Identification of the earliest NK-cell precursor in the mouse BM

Although the developmental stages leading to committed B and T cells have been dissected in great detail,¹  the developmental intermediates downstream of the common lymphoid progenitor (CLP) leading to a committed natural killer (NK)–cell precursor remain poorly defined. The earliest committed NK-cell progenitor (NKP) identified to date is characterized through the expression of the IL-2 receptor β chain (CD122) and the lack of pan-NK–cell surface markers (NK1.1 and CD49b), but this population is likely to be heterogeneous because only a minority of cells give rise to NK cells in in vitro assays.^2,3 

The transcription factor inhibitor of DNA binding 2 (ID2) is essential for NK-cell development. In its absence, fetal and adult NK-cell development is blocked at an early stage.^4-7  Id2 inhibits E proteins in NK cells and is itself regulated by E4BP4, which is also required for the differentiation of NKPs to immature NK cells.^8,9  Using a reporter line in which the green fluorescent protein (GFP) gene has been introduced into the Id2 locus, we report the identification of the earliest committed NK-cell precursor in adult BM. By analogy to B-cell development, we termed these pre-pro NK cells, and they express high levels of ID2, IL-7Rα (CD127), and lack expression of most, but not all, NK-cell markers. Furthermore, we demonstrate that ID2^gfp expression can be used to isolate the committed NK-cell progenitors within the heterogeneous NKP population.

We generated ID2^gfp reporter mice by inserting an internal ribosome entry site-GFP cassette into the 3′ untranslated region of the Id2 locus (G.T.B., S.C., and S.L.N. manuscript submitted April 2011). PU.1^gfp and Rag1^gfp mice have been described previously.^10,11  All mice were maintained under specific pathogen-free conditions at the Walter and Eliza Hall Institute according to institute guidelines.

Lineage (lin) marker–positive BM cells were depleted and analyzed by flow cytometry as described in supplemental Methods (available on the Blood Web site; see the Supplemental Materials link at the top of the online article).

To determine precursor frequencies, sorted progenitors cells were seeded in limiting dilution on OP9 or OP9-DL1 cells in the presence of 2% IL-7 supernatant and 5 ng/mL FMS-like tyrosine kinase 3 (Flt3) ligand for B- and T-cell development, respectively, and on OP9 cells plus 0.5% IL-7 supernatant and 50 ng/mL human IL-15 (R&D Systems) for NK-cell differentiation as described.¹²  After 7-10 days, wells were scored for the presence of colonies. Colonies were harvested for flow cytometric analysis, and precursor frequency was calculated according to Poisson statistics with the use of ELDA (extreme limiting dilution analysis) software Version 1.0 (http://bioinf.wehi.edu.au/software/elda/).¹³ 

The earliest committed NKP identified to date expresses CD122 but not the pan-NK cell–specific markers NK1.1 and CD49b.²  This population is likely to be heterogeneous because only 8%-40% of NKPs are reported to have unilineage NK-cell potential.^2,3,14  To test whether ID2 expression can be used to identify committed NK-cell progenitors, we analyzed the NKP population by using mice that have GFP introduced into the Id2 locus (ID2^gfp). Flow cytometric analysis showed that lin⁻CD122⁺NK1.1⁻CD49b⁻ NKPs could be subdivided into low and high ID2-expressing cells (Figure 1A; supplemental Figure 1A). To examine if ID2^gfp expression marks, the NK cell–committed NKPs, ID2^gfp-high, and ID2^gfp-low NKP cells were sorted and cultivated on OP9 stromal cells in the presence of IL-15, conditions that induce NK-cell development. Strikingly, only ID2^gfp-high NKPs efficiently generated NK1.1⁺CD49b⁺ NK cells (Table 1).¹³  The authors of a recent report¹⁴  showed that the NKP population consists of NK cell–restricted and NK/T bipotent precursor cells. Although we obtained similar results with unfractionated NKPs, we found that only NK-cell potential was enriched in ID2^gfp-high cells (data not shown). Thus, the NK-cell lineage–committed fraction of the NKP can be readily identified by the use of ID2^gfp-high expression.

To determine the earliest expression of ID2^gfp within the NK-cell lineage, and thus potentially the earliest committed BM-derived NK-cell precursor, hematopoietic stem and progenitor cell populations were analyzed for ID2^gfp expression by flow cytometry. Although very low ID2^gfp expression was observed in the hematopoietic stem cell–enriched lin⁻Sca-1⁺CD117⁺ population (supplemental Figure 1A), pronounced ID2^gfp fluorescence was found within the gate containing the lymphoid-restricted CLP population, defined as lin⁻Sca-1⁺CD117^intCD127⁺ (Figure 1B). ID2^gfp+ cells, however, differed from the CLP¹⁵  in that they expressed very high levels of CD127 and lacked expression of CD135 (Flt3; Figure 1C-D). Most ID2^gfp+ cells were found outside of the CLP gate in the lin⁻Sca-1^highCD117⁻ fraction, which also expressed high levels of CD127 and lacked CD135 (Figure 1C-D). This population has been observed previously by other groups. Although they found B and T lineage potential in the CD135⁺ fraction, the lineage potential of CD135⁻CD127^high cells has remained elusive.^16-18  When cultured on OP9 cells plus IL-7 and IL-15, both ID2^gfp-expressing populations efficiently generated NK1.1⁺ NK cells within 1 week of cultivation, whereas sorted multipotent progenitors and CLP were delayed in their capacity to differentiate into NK1.1⁺ cells (Figure 1E). Given their immature surface marker phenotype and NK-cell potential, we termed these cells pre-pro NK a and pre-pro NK b cells by analogy to the hierarchical B-cell developmental system.

We further characterized the 2 pre-pro NK-cell populations by analyzing the expression of surface makers characteristically expressed on maturing NK cells (for a summary of surface markers, see Huntingon et al¹  and Boos et al¹⁹ ). Pre-pro NK cells did not express NK-cell markers such as CD122, NK1.1, CD49b, or NKp46, but they did express the NK receptors CD244 (2B4) and NKG2D (supplemental Figure 1B). On cultivation under NK cell–specific conditions, pre-pro NK-cell subsets sequentially up-regulated CD122 and NK1.1 expression (supplemental Figure 1C). These data indicate that pre-pro NK cells represent more immature precursors than the NKPs, given their lack of CD122 expression and may represent the direct precursor of the NKP population.

To characterize pre-pro NK cells in an ID2^gfp-independent manner, a gating strategy was established that took advantage of the uniquely high expression of CD127 and absence of CD135 in these populations (as described in Figure 1C). By using this strategy, we found that PU.1 and Rag1, 2 factors that are expressed in CLPs and developing B and T cells,^10,20-22  were already silenced in pre-pro NK cells (supplemental Figure 2). These data show that pre-pro NK cells have established a unique gene expression signature within the lymphoid lineages and argue against a role of either protein in the NK-cell lineage.

To determine the lineage potential and purity of the identified pre-pro NK-cell populations, limiting dilution experiments under B cell–, T cell–, and NK cell–specific conditions were performed. Sorted CLP and pre-pro NK a and b cells were cultivated either on OP9 stroma plus specific cytokines for B- and NK-cell development or OP9-DL1 stromal cells plus IL-7 and Flt3L for T-cell development.¹²  As expected, CLP efficiently differentiated into all 3 lymphoid lineages. In contrast both pre-pro NK-cell populations gave rise to only NK-cell colonies at a high frequency (Table 1). These data suggest that despite the differences in CD117 expression the 2 pre-pro NK-cell subsets a and b are phenotypically and functionally similar, and we propose to summarize these 2 fractions into one pre-pro NK-cell precursor population.

In conclusion, by initially using the ID2^gfp reporter mouse model and later high levels of CD127, we have prospectively identified and characterized the pre-pro NK cells that likely represents the direct precursor of the earlier described NKP, whose purity can also be greatly enhanced by use of the ID2^gfp reporter. Identification of the earliest steps in NK-cell formation will now enable a detailed molecular characterization of the lineage commitment process in NK cells.

We are indebted to the facilities of our institute, particularly those responsible for animal husbandry and flow cytometry.

This work was supported in part by grants from the National Health and Medical Research Council (NHMRC) of Australia. G.T.B. is supported by fellowships from the Sylvia and Charles Viertel Charitable Foundation and Howard Hughes Medical Institute; S.L.N. is supported by a Pfizer Senior Research Fellowship (Australia); S.H.M.P. is supported by a Leukemia Foundation of Australia Scholarship; and S.C. is supported by an NHMRC fellowship.

Howard Hughes Funding

Contribution: S.C., S.H.M.P., S.L.N., and G.T.B. designed research, performed research, prepared an initial draft based on a systematic review of published literature, and discussed the draft.

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

Correspondence: Sebastian Carotta or Gabrielle Belz, Walter and Eliza Hall Institute of Medical Research, 1G Royal Parade, Melbourne, Victoria 3052, Australia; e-mail: carotta@wehi.edu.au or belz@wehi.edu.au.