De-coding Group 3 medulloblastoma biology with noncoding RNA

See the article by Katsushima et al in this issue pp. 572–585.

Medulloblastoma is one of the most common forms of brain cancer in children. While these tumors are unified by their common location in the posterior fossa and characteristic primitive histomorphologic features, they are a biologically and clinically heterogenous disease. Molecular analysis recently revealed that medulloblastoma can be better subdivided into 4 major subgroups that have distinct biology, prognoses, and therapeutic implications.¹ One subgroup, characterized by WNT (wingless) pathway activation, has a relatively good prognosis, where patients benefit from treatment de-escalation aimed at reducing long-term sequelae of chemoradiation on the brain. Another subgroup, characterized by sonic hedgehog (SHH) pathway activation, shows an intermediate outlook and has garnered significant interest for the development of SHH-targeted therapies. Conversely, groups 3 and 4 medulloblastomas, which make up 60% of cases, are still less well-characterized and often show more aggressive clinical behavior. Specifically, Group 3 medulloblastomas are characterized by gene amplification and increased expression of MYC, very frequently show metastatic disease at presentation, and have an extremely poor prognosis. Despite this well-known biology, therapy design for Group 3 medulloblastoma has been challenging as transcription factors like MYC and downstream partners (eg, HLX) have proven to be difficult targets. Recent studies have attempted to manage MYC dysregulation by using small molecules like JQ1 that inhibit upstream bromodomain and extra-terminal motif (BET) proteins involved in epigenetic regulation of this pathway.^2,3 While successful in reducing MYC activation and prolonging survival in preclinical models, emergence of treatment-resistance is known to occur.^4,5 Targeting canonical downstream targets of MYC also poses unique challenges, due to the pleiotropic effector pathways involved, many of which are essential to both neoplastic and non-neoplastic processes. There is therefore a need for alternative strategies to appropriately manage the MYC-driven biology of Group 3 medulloblastoma.

Recently, the central dogma of biology, in which informative DNA regions are transcribed to RNA and, then coded into effector proteins has been substantially revised by the discovery of a plethora of noncoding RNA species that drive biological processes without translated protein products. Specifically, long noncoding RNAs (lncRNA), defined as non-protein-coding RNA molecules over 200 nucleotides in length, have been shown to play important regulatory roles in cancer including proliferation, differentiation, apoptosis, and invasion.⁶ Given these critical processes, exploring the contributions of the noncoding genome in current models of cancer may help provide novel insights and targets for cancer and other diseases.

By using publicly available RNA sequencing (RNAseq) datasets of 175 medulloblastomas spanning all 4 major molecular subgroups, Katsushima et al. identified lncRNAs enriched in Group 3 medulloblastomas, many of which belonged to the lnc-HLX family.⁷ One of these candidates, lnc-HLX-2-7, located in a known enhancer region in medulloblastoma,⁸ was characterized further and consistently observed to be enriched in Group 3 medulloblastoma across clinical tissue samples and experimental in vitro and in vivo models. Functional knockdown of lnc-HLX-2-7 showed a reduction of the cis neighboring HLX gene, a decrease in tumor growth potential, and higher levels of apoptosis. Moreover, lnc-HLX-2-7 knockdown decreased tumor growth of mouse xenografts. The authors importantly showed that lnc-HLX-2-7 is highly correlated with, and driven, by MYC through direct interaction with this lncRNA’s promoter. Therapeutically, JQ1 not only reduced MYC expression levels, as was previously shown, but also downstream lnc-HLX-2-7 abundance. Finally, at a systems level, lnc-HLX-2-7-deleted medulloblastoma lines showed alternations in mitochondrial and metabolism signaling including oxidative phosphorylation by regulating HLX transcription factor expression. Together, this places lnc-HLX-2-7 as a key downstream effector of the MYC pathway in Group 3 medulloblastoma. In addition to helping decipher biological mechanisms by which MYC activation promotes malignancy in medulloblastoma, revising our models of cancer to include these non-canonical regulators may also provide potential new targets for therapy design.

Therapeutic approaches to target noncoding genome products have so far relied on antisense oligonucleotides, morpholinos, and small molecules that promote ncRNA degradation, alter splicing, and uncouple critical RNA-protein interactions. Encouragingly, animal studies, injecting an antisense oligomer directed against MYC, have shown that they could target and inhibit MYC translation, supress growth and promote apoptosis in human prostate cancer xenograft mouse models.⁹ These models also showed decreased tumor cellularity, vascularity and increased degeneration. The same study showed such RNA-based therapies, against the MYC pathway, can be relatively safe in humans.⁹ While these studies did not target lncRNA specifically, they demonstrate the potential of targeting non-translated species for the efficient regulation of MYC biology.

With these exciting implications in mind, it is important to emphasize that noncoding RNA precision medicine trials in cancer are only now beginning to be formally explored. Many of these early clinical trials are still notably observational and focused on quantifying ncRNAs in cancers, including gliomas, and exploring their utility as prognostic and predictive biomarkers.¹⁰ Most therapeutically-oriented ncRNA clinical trials are still very early in their evolution and are concentrated on assessing safety, pharmacokinetic and pharmacodynamic aspects. While it is exciting to see how this field evolves, it is important to stress that most preclinical studies suggest that delivery technologies for these therapeutics, with improved efficiencies, are likely a prerequisite before any positive results can be observed in human cancer trials. Similarly, solutions to improve bioavailability by reducing nuclease degradation, balancing off-target effects, and promotion of effective uptake of these therapies into cancer cells are still needed. These practical issues with targeting the noncoding genome in cancer may therefore take many years to catch up with the excitement of these emerging biological targets. With these current limitations in mind, small interfering RNA-based therapies such as those developed by Alnylam are now beginning to be approved for a variety of human diseases including hyperoxaluria type 1, acute hepatic porphyria, and hereditary transthyretin-mediated amyloidosis. The characterization of the noncoding cancer genome, as we continue to learn from pioneering trials, offers hope and new solutions for nominating new targets for tumor subtypes that are proving challenging to address with more classical approaches. It will be interesting to see how this new layer of information contributes to our understanding and management of not only medulloblastoma but also other cancers and diseases.

Conflict of interest statement. Not applicable.