Taking stock of medulloblastoma cell lines

What is it about?

Medulloblastoma is a brain tumour in children, diagnosed in around 5000-8000 children each year worldwide. Scientific discoveries in the genetic makeup of medulloblastoma tumours and their biological circuitry have divided medulloblastoma in four molecular subtypes- WNT, SHH, Group 3 and Group 4. Patients with WNT tumours have >90% survival rates, while Group 3 patients fare poorly with 55% 5-year survival despite multimodal therapy. The current challenge is to move away from historically uniform medulloblastoma chemotherapy treatment and rationally select drugs which may be beneficial for each subgroup. However, the small number of patients within each subgroup calls for a targeted strategy for preclinical testing of future therapies so that only safe and effective treatments are trialled in patients. Tumour cells from patients can help in the rational process of preclinical screening for novel anticancer therapies. Patient-derived cell lines can be grown in the laboratory for years and expanded to large numbers relatively quickly. Some cell lines, such as DAOY and D283, have been cultured for over 30 years, and new ones are derived every year. With 44 existing medulloblastoma cell lines there is an urgent need to link new and long-established cell lines with the molecular subtypes they represent and ultimately the patient tumour they model most closely. This review arranges the available medulloblastoma cell types by molecular subtype and proposes strategies to increase the impact and usefulness of in vitro medulloblastoma models. We review approaches such as three-dimensional cell culture, adding in normal brain cells and brain-like extracellular matrix to improve the relevance of in vitro cultures. Moreover we advocate the establishment of an online database and linked cell bank to facilitate the use of well-characterised and patient-relevant cells.

Featured Image

Why is it important?

While long-established cell lines may have undergone selection in culture or genetic and phenotypic drift, they are widely used to model medulloblastoma tumours and their worldwide availability facilitates reproducibility of results. We have reviewed the published literature and have arranged all established medulloblastoma cell lines by subtype and molecular features. We have examined the citation numbers for all cell lines and have identified that high-risk tumours are over-represented in preclinical models of medulloblastoma. Moreover, the vast majority of patients with WNT, Group 4 tumour and Group 3 tumours without myc amplification remain poorly represented by cell lines and mouse models. Eighteen out of 44 cell lines have been subtyped and 11 of these belong to Group 3 medulloblastoma with myc-amplification. Moreover, 50% (2 out of 4) of the 4 SHH cell lines harbour a mutation in TP53. Twenty-three cell lines have not been subtyped yet and it would be vital to thoroughly characterise them in order to continue using them. We have also reviewed strategies to improve the relevance of in vitro models by tailoring culture conditions to include three-dimensional culture, the normal brain and physiologically-relevant extracellular matrix. Our vision is that medulloblastoma models would benefit from an online database and a linked cell bank matching genetic makeup, phenotype and drug sensitivity.

Perspectives