inhibitors are, however, ineffective in SHH MB with downstream aberrations (eg, GLI2 amplification).30 Furthermore, upregulation of the PI3K/Akt/mTOR pathway in SMO-resistant tumors has been identified as another potential mechanism of resistance.43 Dual inhibition of PI3K and SMO showed a markedly delayed development of resistance in MB.44 Specific protocols should also be designed for the high-risk group of patients with p53-mutated SHH-MB. GSK-3β inhibitors (acting as a radiosensitizer),45 bromodomain inhibitors modulating GLI1 transcription,46 and GLI1 inhibitors47 demonstrated efficacy in vivo and are promising agents for further clinical trials. There is a paucity of therapeutic targets for group 3 and group 4 MB; molecular profiling studies failed to identify targetable molecular drivers in the majority of tumors,24 and the lack of spontaneous mouse models of the disease further hampers hypotheses-generating studies. Pemetrexed and gemcitabine have been tested as potential therapeutic agents in preclinical MYC-driven models48 and were introduced in the current SJMB12 trial (NCT01878617) for patients in the high-risk stratum of the non-WNT, non-SHH arm. Preclinical studies showed efficacy of histone deacetylase inhibitors,49,50 PI3K inhibitors,50 and BET-bromodomain inhibitors51-53 in the treatment of MYC-amplified MB, and these agents should be prioritized for clinical trials. Metastatic MB About 30% of all pediatric MB patients present with leptomeningeal dissemination at diagnosis.54 The degree of metastasis is classified as M1 if there is evidence of tumor cells in the cerebrospinal fluid, M2 if there is radiographic evidence of gross nodular seeding in the cerebellar/cerebral subarachnoid space, or in the third or lateral ventricles or M3 if there is radiological evidence of spinal metastases. WNT tumors rarely metastasize and data on prognostic impact of metastases in this subgroup are scarce. Approximately 20% of SHH MB, 40% to 50% of group 3 MB, and 30% of group 4 tumors are metastatic at presentation.26 The presence of M2/M3 disease is a strong predictor of poor prognosis across all non-WNT subgroups.28 Our current knowledge of MB tumorigenesis derives mostly from studies of the primary tumor. The metastatic compartment—which accounts for significant morbidity and treatment failure—remains understudied, largely due to the limited availability of samples, given that metastases are not typically biopsied. Indeed, metastases significantly diverge from the primary tumor and targeted therapies tailored to the primary tumor will likely be unsuccessful.20 The fact that most group 3 and group 4 MB recur metastatically rather than at the primary site (Figure) strongly suggests that our drug development strategies need to consider differences between primary and secondary disease, taking into account both local and metastatic compartments.FIGURE: A, Localized group 3 MB at diagnosis with characteristic location in the midline and contrast enhancement. B, Same patient at relapse after completion of treatment with surgery, craniospinal irradiation, and chemotherapy with extensive intracranial and spinal leptomeningeal dissemination.Recurrent MB The frequency and pattern of disease at recurrence is subgroup-specific; the majority of SHH tumors recur at the primary tumor site in the posterior fossa, while group 3 and group 4 tumors almost always recur as disseminated leptomeningeal disease without evidence of tumor in the tumor bed. There has been no obvious pattern found for the rare recurrences in WNT MB.17 Although the molecular subgroup remains stable,55 recurrent MB—regardless of subgrouping—is an incurable disease and most therapeutic approaches at relapse have a palliative intent, with the exception perhaps of craniospinal irradiation with or without high-dose chemotherapy and stem cell rescue for young children with relapsed MB that were not irradiated upfront and may be salvaged at relapse.56 For each individual patient, different approaches—repeated surgery, irradiation, chemotherapy, or high-dose chemotherapy with stem cell rescue and molecular targeted therapies—should be carefully considered and discussed with the family at the time of relapse, as data are scarce and there is no evidence-based standard of care. Whenever a targeted agent is considered, patients should undergo rebiopsy of the recurrent tumor and be enrolled in phase I/II clinical trials whenever possible. Targets for molecularly tailored therapy discovered at presentation may not be present at recurrence and targeted therapies based on analysis of the primary tumor are thus doomed to failure at recurrence. Clinicians have observed for decades that MB at relapse is a “new” and deadly disease with a very distinct behavior, refractory to therapy. Indeed, a recent study showed striking clonal divergence of MB at relapse, with the dominant clone at recurrence rarely being the dominant clone seen at diagnosis, reflecting the clonal selection and emergence of resistant clones during treatment.20 This fact has huge clinical implications and challenges us to rethink our traditional practices, and to start rebiopsying tumors at recurrence in order to offer appropriate targeted therapy.13,20 CONCLUSION Our understanding of the heterogeneity of MB has dramatically increased in the recent past. New approaches of clinical risk stratification of MB patients and advances in mapping of genetic drivers may open doors for new ways of adjusted therapies in the future, which will potentially be more effective and less toxic. Targeted treatment is based on the identification of therapeutic targets present exclusively in the cancer cells but absent from the normal host cells. Novel oncological agents are usually tested in children previously highly treated with radiation and chemotherapy, in the context of phase I/II clinical trials. Paradoxically, most of the basic and translational research on MB is based on treatment-naïve xenografts and mouse models. The current clinical studies in which new drugs are tested at recurrence are therefore based on the assumption that biology at recurrence is largely similar to biology at presentation. This hypothesis appears to be wrong as current data show that MB demonstrates striking evolution over time and metastases genetically diverge from the primary tumor. Targets for rational therapy discovered and documented at presentation may no longer be present in the dominant clone at the time of recurrence. Therefore, novel therapies developed through study of the therapy naïve primary tumor may not increase survival rates and seem doomed to failure in case of recurrence. These observations have clinical implications and although historically surgery has not been indicated at relapse, rebiopsy should be considered if targeted therapies are a therapeutic option. Disclosure The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.