Advances in RNA editing in hematopoiesis and associated malignancies Free

To date, >170 RNA modifications have been identified across various RNA types, playing a crucial role in enhancing transcriptomic and proteomic diversity by regulating splicing, messenger RNA (mRNA) stability, codon decoding, and the functions of noncoding RNAs.¹ RNA editing is one of the most prevalent posttranscriptional RNA modification that serves to introduce changes to RNA sequences.² In vertebrates, RNA editing includes adenosine-to-inosine (A-to-I) editing, catalyzed by the ADAR family of enzymes, and cytidine-to-uracil (C-to-U), catalyzed by Apolipoprotein B mRNA editing catalytic polypeptide-like (APOBEC) family of enzymes.³ This review primarily focuses on A-to-I RNA editing, with a brief mention of C-to-U RNA editing.

The ADAR family comprises 3 members: ADAR1, ADAR2, and ADAR3.⁴ ADAR1 and ADAR2 are broadly expressed and share common functional domains, including 2 to 3 repeats of the double-stranded RNA (dsRNA)–binding domain and a catalytic deaminase domain.^5-7 ADAR3 is specifically expressed in the brain and has no documented deaminase activity.⁸ ADAR1 has 2 isoforms: an interferon (IFN)–inducible p150 isoform and a constitutively expressed p110 isoform.⁹ ADAR1 p110 is predominantly localized in the nucleus; however, studies have shown that it can shuttle between the nucleus and cytoplasm under specific conditions, although less dynamically than ADAR1 p150.¹⁰ The dynamic nucleocytoplasmic shuttling of ADAR1p150 is facilitated by a nuclear export signal located within helix α1 of its Zα domain.¹¹ 

The primary function of ADAR enzymes is to regulate immune responses by editing host RNA to mark it as “self” while leaving viral RNA unedited as "nonself." Unedited dsRNAs are recognized by immune sensors such as melanoma differentiation-associated protein 5 (MDA5) and dsRNA-activated protein kinase R (PKR), which trigger signaling cascades that lead to the production of IFNs and IFN-stimulated genes (ISGs).^12,13 ADARs can also facilitate immune evasion by introducing coding changes that promote tumor growth while concurrently shielding tumor cells by suppressing ISG-driven immune responses.^14-16 Additionally, ADARs have been implicated in regulating stress response¹⁷ and stem cell fate.^18,19 

Hematopoiesis is a finely regulated process orchestrated by lineage-specific transcription factors that direct the generation of mature cells from hematopoietic stem and progenitor cells (HSPCs). HSPC maintenance, self-renewal, differentiation, and lineage commitment are controlled by diverse posttranscriptional regulators.^20,21 Posttranscriptional perturbations are, therefore, closely associated with the development of hematologic malignancies.²² A key example is seen in the transition from pre–leukemic stem cells (pre-LSCs) to LSCs.^20,21,23 ADAR has been extensively studied in this context, being implicated in both normal and aberrant hematopoiesis.^24,25 Among the ADAR family, ADAR1 is the most well established in its roles, whereas ADAR2, although less studied, is gradually emerging as a molecule of increasing interest in leukemia research. Despite mounting evidence, a narrative review of ADAR’s role within the hematopoietic system is lacking. Here, we provide an overview of RNA editing in normal and abnormal hematopoiesis. We focus on critical RNA editing sites and RNA editing detection methods. We then highlight the latest evidence that show the relevance of RNA editing in normal and pathological hematopoiesis. Finally, we speculate on how RNA editing might be leveraged as a potential therapeutic strategy.

To dissect the function and implications of A-to-I editing in hematopoiesis, it is important to first understand the underlying mechanism of this RNA editing event (Figure 1). ADARs preferentially edit mismatched adenosines within imperfect dsRNA, converting them to inosines.^26,27 The inosine is then recognized as guanosine by the cellular translation and splicing machinery.²⁸ This editing activity is particularly prominent in Alu repeats in humans, which constitute ∼11% of the genome. In contrast, the mouse genome is devoid of Alu repeats.²⁹ 

The effects of A-to-I editing are impactful when it occurs in coding regions, because they can alter the amino acid sequence and modify protein function.^24,30-32 RNA editing events occur more frequently, however, in noncoding regions that consist of inversely oriented repetitive elements. RNA editing in introns or in exon-intron junctions can lead to aberrant splicing, generating new isoforms.^33-36 Additionally, A-to-I modification in the 3' untranslated region (3′ UTR) can affect mRNA translation or expression by influencing microRNA (miRNA) binding specificity.^37,38 In some cases, RNA editing occurs after splicing (mature mRNA) can also cause amino acid substitutions.³⁹ RNA editing in long noncoding RNAs can affect their secondary structure, abundance, stability, and interactions with miRNAs.^40-46 Gong et al have created a database, LNCediting, to catalog the functional roles of RNA editing in long noncoding RNAs.⁴⁷ 

RNA editing in primary microRNAs (pri-miRNAs) typically inhibits miRNA maturation^48-50 but can promote biogenesis in rare cases (eg, miR-197 and miR-203).^51,52 Editing in the seed sequence alters target gene recognition^53,54 and can also repress miRNA incorporation into RNA-induced silencing complex.⁵⁵ RNA editing in Alu repeats of circular RNAs can unwind the dsRNA structure, promoting linear mRNA formation.^56,57 In transfer RNAs (tRNAs), it enables wobble pairing for codon recognition.⁵⁸ Pathogenic viral RNAs also exploit RNA editing to influence replication and pathogenesis,⁵⁹ highlighting its role in diverse biological processes.

RNA editing can be quantified through various methods (Figure 2). Conventional approach analyze G/(G+A) or T/(T+C) peak height ratios in Sanger sequencing.⁶⁰ The emergence of deep sequencing enables transcriptome-wide identification of RNA editing sites.^61-64 Bulk-RNA sequencing, often using tools such as REDItools, maps the RNA editome across tissues.⁶⁵ Microfluidics-based multiplex polymerase chain reaction (PCR) followed by deep sequencing is a targeted sequencing method developed to enable uniform and simultaneous amplification of nearly 1000 RNA editing loci for numerous samples.^29,66 Single-cell RNA sequencing (scRNA-seq) data, although challenging due to lower quality and coverage, can identify RNA editing sites with high-depth, full-length data.⁶⁷ Millions of RNA editing events have been curated and annotated in databases such as REDIportal,⁶⁸ RADAR,⁶⁹ and REDH,⁷⁰ facilitating the exploration of the relevance between RNA editing loci, normal cellular processes, and associated disease pathogenesis.

To validate these RNA editing sites, sensitive and cost-efficient strategies known as site-specific quantitative PCR (qPCR) and digital PCR have been developed and proven effective in clinical samples.^32,71-75 Finally, specialized methods, such as 2-tailed reverse transcription qPCR (RT-qPCR) and RT-qPCR with locked nucleic acid primers, have been developed to analyze RNA editing in miRNAs.^76-78 Together, this comprehensive tool kit for identifying and quantifying RNA editing provides researchers with unparalleled opportunities to gain in-depth insights into RNA biology and disease mechanisms. In the following sections, we focus on our explorations into the RNA editing landscape in the context of hematopoiesis.

There is now extensive evidence to suggest a role for RNA editing in hematopoiesis (Table 1). Jameson et al showed that lentiviral overexpression of ADAR1 p150 in cord blood cells skews differentiation toward the myeloid lineage⁷⁹ and affects self-renewal and pluripotency genes.⁸⁰ Wang et al⁸¹ published a seminal paper showing that ADAR1 haploinsufficiency leads to embryonic lethality before embryonic day 14 (E14), likely due in part to observed hematopoietic system abnormalities. Hartner et al⁸² also observed lethality around E12 using ADAR1-null mutant model, with liver disintegration and major defects in definitive hematopoiesis. It seems that widespread stress-induced apoptosis occurs in ADAR1 knockout mice.⁸³ In the sections that follow, we discuss the role that ADAR1 plays during normal hematopoiesis.

Inspired by the gross animal-level effects of an ADAR1 knockout, researchers have conducted efforts to identify the cell-specific impact of ADAR1 ablation. For example, deletion of ADAR1 in fetal liver hematopoietic stem cells (HSCs) (Mx1-Cre-Adar^f/−) resulted in notably low engraftment.⁸⁴ ADAR1 deletion in adult HSCs (SCL-Cre-ER-Adar^f/−) revealed that ADAR1 is essential for preventing the exhaustion of long-term HSCs (Lin^–c-Kit⁺Sca-1⁺CD34^lo) and IFN signaling.⁸⁴ However, in our laboratory, using ER-Cre-Adar1^f/f mice, we showed that ADAR1 plays a preferential role in promoting the proliferation of hematopoietic progenitor cells (Lin^–c-Kit⁺Sca-1^–) compared with more primitive HSCs (Lin^–c-Kit⁺Sca-1⁺).⁸⁵ ADAR1-deficient HSCs successfully engrafted and were sustained for >6 months. This inconsistency may stem from variations in mouse models, gene excision efficiency, and the manipulation process.

ADAR1 also plays a crucial role in the development of differentiated hematopoietic cells. Apart from its role in fetal erythropoiesis,^81,86 ADAR1 knockout in Epor-Cre mice (Epor-Cre-Adar1^Δ/−) induces macrocytic anemia or splenomegaly in adult, whereas an editing-deficient model (Epor-Cre-Adar1^Δ/E861A) was underrepresented from around E15.5, indicative of a vital role of catalytic function of ADAR1 in murine erythropoiesis.⁸⁶ Results from LysM-Cre-Adar^Δ/Δ mice, however, suggested that ADAR1 was dispensable for myelopoiesis based on peripheral blood parameters.⁸⁶ In contrast, others later demonstrated that specific deletion of ADAR1 in myeloid cells using LysM-Cre-Adar^Δ/Δ mice led to significantly reduced numbers of alveolar macrophages.⁸⁷ The same mouse model also proved that the loss of ADAR1 in macrophages promoted an antitumor microenvironment in combination with IFN-γ.⁸⁸ In dendritic cells (DCs), CD11c-Cre-Adar1^–/– mice exhibited impaired CD103⁺ DC expansion.⁸⁷ Additionally, ADAR1-deficient antigen-presenting cells (CD11c-Cre^+/−Adar1^f/fR26^YFP) showed an increased number of inflammatory type-2 conventional dendritic cells (cDC2)-like cells, elevated activated tissue-resident memory T cells, and an antiviral gene signature.⁸⁹ 

Numerous studies have also investigated the role of ADAR1 in lymphocyte development (Table 1). Deletion of ADAR1 at the early T-cell stage (Lck-cre-Adar1^f/f) resulted in severe thymic atrophy and impaired transition from DN3 (CD44^-CD25⁺) to DN4 (CD44^-CD25^-) stage, along with reduced T-cell receptor β expression and increased expression of type I ISGs.^90,91 Aberrant thymic T-cell maturation and excessive ISG expression were also observed in Cd4-Cre-Adar1^f/f mice when Adar1 was deleted at the late T-cell stage.⁹² Luca et al⁹³ demonstrated that deletion of Adar1(Foxp3-Cre-Adar1^f/f) in T regulatory cells led to reduced numbers of peripheral T regulatory cells, which ultimately resulted in compromised regulatory functions and contributed to autoimmune disorder progression.

The ADAR1 p150 isoform has a critical role in B lymphopoiesis. Marcu-Malina et al⁹⁴ showed that CD19-Cre–mediated Adar1 ablation (in late B-cell stage) led to a significant reduction in immature and mature B cells, primarily due to increased apoptosis. Focusing on early B lymphopoiesis, Chen et al⁹⁵ used Mb1-Cre-Adar1^f/f mice (from pre to pro B cells onward) to prove that ADAR1 regulates the transition from pro-B to large pre-B cells. In Mb1-Cre-Adar1^f/fMD4^Tg/+ mice, in which transgenic B-cell development bypasses the early B-cell stage, ADAR1 plays a critical role in regulating pre–B-cell receptor expression. This regulation is crucial for the transition from large pre-B cells to small pre-B cells. Li et al⁹⁶ used Aicda-Cre-Adar1^f/f mice, in which ADAR1 was specifically deleted in activated B cells, to investigate the role of ADAR1 in adaptive immunity. They discovered that ADAR1-deficient mice had impaired T-cell–dependent antibody responses and a significant reduction in germinal center B cells. In summary, studies demonstrate that ADAR1 plays a critical role in T and B lymphopoiesis across multiple stages.

Certain key ADAR1-mediated RNA editing events play a pivotal role in regulating hematopoiesis (Table 2). A genome-wide analysis of ADAR1 RNA substrates in fetal liver during erythropoiesis revealed that 40% of RNA editing events occur in the long 3′ UTR of 3 genes: Klf1, Oip5, and Optn.¹³ These edited events generate inosine-uracil (I-U) mismatches and prevent the activation of inflammatory responses. While in editing-deficient knockin mutant mice (Adar1^E861A/E861A), unedited transcripts can be bound by MDA5 and/or retinoic acid inducible gene I (RIG-I), leading to the activation of the cellular dsRNA response and immune cascades.

Wang et al²⁴ performed a comprehensive study of the RNA editome in murine hematopoietic cells. Their study identified 8 genes that undergo A-to-I RNA editing specifically in HSPCs, including Azin1, Cog3, Taf1c, Rtkn, Igbp1, Rrp15, H19, and Mri1. Among them, Azin1 can modulate HSC differentiation through interaction with Ddx1. Fully edited Azin1(S367G) enhances long-term hematopoietic reconstitution. Interestingly, Azin1 editing has been observed in patients with hepatocellular carcinoma and prostate cancer, in which it promotes a tumorigenic phenotype through mechanisms involving Azin1-Ddx1 nuclear translocation.^30,97,98 This indicates that RNA editing at the same site can lead to divergent outcomes depending on the cell type.

Human HSPCs exhibit distinct RNA editing patterns that regulate lineage commitment and self-renewal. A study using single-cell RNA sequencing dataset identified several editing events linked to these processes.⁶⁴ In EIF2AK2, editing at 4 miRNA binding sites in the 3′ UTR disrupts miR-23a/miR-23b binding, increasing EIF2AK2 expression and activating the stress response pathway. This mechanism initiates global translational attenuation as a protective mechanism to maintain cellular homeostasis during HSPC differentiation.⁶⁴ RNA editing in genes such as CASP8 and ZBTB1 also plays a role in HSPC differentiation.

Given the critical role of RNA editing in normal hematopoiesis, the REDH database was developed to explore RNA editing events in both normal and abnormal hematopoiesis.⁷⁰ 

RNA editing events not only maintain steady-state hematopoiesis but also facilitate cellular adaptation to environmental stimuli such as temperature changes, hypoxia, inflammation, and aging.^99-101 For instance, APOBEC3A-mediated C-to-U editing occurs during M1 macrophage polarization and in monocytes exposed to hypoxia and IFN.^102,103 In the SDHB gene, a rapid increase in the C136U editing event creates a premature stop codon (R46X), inactivating the SDHB protein and potentially enhancing monocyte and macrophage responses to low oxygen.^102,104,105 

In natural killer cells and lymphocytes, hypoxic stress or reduced mitochondrial activity can trigger APOBEC3G-mediated RNA editing of multiple mRNA substrates, including TM7SF3, RPL10A, and RFX7, which may lead to translational reprogramming, altering protein expression profiles to respond to cellular stress.¹⁰⁶ Beyond C-to-U editing, ADAR1-mediated A-to-I editing can also be triggered by external factors.^99,107 Collectively, RNA editing helps cells better respond to environmental changes.

Aberrant RNA editing of specific transcripts has been implicated in various cancers, including hematologic malignancies.^108-110 We believe that elevated ADAR1, particularly p150, in hematologic malignancies shares a common mechanism. Specifically, in disease progression, inflammatory cytokines (interleukin-3 [IL-3], IL-6, IFN, and tumor necrosis factor -α) activate the JAK/STAT pathway, promoting ADAR1 p150 and other ISG transcription. This mechanism is validated in blast crisis (BC) chronic myeloid leukemia (CML)⁵⁰ and multiple myeloma (MM).¹¹¹ In MM, ADAR1 is amplified on 1q21, explaining its increased levels. Additionally, the IL6R-STAT3-ADAR1¹¹² and NOTCH/IL-6/ADAR1¹¹¹ axes sustain ADAR1 expression. In this section, we highlight the oncogenic roles of ADAR and site-specific RNA editing events involved in hematologic cancers (Table 3).

Several studies have explored the role of ADAR in acute myeloid leukemia (AML).^31,113 In 2000, researchers found that protein tyrosine phosphatase nonreceptor type 6 (PTPN6), known for splicing mutations in hematopoietic disorders, undergoes RNA editing in AML. An A-to-I edit at position 7866 in intron 3 leads to an extended isoform. This hyperediting in AML may cause PTPN6 haploinsufficiency, weakening its tumor suppressor role in cKit signaling.^33,114,115 Nishikura et al⁴⁸ discovered that pri-miR-142-3p was the first microRNA edited by ADAR1 p110 and ADAR2, blocking its maturation. Another study¹¹⁶ reported miR-142 downregulation in patients with AML, indicating its potential as a prognostic marker. A study¹¹⁷ on RNA editing events during the transition from pre-LSCs to LSCs in AML showed that the bone marrow’s proinflammatory microenvironment increased ADAR1 p150 expression, leading to higher editing in STAT3 intron 21. This produced a shorter form of STAT3β isoform,³⁵ which inhibits β-catenin degradation¹¹⁸ and helps LSCs maintain self-renewal.¹¹⁹ 

Meduri et al¹²⁰ analyzed The Cancer Genome Atlas (TCGA) AML data and found that SOCS2-AS1 editing is linked to poor survival negatively correlated with mRNA levels. They also observed reduced ADAR2 expression in AML, an unexpected phenomenon also confirmed in other studies.³¹ In this specific case, transcription of the ADAR2 gene is disrupted by the t(8;21) translocation. The reduction of ADAR2 levels results in the loss of RNA editing in mRNAs encoding components of the oligomeric Golgi complex 3 (Cog3) and coatomer subunit α (COPA). Overexpression of the edited forms inhibited the colony-forming ability of Kasumi-1 cells, suggesting ADAR2-mediated RNA editing plays a crucial role in preventing the development of therapy-resistant forms of core-binding factor AML.

Jamieson et al⁷⁹ first demonstrated the role of ADAR1 p150 in reprogramming malignant progenitor cells in CML. Subsequent studies have explored the involvement of ADAR1-mediated A-to-I editing in cell cycle transit, self-renewal, and the maintenance of LSCs. Overexpression of ADAR1 p150 induced myeloid skewing via PU.1 upregulation and created an aberrant spliced isoform (an in-frame splice deletion of exons 8 and 9) of glycogen synthase kinase 3β in CML progenitors, enhancing β-catenin expression and leukemia engraftment potential.¹²¹ Additionally, increased editing events in genes such as MDM2, APOBEC3D, GLI1, and AZIN1⁷² have been identified and validated from chronic-phase progenitors to BC AML, which have been associated with stem cell fate¹²² and DNA mutagenesis.¹²³ 

Zipeto et al⁵⁰ further proved the active role ADAR1 played in sustaining LSC self-renewal. Mechanistically, the activated JAK2/STAT and BCR-ABL1 (breakpoint cluster region-Abelson 1) signaling in CML converged on ADAR1 activation, mediating RNA editing of pri-let-7 miRNA family, thereby impairing their maturation. The reduced let-7 levels fail to target LIN28B, thus promoting self-renewal.^124,125 Jiang et al³⁷ found that editing at the DROSHA (human protein) cleavage site of pri-miR-26a inhibits miR-26a biogenesis, which normally suppresses CDKN1A via its target gene EZH2, thus promoting cell cycle progression and increasing self-renewal of CD34⁺ HSPCs. They also identified a ∼600-nucleotide region in the 3′ UTR of MDM2 that serves as a hot spot for RNA editing, where several miRNAs (such as miR-155) have their target sites. This editing increases MDM2 transcript abundance and accelerates the degradation of p53, ultimately promoting the transformation from chronic phase to BC CML.¹²⁶ 

In samples from patients with chronic lymphocytic leukemia (CLL), ADAR1 is presumed to interact with DROSHA, thereby inhibiting the biogenesis of pri-miR-16-1,¹²⁷ a validated tumor suppressor in CLL.¹²⁸ Notably, it is not ADAR1’s catalytic RNA-editing activity but rather its RNA-binding domain and nuclear localization domain that are responsible for affecting miRNA processing. Furthermore, a comprehensive analysis of miRNA editing in CLL and normal B cells identified 4 clinically relevant miRNAs that undergo robust RNA editing in CLL samples: miR-184, miR-589, miR-3157, and miR-6503. These edited miRNAs contributed to a dysregulated mRNA network.¹²⁹ 

In addition to the miRNA editome analysis, researchers conducted matched RNA sequencing and whole-exome sequencing of samples from patients with CLL. They identified 19 recurrent editing sites in genes such as COG3, CDK13, and COPA and revealed that RNA editing patterns are specific to CLL cohorts with different clinical outcomes. Moreover, ADAR knockout increased the sensitivity of CLL cell line MEC1 to treatments such as fludarabine and ibrutinib in vitro. These findings highlight ADAR as a promising target for future combination treatment strategies in CLL.²³ 

Jiang et al recently uncovered the critical role of ADAR1 in patients with relapsed T-cell acute lymphoblastic leukemia (T-ALL).¹³⁰ ADAR1 maintained leukemia-initiating cell survival and self-renewal by mediating malignant A-to-I editing, which suppresses MDA5-directed dsRNA sensing. Loss of ADAR1 not only alleviated the leukemia burden in a xenograft model derived from patients with T-ALL but also reduced hyperediting events, such as those in the 3′ UTR of MAVS and IL17RA and in the intronic region of MYB.¹³⁰ These findings suggest that targeting of ADAR1 could be a promising therapeutic strategy for eradicating T-ALL leukemia-initiating cells.

RNA editing events have been linked to subtype-specific non-Hodgkin lymphoma regulation.¹³¹ Wayne et al³⁴ first elucidated a potential role for RNA editing in lymphoma development via a mutational analysis of PRDM1. PRMD1 editing leads to its premature translation termination, which partially explains the pathogenesis of diffuse large B-cell lymphoma (DLBCL), given that PRDM1 is a well-known transcription suppressor participating in plasma cell differentiation.¹³² 

Pecori et al¹³³ studied ADAR1’s role in immune evasion in DLBCL by profiling genomic and transcriptomic data from 106 patient samples. They identified numerous edited or mutated genes involved in the apoptosis/p53 signaling pathway, NF-κB, and RIG-I–like receptor signaling pathways. An intriguing candidate is the RNA editing of MAVS 3′ UTR, which positively correlated with increased MAVS protein levels, inflammation cascade activation, and T-cell exhaustion, predicting poor prognosis in affected patients. Most excitingly, restoring MAVS editing via targeted base editing in ADAR-deficient RC-K8 cells (DLBCL-derived cell line) enhanced NF-κB and type I IFN signaling cascades.

RNA editing influences the life cycle of Kaposi sarcoma–associated herpesvirus (KSHV), the cause of primary effusion lymphoma. In this context, editing in the stem region of pri-miR-K12-4-3p affects miRNA maturation, whereas changes in the seed region alter target specificity, promoting more efficient KSHV infection. These findings deepen our understanding of KSHV-induced primary effusion lymphoma.¹³⁴ 

A characteristic feature of MM is the amplification of chromosome 1q21, in which the ADAR1 gene is located.^135,136 This amplification frequently leads to ADAR1 overexpression and is associated with increased RNA editing events.¹³⁷ Notably, GLI1, a key Hedgehog (Hh) pathway activator, is edited in relapsed MM and plasma cell leukemia,³² enhancing Hh signaling and contributing to therapeutic resistance.^138-140 Additionally, ADAR1 editing of NEIL1, a base-excision repair gene, boosts MM cell proliferation, reduces DNA repair, and increases drug resistance to standard MM therapies.^131,141 

In a comparative analysis of whole-exome sequencing and RNA sequencing data, Tasakis et al¹⁴² further illustrated the dual role of ADAR1 in facilitating MM progression, which is driven by a combination of increased RNA events and the introduction of novel DNA mutations. Finally, among the edited cancer-related targets, Mdm4 and Eif2ak2 have significant roles in MM by interacting with the p53 and NF-κB pathways, respectively. Although these transcripts have been identified as edited in relevant databases, the specific mechanisms underlying the editing effects remain unclear.^68,69 

Studies have also explored RNA editing patterns in pediatric hematologic malignancies and other blood cancers, although we will not delve into those details here.¹⁴³ 

Although the role of A-to-I editing in blood cancers is well documented, APOBEC-mediated C-to-U editing has also been implicated in malignancies. For instance, activation-induced cytidine deaminase (AID)-mediated mutations contribute to B-cell lymphoma, and overexpression of AID has been shown to induce T-lymphoma or B-leukemia/lymphoma.^144,145 APOBEC3A overexpression has been associated with C-to-U editing of SF3B1 and KMT2A in hematologic neoplasms.¹⁴⁶ Furthermore, APOBEC3B is upregulated in lymphoma cell lines,¹⁴⁷ and APOBEC3G acts as a prosurvival factor that sensitizes lymphoma cells to radiation therapy.¹⁴⁸ 

As outlined in this review, growing evidence suggests that dysregulated ADAR activity and RNA editing events are involved in both hematopoiesis and abnormal hematopoiesis (Figure 3), particularly in cancer progression, therapy resistance, and poor prognosis.

Indeed, numerous studies have shown that ADAR1 knockdown impairs the self-renewal capacity of malignant cells,^50,130 and ADAR1 deletion can lead to regression of AML and CML in murine models^149,150 or treatment sensitivity in CLL cell lines.²³ These findings open avenues into potential treatment approaches that might involve (1) the selective inhibition of ADAR activity or (2) the prevention of dysregulated A-to-I editing at specific sites. We discuss these 2 potential approaches in the following sections.

Despite considerable efforts, there are currently no US Food and Drug Administration–approved drugs targeting ADAR. For example, adenosine analogs such as 8-azaadenosine (8-aza-A) and 8-chloroadenosine (8-Cl-A) can suppress ADAR activity in various cancer contexts.^50,151 However, results showed that these inhibitors are neither selective for ADAR nor effective in inhibiting A-to-I RNA editing of ADAR substrates.¹⁵² ZYS-1 is a newly synthesized small molecule that targets the deaminase activity of ADAR1, showing promising antitumor efficacy and favorable safety for patients with prostate cancer.¹⁵³ It may be worthwhile to explore the application of ZYS-1 in hematologic malignancies. Another promising inhibitor, Rebecsinib, has shown potential in treating high-risk myelofibrosis and secondary AML. It prevents the expression of the ADAR1 p150 splice isoform and blocks malignant A-to-I editing–mediated LSC self-renewal, while sparing normal HSPCs.⁷³ Additionally, research is ongoing into strategies for inhibiting ADAR deaminase activity,¹⁵⁴ marking this as a promising therapeutic area. RNA editing inhibition may treat blood malignancies, but side effects on hematopoiesis and immunity must be carefully considered, with an optimized therapeutic window balancing efficacy and safety.

Antisense oligonucleotide (ASO) therapy, designed to block RNA editing by targeting complementary sequences, shows promise for correcting disease-causing mutations. Chemically modified oligonucleotides enhance stability, binding affinity, specificity, and reduce nuclease cleavage and immune response. For instance, ASOs with 2’-O-methyl/locked nucleic acid modifications have shown efficacy in inhibiting NEIL1 editing.¹⁵⁵ Meanwhile, ASO3.2 has been shown to abolish AZIN1 editing without affecting splicing and translation and inhibits cancer cell viability in vitro or in patient-derived xenograft models.¹⁵⁶ Progress in this area is restrained, however, given the promiscuous RNA editing events that occur in cancer context; halting disease progression by targeting only a few specific RNA editing sites remains unlikely. Consequently, further studies are needed on the application of site-specific RNA editing inhibitors.

RNA editing holds significant potential for clinical diagnostics, with studies exploring its application in diseases such as neurological disorders and cancers.¹⁵⁷ Efforts to diagnose hematologic malignancies have focused on detecting miRNA editing in plasma and extracellular vesicles from patients with myelodysplastic syndrome.¹⁵⁸ A study found increased RNA editing in early-stage myelodysplastic syndrome, but only a few miRNAs showed edited forms with low rates of A-to-I changes, highlighting the need for improved detection sensitivity.

ADARs, with their ability to bind dsRNA and perform A-to-I deamination, have potential therapeutic applications.¹⁵⁹ Recent studies have explored by the use of antisense guide RNAs to harness endogenous ADARs^160-163 or delivering exogenous ADARs for targeted RNA editing.^162,164 Although DNA editing can provide durable and potentially permanent cures for certain diseases,¹⁶⁵ RNA-targeted modifications without genome integration offer flexible and temporary changes, making them safer for conditions requiring transient effects, such as acute pain, viral infections, obesity, and inflammation.¹⁶⁶ 

To further explore the therapeutic potential of RNA editing, we queried the ClinVar website and identified nearly 14 000 G>A missense and nonsense mutations among pathogenic single-nucleotide variants to date, which represent prime candidates for targeting and correction to the wild-type sequence using ADAR-mediated RNA editing therapy.^166,167 Of course, the candidate mutations are not limited to G>A mutations, given the redundancy of codons in the genetic code. This flexibility allows for various types of nucleotide substitutions to be targeted for therapeutic correction, as demonstrated by successful practices in the treatment of inherited retinal diseases.¹⁶⁸ 

In addition to their RNA editing–dependent effects, ADARs can also exert RNA editing–independent functions by binding to DNA or interacting with proteins involving DORSHA, DICER1, and HuR.⁵⁶ Furthermore, the Zα domain of ADAR1 can bind to the G-quadruplex in the c-Myc promoter.¹⁶⁹ Given the critical role of c-Myc in regulating HSPC function and hematopoietic malignancies,¹⁷⁰ this interaction warrants further investigation.

Future research targeting RNA editing enzymes or specific editing events holds promise for advancing cancer treatment. We envisage RNA editing therapy as a promising tool for correcting disease-causing mutations by leveraging endogenous or engineered ADAR enzymes. First, we need to overcome limitations in delivery systems, target specificity, and off-target effects and take time to explore the molecular mechanisms governing RNA editing efficiency and accuracy to better understand how to safely and effectively harness this technology for therapeutic applications.

This work was supported by the Ministry of Science and Technology of China (grant number 2021YFA1100900), the National Natural Science Foundation of China (grant numbers 92468206 and 92368202), the Chinese Academy of Medical Sciences (CAMS) Innovation Fund for Medical Sciences (grant numbers 2021-I2M-1-040 and 2021-I2M-1-019), the CAMS Fundamental Research Funds for Central Research Institutes (grant number 3332021093), and the Haihe Laboratory of Cell Ecosystem Innovation Fund (grant number 24HHXBSS00006).

Contribution: S.P. wrote the initial draft of the manuscript and conducted the literature review; and H.C. and T.C. provided revisions and helped refine the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Hui Cheng, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, 288 Nanjing Rd, Heping District, Tianjin 300020, China; email: chenghui@ihcams.ac.cn; and Tao Cheng, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, 288 Nanjing Rd, Heping District, Tianjin 300020, China; email: chengtao@ihcams.ac.cn.