B progenitor acute lymphoblastic leukaemia, B-ALL for short, is a most common paediatric cancer. It is a very heterogeneous disease, it has dozens of subtypes and so its classification system plays a critical role for diagnosis, prognosis and even tailored therapies. Over the past few years many novel subtypes have been identified but still 10-30% of the cases cannot be classified into any of the subtypes. For those cases a lot of them have very poor outcome. In our study we tried to identify novel subtypes and the germline genetic lesions in this group of cases.
In this study we tried to identify the novel subtypes in this group of cases. We’re aware of two issues: first the novel subtypes could be minor subtypes, that means we need to have a very large cohort to address these minor subtypes. So through collaboration we collected around 2,000 B-ALL cases; we covered different age groups. The other issue is the subtype defined lesions, they could be cryptic or challenging to be detected so I proposed to do integrative genomic analysis. That means all the cases have RNA-Seq data to identify their gene arrangements and gene expression profile. Also the novel subtypes will have whole genome or whole exome sequencing data to address their mutation spectrum. Over half of the cases will have SNP array to analyse their copy-number alterations. This distinct gene expression profile is the key feature of a subtype so I proposed to use three different algorithms to analyse their gene expression profile. After the identification of the novel lesions we’ll perform functional experiments to validate their role in driving leukaemia.
For this study we demonstrate the power of using RNA-Seq data to identify novel subtypes. We established a revised B-ALL classification pipeline, we classified B-ALL into 23 different subtypes. Among these subtypes eight of them are novel discovered in this study; two are defined by PAX5 genetic alterations. PAX5 is a very important transcription factor in B-cell development. Besides that, we performed a functional experiment to prove that PAX5 or P80R missense mutation could block B-cell development and initiate leukaemia. That is a subtype we define based on PAX5 P80R missense mutation.
Together with another subtype defined by missense mutation, IKZF1 N159Y we proved that that transcription factor missense mutation could be subtype defining lesions.
What could the implications be for treatment in the future?
Through this study we classified two-thirds of the previously uncategorised cases into distinct subtypes. We clarified their genetic mutation background and also evaluated their clinical outcome. In the future researchers could develop functional experiments and clinical treatments for those cases to improve their clinical outcome. Also, St Jude’s is implementing this pipeline to benefit the patients in house and ultimately it will launch this pipeline on the St Jude cloud to serve the patients around the world.