Focusing on frequent ASXL1 mutations in myeloid neoplasms, and considering rarer ASXL2 and ASXL3 mutations

In humans, additional sex combs-like (ASXL) genes consist of three family members which are involved in epigenetic regulation. Mutations in the family of ASXL proteins have been identified in increased frequencies in myeloid neoplasms. Recently Alvarez-Argote et al. published a comprehensive review of the significance of somatic mutations in one member of this family of genes, ASXL1, in myeloid malignancies¹. In this review the authors describe how ASXL1 mutations are commonly seen in myeloid neoplasms including chronic myelomonocytic leukemia (CMML), acute myeloid leukemia (AML), myelodysplastic syndrome (MDS) and myelodysplastic/myeloproliferative neoplasm (MDS/MPN). Additionally, these mutations are predominantly frameshift and nonsense mutations and across all myeloid malignancies are associated with worse prognosis and higher rate of relapsed disease.

While we are becoming aware that the ASXL1 gene is critical in the progression and initiation of myeloid malignancies, this gene is only one in a larger family. Indeed, the human ASXL gene family consists of three genes in total: ASXL1, ASXL2, and ASXL3. ASXL gene family members are all epigenetic regulators potentially interacting with proteins including BAP1, EZH2, NCOA1, nuclear receptors, and WTIP to target specific genomic loci via histone modifications². Members of this family share a common domain architecture including highly conserved ASXN and ASXH domains at the N-terminal region, the ASXM1 and ASXM2 domains, and a plant homeodomain (PHD) finger at the C-terminal region. Most ASXL1 mutations in myeloid neoplasms are nonsense point mutations or frameshift mutations that result in functional loss of the PHD domain. The PHD domain of ASXL family members has been suggested to bind histones or DNA-binding modules of transcription factors based on homology studies, but functional studies are few and require more study to understand this proposed relationship. Similar to ASXL1 mutations, mutations in ASXL2 and ASXL3 are also predominantly frameshift or nonsense mutations.

ASXL2 is present on human chromosome 2p23.3 and is implicated to play a role in cell cycle regulation, cell senescence, cardiovascular and bone development, and adipogenesis^3–5.

Mutations in ASXL2 are found in developmental disorders and in a variety of malignancies including acute myeloid leukemia, and cancers of the bladder, prostate, breast, and pancreas^2,6. In the pediatric population, ASXL2 mutations are limited to hematologic malignancies with a predominance of AML with mutations in core binding factors⁷. A larger study expanding on this finding demonstrated that 23% of AML with t(8;21)(q22;q22);RUNX1-RUNX1T1 gene fusions harbor ASXL2 mutations, which makes it the second most common gene to be mutated after the KIT gene⁸. Interestingly mutations in ASXL2 are rare in the other core-binding factor AML, AML with inv(16)(p13q22);CBFB-MYH11. Additionally, ASXL2 mutations may be frequently associated with del(9q) but never co-occur with trisomy 8⁹. Unlike ASXL1 mutations, which are predominantly present in patients over 60 years, ASXL2 mutations are present in similar frequencies amongst pediatric and adult cases. Similar to cases with ASXL1 mutations, patients with ASXL2 mutations had no significant change in overall survival but had an increased rate of relapse compared to patients with wildtype ASXL1/2 with event free survival rate of 36% versus 25%^9,10.

Relatively little is known of the newest member of the ASXL gene family ASXL3, which is present on 18q12.1. Compared to ASXL1 and ASXL2, ASXL3 is significantly larger and shows more phylogenetic divergence and has a unique proline rich region adjacent to the PHD domain². Genetic studies suggest that ASXL3 and ASXL1 share some functional redundancy, as de novo truncating mutations in ASXL3 result in a developmental disorder that resembles Bohring–Opitz syndrome, which is also associated with de novo truncating mutations in ASXL1¹¹. ASXL3 mutations are present in a variety of malignancies, including 10% of melanomas and lung adenocarcinoma, but its role in hematologic malignancies is unclear^2,11. ASXL3 is expressed in the bone marrow¹¹ but ASXL3 mutations are rare in myeloid neoplasms: one patient with an ASXL3 mutation was identified in the AML TCGA study¹² and one case of pediatric AML with a point mutation clone that significantly expanded from less than 1% at initial diagnosis to 30% at relapse has also been identified¹³. Unlike ASXL1/2, ASXL3 mutations are rare in AML with t(8;21), where a single case was identified in a cohort of 110 cases¹⁴. Interestingly, this patient harbored a concurrent ASXL1 mutation, which contrasts with the finding that ASXL1 and ASXL2 mutations are mutually exclusive¹⁰.

Mutations among ASXL family members, despite having structural similarities, are not equally distributed amongst myeloid neoplasms. Overall ASXL1 mutations are much more common and seen in wide range of myeloid neoplasms as comprehensively reviewed by Alvarez-Argote et al., while current studies show only a single case of ASXL2 mutation in CMML and no ASXL3 mutations in MDS or MPN (Table 1)¹⁵. In fact, ASXL2 mutations are documented almost exclusively in AML with t(8;21)(q22;q22);RUNX1-RUNX1T1 and may be functionally linked to the RUNX1-RUNX1T1 gene fusion, since ASXL2 mutations and RUNX1 point mutation or inversion of chromosome 16 are mutually exclusive¹⁶. In some respects this is similar to ASXL1 where mutations are seen in AML with t(8;21)(q22;q22);RUNX1-RUNX1T1 but not in AML with inv(16)(p13q22);CBFB-MYH11; however, dissimilar to ASXL2, ASXL1 mutations may be associated with RUNX1 mutations^17,18,19,20. ASXL1 and ASXL2 mutations are also mutually exclusive in t(8;21) AML, suggesting that the two proteins may have functional redundancy though concurrent ASXL1/3 mutations have been documented in other instances^10,14. Indeed, in at least one in vitro study ASXL1 and ASXL2 were demonstrated to competitively bind BAP1 and this interaction believed to be crucial for the stability of ASXL2; however, this possible functional relationship is not present across all studies^21,22.

Table 1. Frequency of ASXL1, ASXL2 and ASXL3 mutations in myeloid malignancies.

Disease	ASXL1 mutations	ASXL2 mutations	ASXL3 mutations
AML (all types)	5–15%	1–2%	<1%
AML-MRC	37%	None described	None described
AML with t(8;21)	11.5%	23%	<1%
MDS (all types)	10–20%	None described	None described
CMML	40–50%	<1%	None described
PMF	20–25%	None described	None described
PV	7%	None described	None described
ET	4%	None described	None described

Adapted from references^{1,10,12,14,15,17,18}.

AML, acute myeloid leukemia; AML-MRC, acute myeloid leukemia with myelodysplasia related changes; AML with t(8;21), acute myeloid leukemia with t(8;21)(q22;q22); RUNX1/RUNX1T1, MDS, myelodysplastic syndrome; CMML, chronic myelomonocytic leukemia; PMF, primary myelofibrosis; PV, polycythemia vera; ET, essential thrombocythemia.

The work by Alvarez-Argote et al. highlights the importance of ASXL1 and also indirectly calls to attention to the entire family of ASXL genes. Further studies are needed to answer critical questions, some that the authors themselves raise such as: 1) what is the significance of mutations in ASXL1 with regard to patients treated with stem cell therapies, 2) what is the mechanism whereby ASXL1 induces oncogenesis, 3) why are ASXL1 mutations more ubiquitous than ASXL2 and ASXL3 mutations in myeloid neoplasms, and finally 4) what is the mechanistic role of ASXL2 and ASXL3 in oncogenesis? The fact that mutations in all ASXL family members are present in a range of malignancies suggests that these proteins, and the biological pathways they affect, likely play a fundamental role in oncogenesis; however, further functional studies are necessary²³.

Jean S. Oak and Robert S. Ohgami Stanford University, Stanford, CA, USA [email protected]

Transparency

Funding

This editorial was not funded.

Declaration of financial/other relationships

J.S.O. and R.S.O. have disclosed that they have no significant relationships with or financial interests in any commercial companies related to this study or article.

CMRO peer reviewers on this manuscript have no relevant financial or other relationships to disclose.