The miR-144/451^eGFP allele, a novel tool for resolving the erythroid potential of hematopoietic precursors

The study of blood cell formation has for many years spearheaded the investigation of adult stem cells. In the bone marrow an enigmatic, self-renewing, and rare hematopoietic stem cell (HSC) population gives rise to all blood cell lineages.¹  Traditionally, the study of hematopoietic stem and progenitor cells has profited from the relative ease of transplanting small amounts of stem cells into lethally irradiated recipients. Usage of mouse transplantion models enables the in vivo long-term qualitative and quantitative analysis of hematopoietic lineage determination.²  The CD45/Ly5 congenic strains constitute the most widely adopted transplantion mouse model for tracking lineage output in peripheral blood.³  However, the Ly5 antigen is expressed exclusively on nucleated cells, precluding the evaluation of donor-derived contribution to the erythrocytic and thrombocytic lineages. Several other transplant models are available that take advantage of naturally occurring isoforms (eg, Hbb^s/d, Gpi1^a/b),^4,5  as well as transgenes under control of ubiquitous promoters (eg, Actb^GFP, H2B^GFP, and H2K^b-GFP).^6-8  However, these methods have proved to be unsuitable or relatively inefficient in assessing erythroid output because of an absence of expression in mature RBCs or laborious methods of detection.

The vertebrate-specific miR-144/451 cluster, encoding miR-144 and miR-451, regulates erythroid development, homeostasis and susceptibility to oxidative stress.^9-11  The maturation and accumulation of miR-451 is mediated by a unique biogenesis pathway that is strictly dependent on Ago2's endonuclease activity.^12-14  Here we report a novel erythroid reporter allele based on the miR-144/451 noncoding RNA locus.

Cells were stained and analyzed and/or sorted using FACSCanto or FACSAria cytometers (BD Biosciences) as described.^15-23  Where appropriate, RBCs were lysed with an NH₄Cl-based solution (ACK lysis buffer) and nonspecific binding was reduced by preincubation with Fc receptor block (FcγII/III). Dead cells and duplets were excluded by 7-aminoactinomycin D (Sigma-Aldrich) staining and FSC/SSC profile, respectively. Data were analyzed off-line using FlowJo Flow Cytometry analysis software (TreeStar Inc). Supplemental Tables 1 and 2 detail surface marker profiles and antibody information (available on the Blood Web site; see the Supplemental Materials link at the top of the online article).

Young (8-12 weeks) female C57Bl/6 Ly5.1/Ly5.2 recipient mice were lethally irradiated in a single dose (875 Rad). Donor bone marrow cells were isolated from age- and sex-matched (8-10 weeks) congenic Ly5.1 or miR-144/451^+/eGFP: Ly5.2 mice and 2 × 10⁶ nucleated cells were transplanted by tail vein injection. Bone marrow reconstituted mice were maintained on medicated water (Baytril; Bayer) for 4 weeks after reconstitution and peripheral blood was analyzed 8 weeks after transfer.

The miR-144/451 cluster represents a promising candidate locus to engineer a reporter for specifically marking terminal differentiated erythrocytes, and erythroid lineage-committed hematopoietic precursors. Firstly, expression of miR-144/451 are dependent on the erythroid transcription factor GATA-1.²⁴  Secondly, they are restricted to the erythroid lineage.⁹  And finally, miR-451 accumulates to high levels in developing erythroblasts indicating extensive transcriptional activity of the locus.^9,14  To investigate the expression pattern in detail, we generated mice carrying an enhanced green fluorescent protein (eGFP) reporter cassette in the stem loop of precursor-miR-144 (Figure 1A, supplemental Figure 1). The insertion of the cassette disrupted expression of both miR-144 and miR-451 from the locus (supplemental Figure 1). As expected, homozygous miR-144/451^eGFP/eGFP mice displayed the same phenotype as the previously described miR-144/451^−/− animals⁹  (supplemental Figure 2). However, despite disruption of expression from one allele of the miR-144/451 locus and concomitant decrease in miRNA dosage, heterozygous miR-144/451^+/eGFP animals did not show any physiologic and erythropoietic abnormalities (supplemental Figure 2). More importantly, we found that expression of the reporter faithfully distinguished hematopoietic cells expressing miR-144 and miR-451. Strong reporter expression was observed exclusively in Ter119⁺ erythroblasts among terminal differentiated blood cell lineages and conversely, miR-144 and miR-451 expression fractionated with eGFP expression in sorted miR-144/451^+/eGFP bone marrow cells (supplemental Figure 3).

We next wanted to perform an in-depth analysis of reporter expression in miR-144/451^+/eGFP animals. Reporter expression was absent in the HSC population and all lymphoid- and GM-committed precursors in the bone marrow, as well as in developing T-cell progenitors in the thymus (supplemental Figure 4). Instead, we could detect minor reporter expression in the common myeloid precursor (CMP) and pre-megakaryocyte-erythrocyte precursor (PreMegE) populations (Figure 1B-C). Notably, as cells committed to the erythroid lineage by proceeding through PreMegE, PreCFU-E, and finally CFU-E stages, a dramatic increase in eGFP expression from ∼ 9% to ∼ 93% of the population was seen (Figure 1C). On the contrary, the low level of eGFP expression at the PreMegE stage was completely extinguished as cells differentiated further into megakaryocytic progenitors (MkP; Figure 1D). The strong reporter expression observed at the CFU-E stage was fully maintained as cells went through terminal erythroid differentiation in the bone marrow (Figure 1E). Finally, we bled the mice and examined the peripheral blood for traces of eGFP. In agreement with the absence of reporter expression in the megakaryocytic precursors, also fully differentiated platelets were GFP-negative (Figure 1F). Astoundingly, when examining mature erythrocytes in the peripheral blood we found that all cells (> 99%) were GFP-positive (Figure 1G-H).

To examine whether the properties of the miR-144/451^+/eGFP hematopoietic system could be faithfully recapitulated on bone marrow transplantation we took advantage of the widely used CD45/Ly5 transplantion model. Lethally irradiated recipients were injected with either control (Ly5.1) or experimental (miR-144/451^+/eGFP; Ly5.2) bone marrow followed by peripheral blood analysis 8 weeks after transplantation. Similar to WT, miR-144/451^+/eGFP cells could fully reconstitute the GM lineage and the B- and T-cell lineage (Figure 2A-B). Importantly, the reconstitution with miR-144/451^+/eGFP cells also resulted in > 99% GFP-positive erythrocytes in the peripheral blood of the recipients (Figure 2C). Next, we wanted to examine the quantitative nature of the miR-144/451^+/eGFP reporter in a competitive transplantation assay. Bone marrow transplantations were performed with diverse mixtures of control (Ly5.1) or experimental (miR-144/451^+/eGFP; Ly5.2) cells (Figure 2D). Remarkably, analysis of peripheral blood in 4 independent experiments revealed a near-perfect linear correlation (r² > 0.9) between the relative contribution of leukocytes, as measured by Ly5.2 chimerism, and GFP-positive RBCs (Figure 2E). The 2 parameters are directly proportional as the slopes approximate a value of 1 in each experiment (Figure 2E). Hence, the relative fraction of GFP-erythrocytes can be used as a novel direct measure of the HSC output in the erythroid lineage.

In summary, analysis of the miR-144/451^eGFP allele revealed unprecedented reporter expression in the entire pool of mature erythrocytes both in terms of specificity and reporter intensity. High transcriptional activity of the locus during terminal erythroid differentiation results in a substantial accumulation of eGFP protein, which in turn concedes detection of RBCs throughout their extended lifetime in circulation. Equivalent to Ly5 mediated tracking of leukocyte output, the miR-144/451^eGFP reporter enables single-cell resolution and quantification of in vivo erythroid lineage output from a subset of hematopoietic stem and progenitor cells. Moreover, the absence of reporter expression in early progenitors makes the allele compatible with the usage of already existing GFP-based reporters to isolate HSC subsets for subsequent analysis of erythroid potential. Thus, the miR-144/451^eGFP allele represents a novel quantitative in vivo erythroid lineage assay that will be a valuable tool in the study of blood formation and hematopoietic stem and progenitor cell heterogeneity.

The authors acknowledge the services of Daniel Bilbao-Cortes and Jeanette Rientjes at EMBL Flow Cytometry Core facility and Monash University's Gene Recombineering facility, respectively.

Contribution: K.D.R. contributed to the design, execution, and analysis of the experiments, and wrote the final version of the manuscript; and D.O.C. conceived and supervised this study.

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

Correspondence: Dónal O'Carroll, EMBL, Mouse Biology Unit, Via Ramarini 32, Monterotondo Scalo 00015, Italy; e-mail: ocarroll@embl.it.