Multi-omics analysis defines highly refractory RAS burdened immature subgroup of infant acute lymphoblastic leukemia

Tomoya Isobe, Masatoshi Takagi, Aiko Sato‐Otsubo, Akira Nishimura, Genta Nagae, Chika Yamagishi, Moe Tamura, Yosuke Tanaka, Shuhei Asada, Reina Takeda, Akiho Tsuchiya, Xiaonan Wang, Kenichi Yoshida, Yasuhito Nannya, Hiroo Ueno, Ryo Akazawa, Itaru Kato, Takashi Mikami, Kentaro Watanabe, Masahiro Sekiguchi, Masafumi Seki, Shunsuke Kimura, Mitsuteru Hiwatari, Motohiro Kato, Shiro Fukuda, Kenji Tatsuno, Shuichi Tsutsumi, Akinori Kanai, Toshiya Inaba, Yusuke Shiozawa, Yuichi Shiraishi, Kenichi Chiba, Hiroko Tanaka, Rishi S. Kotecha, Mark N. Cruickshank, Fumihiko Ishikawa, Tomohiro Morio, Mariko Eguchi, Takao Deguchi, Nobutaka Kiyokawa, Yuki Arakawa, Katsuyoshi Koh, Yuki Aoki, Takashi Ishihara, Daisuke Tomizawa, Takako Miyamura, Eiichi Ishii, Shuki Mizutani, Nicola K. Wilson, Berthold Göttgens, Satoru Miyano, Toshio Kitamura, Susumu Goyama, Akihiko Yokoyama, Hiroyuki Aburatani, Seishi Ogawa, Junko TakitaNature Communications 2022 · DOI

TL;DR

This study uses multi-omics data (RNA-seq, methylation, WES/targeted sequencing) to define five distinct molecular subgroups of KMT2A-rearranged infant acute lymphoblastic leukemia (ALL). A particularly refractory subgroup (IC2) is characterized by a near-universal and heavy burden of RAS pathway mutations, suggesting a new target for individualized therapy.

What Problem Does This Paper Address?

Infant acute lymphoblastic leukemia (ALL) with KMT2A rearrangement is a particularly aggressive and difficult-to-treat form of childhood cancer, with a dismal event-free survival rate. Despite this, current molecular stratification is limited, meaning a better understanding of the disease's heterogeneity is crucial for developing more effective, personalized treatments. The authors aim to uncover this molecular heterogeneity by integrating multiple types of genomic data.

Methods

The researchers employed a multi-omics approach, integrating:

  • RNA sequencing: To analyze gene expression patterns and identify fusion transcripts.
  • DNA methylation array analysis: To assess epigenetic modifications.
  • Whole exome sequencing (WES) and targeted deep sequencing: To identify somatic mutations and copy number alterations.

They utilized Similarity Network Fusion (SNF), a machine learning algorithm, to integrate these diverse datasets and identify robust patient subgroups. Single-sample and single-cell inference methods were used to map these subgroups onto hematopoietic differentiation stages.

Datasets

The study analyzed diagnostic samples from 84 infants with KMT2A-rearranged leukemia. A discovery cohort of 61 infants was used for the primary multi-omics clustering and analysis. An independent cohort of 23 infants was used for validation. Additionally, they leveraged publicly available datasets for normal B-cell progenitors and a single-cell atlas of murine gastrulation.

Key Insights and Ideas

  • Five distinct molecular subgroups: The multi-omics approach revealed five robust integrative clusters (ICs) that go beyond single-omics classifications, highlighting a more nuanced view of KMT2A-r infant ALL.
  • IRX vs. HOXA axis: The study confirms and refines the previously described IRX and HOXA transcriptional subtypes, further subdividing them based on multi-omics data and clinical outcomes.
  • RAS pathway mutations as a critical driver: The most refractory subgroup (IC2) is defined by a near-universal (100%) and heavy burden of RAS pathway mutations, suggesting this pathway is a key driver of this aggressive phenotype.
  • Developmental stage matters: The identified subgroups correlate with distinct stages of B-lymphopoietic and early hemato-endothelial development, indicating that the cell of origin and its differentiation state play a significant role.
  • Single-cell insights: Single-cell analysis of a double-positive case revealed heterogeneity within a single patient, mimicking normal developmental trajectories and highlighting the potential for distinct leukemic stem cell populations.

Key Results

  • Five Integrative Clusters (ICs): The SNF algorithm successfully integrated transcriptomic and methylomic data to define five ICs (IC1-5), which significantly correlated with patient prognosis.
  • IC2: The refractory subgroup: This cluster, characterized by an IRX-type differentiation signature, showed the shortest event-free survival (EFS) and was associated with a 100% frequency of RAS pathway mutations, with a significantly higher number of mutations per case compared to other clusters.
  • RAS pathway mutation burden predicts prognosis: The number of RAS pathway mutations was a significant predictor of EFS, with patients having ≥3 mutations showing worse outcomes.
  • Developmental signatures: ICs were enriched for specific B-lineage developmental signatures, with IC1 showing hematoendothelial progenitor enrichment, IC3/IC5 showing pre-pro-B enrichment, and IC4 showing pre-B enrichment. IC2 was the least differentiated.
  • KMT2A fusion's role: ChIP-seq data suggested that KMT2A fusions can directly activate either IRX or HOXA programs, influencing developmental phenotypes.

Limitations and Future Work

  • Rarity of the disease: The small sample size, inherent to infant ALL, limits the statistical power for some analyses and the ability to fully explore rare subtypes.
  • Subclonal mutation detection: While deep sequencing improved detection, the full extent of subclonal diversity and its impact on relapse requires further investigation in larger cohorts.
  • Cell of origin: While evidence points to less differentiated cells (HSCs/MPPs) as the origin for IRX-type ALL, further studies are needed to confirm this definitively, especially considering the in utero origin of KMT2A rearrangements.
  • Therapeutic targets: The identification of the RAS pathway as a critical driver in IC2 opens avenues for developing targeted therapies (e.g., MEK inhibitors), but these require preclinical validation and clinical trials.
  • Inter-patient heterogeneity: The study highlights significant intra- and inter-patient heterogeneity, emphasizing the need for genomics-guided risk stratification and individualized therapy development.

Summary updated

Abstract

KMT2A-rearranged infant acute lymphoblastic leukemia (ALL) represents the most refractory type of childhood leukemia. To uncover the molecular heterogeneity of this disease, we perform RNA sequencing, methylation array analysis, whole exome and targeted deep sequencing on 84 infants with KMT2A-rearranged leukemia. Our multi-omics clustering followed by single-sample and single-cell inference of hematopoietic differentiation establishes five robust integrative clusters (ICs) with different master transcription factors, fusion partners and corresponding stages of B-lymphopoietic and early hemato-endothelial development: IRX-type differentiated (IC1), IRX-type undifferentiated (IC2), HOXA-type MLLT1 (IC3), HOXA-type MLLT3 (IC4), and HOXA-type AFF1 (IC5). Importantly, our deep mutational analysis reveals that the number of RAS pathway mutations predicts prognosis and that the most refractory subgroup of IC2 possesses 100% frequency and the heaviest burden of RAS pathway mutations. Our findings highlight the previously under-appreciated intra- and inter-patient heterogeneity of KMT2A-rearranged infant ALL and provide a rationale for the future development of genomics-guided risk stratification and individualized therapy.
BibTeX
@inproceedings{isobe2022multi,
  author = {Tomoya Isobe and Masatoshi Takagi and Aiko Sato‐Otsubo and Akira Nishimura and Genta Nagae and Chika Yamagishi and Moe Tamura and Yosuke Tanaka and Shuhei Asada and Reina Takeda and Akiho Tsuchiya and Xiaonan Wang and Kenichi Yoshida and Yasuhito Nannya and Hiroo Ueno and Ryo Akazawa and Itaru Kato and Takashi Mikami and Kentaro Watanabe and Masahiro Sekiguchi and Masafumi Seki and Shunsuke Kimura and Mitsuteru Hiwatari and Motohiro Kato and Shiro Fukuda and Kenji Tatsuno and Shuichi Tsutsumi and Akinori Kanai and Toshiya Inaba and Yusuke Shiozawa and Yuichi Shiraishi and Kenichi Chiba and Hiroko Tanaka and Rishi S. Kotecha and Mark N. Cruickshank and Fumihiko Ishikawa and Tomohiro Morio and Mariko Eguchi and Takao Deguchi and Nobutaka Kiyokawa and Yuki Arakawa and Katsuyoshi Koh and Yuki Aoki and Takashi Ishihara and Daisuke Tomizawa and Takako Miyamura and Eiichi Ishii and Shuki Mizutani and Nicola K. Wilson and Berthold Göttgens and Satoru Miyano and Toshio Kitamura and Susumu Goyama and Akihiko Yokoyama and Hiroyuki Aburatani and Seishi Ogawa and Junko Takita},
  title = {{Multi-omics analysis defines highly refractory RAS burdened immature subgroup of infant acute lymphoblastic leukemia}},
  booktitle = {Nature Communications},
  year = {2022},
  doi = {10.1038/s41467-022-32266-4},
}
Figure