In this issue of Blood, Cheng and colleagues report first on intriguingly complex pre–messenger RNA (pre-mRNA) global alternate splicing (AS) events that occur in primary erythroblasts. Predominant regulatory roles then are implicated for a Muscleblind-like (MBNL) splicing factor, Mbnl1 (an AS regulator in pluripotent embryonic stem [ES] cells and myoblasts), and for nuclear distribution element like-1 (Ndel1) as a Mbnl1 target. Via loss-of-function (LOF) analyses, Mbnl1 and Ndel1 further are evidenced to support erythroblast development.1
Editing of primary gene transcripts to yield variant exon-included vs -excluded mRNA forms is an important mechanism for generating protein isoforms with diversified activities.2 To illustrate, one established hematologic example of such AS is FAS in which exon-6 inclusion yields a membrane-anchored proapoptotic ligand, whereas exon-6 excision generates soluble apoptosis antigen-1 with prosurvival properties.3 In higher eukaryotes, the majority of transcripts may, in fact, be subject to AS,2 a process which is regulated by one of the most sophisticated of subcellular machines, the spliceosome (snRNPs U1-U6, plus ∼200 polypeptides).4 In leukemogenic contexts, heightened attention to the spliceosome recently has arisen from associated newly discovered mutations, especially a SF3B1 component. Specifically, SF3B1 is mutated at substantially heightened frequencies in both myelodysplastic syndrome and chronic lymphocytic leukemia.4 Although more work is needed to establish functional significance, SF3B1’s knockdown in hematopoietic progenitors also generates ring sideroblasts.
Preclinical investigations also are providing new insight into selective regulators of AS and functional consequences. Here, major advantages have been provided by the advent of approaches such as RNAseq and CLIP-seq5 as used to define RNA AS cis motifs together with factors acting at exon-intron borders. For erythropoiesis, global analyses of AS have been lacking to date. The present studies by Cheng et al use a murine fetal liver erythroid progenitor cell model first to identify ∼600 AS events for transcripts with diverse gene ontologies.1 Beyond this, via analyses of pentameric regions adjacent to splice sites (and comparisons to published CLIP-seq binding clusters), enrichment among erythroid AS events is defined for clusters predicted to bind MBNL splicing factors (Mbnl 1-3).6
Muscleblind is an AS factor that is involved in selective programs of splicing decisions (and was discovered in Drosophila as a regulator of muscle) and photoreceptor differentiation.6 In human myotonic dystrophies (DMs), MBNLs associate with abnormal 3′ CTP repeats in nuclear pre-mRNAs encoding DM protein kinase, a component of a DM toxic RNA model.6 Mice harboring Mbnl1−/− plus Mbnl2+/− alleles also recently have been shown to exhibit cardinal features of DM muscle disease.7 In additional contexts, extended roles of MBNLs are illustrated by regulation of FOXp1 AS in a context of ES cell pluripotency,8 and in insulin receptor transcript AS.6 In studies by Cheng et al,1 Mbnl1 was expressed in fetal liver erythroblasts, and its knockdown led to increased blast cell size, attenuated hemoglobinization, and decreased frequencies of enucleated erythrocytes, indicative of hindered differentiation in the absence of Mbnl1. Upon ectopic expression of Mbnl1 (exon inclusion form), this LOF phenotype further was partially reversed.
To initially seek Mbnl1 targets, gene expression patterns and published CLIP-seq analyses were analyzed, with Ndel1 pre-RNA as one confirmed target that also bound Mbnl1 in RNA immunoprecipitation and splicing reporter experiments. Ndel1 is a conserved coiled-coil protein (and apparent thiol-activated peptidase) with intriguing binding partners. These include Lis1 as mutated in lissencephaly, and DISC1 as a high-risk factor for schizophrenia.9 In proerythroblasts, Cheng et al go on to show that Ndel1 knockdown inhibits the formation of enucleated red cells, and that a 3′ exon inclusion form (but not exclusion form) partially rescues this knockdown phenotype.1
Overall, work by Cheng et al provides novel insight into global AS events during erythropoiesis, and specifically so in murine fetal liver erythroid progenitors as a robust model system.1 Selective utilization of splicing regulators that target specific sets of AS events also is implicated, with Mbnl1 and Ndel1 as initially defined players. Among Mbnl1-3, additional roles for Mbnl2 in adult bone marrow erythropoiesis will be of interest to define in future studies. For Ndel1, possible effects on erythroblast growth, cytoskeletal features, and/or differentiation also merit extended attention. In a broader scope, RNAseq analyses also have been recently reported for developing human (pro)erythroblasts,10 and such contributions should enable further insight into AS regulation, specificity, and function during normal and dysregulated human erythropoiesis.
Conflict-of-interest disclosure: The author declares no competing financial interests.