Acute promyelocytic leukemia (APL) is successfully treated with PML-RARa-targeted therapies. APL is divided into two groups with distinct PML-RARa subtypes: Long/variant (the L/V type) and short (the S type) isoforms. The L/V type PML-RARa is observed ~70% of APL patients, whereas the S type isoform, which lacks nucleus localization sequence located in exon6 of PML gene, is involved in ~30% of APL cases. PML-RARa contributes to leukemogenesis presumably by repressing transcriptional activities via its DNA binding properties in nucleus. However, recent studies have shown that PML-RARa mainly localizes in nucleus in the L/V type APL, whereas it dominantly distributes in cytoplasms in the S type APL. This study was planned to clarify difference in the role of PML-RARa in leukemogenesis in APL patients with the L/V or the S type.

We performed the transcriptome analysis of APL blast cells from 4 cases each of the L/V type and the S type. Surprisingly, APL cells of the L/V and the S type showed significantly different gene expression pattern by gene set enrichment analysis. For example, pathways for apoptosis involving BCL-2, prostaglandin and T lymphocyte maturation were enriched only in the S type APL, whereas cytokine pathways were mainly enriched in the L/V type APL. More interestingly, the gene sets associated with functional leukemic stem cells (LSCs) (K.Eppert, et al. Nat Med, 2011) were enriched in the S type APL (NES 1.72, FDR q-value 0.083).

In most AML of FAB types except for M3 (APL), self-renewing LSCs are concentrated in the CD34+CD38- population, and transplantation of CD34+CD38- LSCs of AML patients resulted in reconstitution of human AML in xenograft models. Based on this data, it is widely accepted that AML is hierarchically organized and is maintained by LSCs, like normal hematopoiesis is initiated from normal HSCs. However, it has been reported that reconstitution of human APL is usually difficult in xenograft models (D.Thomas and R.Majeti. Blood, 2017), although human APL was reconstituted in rare cases by a recent report (S.Patel, et al. Leukemia, 2012). We have reported that self-renewing AML LSCs express TIM-3, a signaling molecule activating the β-catenin pathway (Y.Kikushige, et al. Cell Stem Cell, 2015), but TIM-3 expression was negative in leukemic cells in most cases of APL. However, because the S type but not the L/V type APL has LSC signatures by transcriptome analysis, we revisited to test the LSC activity of the S type APL.

By FACS analysis, CD34+CD38- cells were rarely found in the L/V type APL patients, and they largely lacked TIM-3 expression and did not have PML-RARa transcript (n=10), suggesting that these CD34+CD38-TIM-3- cells may be residual normal hematopoietic stem cells (HSCs). In contrast, CD34+CD38- cells from the S type APL strongly expressed TIM-3 and PML-RARa transcript (n=12), suggesting that they may be functional LSCs. We then tested whether CD34+ cells of the S type APL can reconstitute human APL in a xenograft model. Purified 0.4-3.0 x106 CD34+ APL cells were transplanted into sublethally irradiated NSG mice. Strikingly in the S type APL, CD34+ but not CD34- cells reconstituted human PML-RARa+ APL in NSG mice in 3 out of 5 patient tested, whereas in the L/V type APL, CD34+ cells were incapable of reconstitution in all 3 cases tested. The hCD45+CD34+TIM-3+ cells harvested from the primary NSG recipients again reconstituted the S type APL in serial xenotransplantion experiments. These results collectively suggest that the S type APL have LSCs capable of self-renewal within the CD34+ leukemic population, and may use TIM-3 signaling for self-renewal as in other FAB types of AML.

Thus, APL with the S type, but not the L/V type PML-RARa have expression signature related to LSCs, and have functional LSCs with TIM-3 expression. Therefore, the leukemic hematopoiesis of APL with the S type is hierarchically organized as in case of AML of other FAB types. The mechanisms by which these 2 types of PML-RARa exert different functions to the LSC property should be critical to understand pathogenesis of APL.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution