Conditional Gata2 inactivation results in HSC loss and lymphatic mispatterning

Abstract

The transcription factor GATA-2 plays vital roles in quite diverse developmental programs, including hematopoietic stem cell (HSC) survival and proliferation. We previously identified a vascular endothelial (VE) enhancer that regulates GATA-2 activity in pan-endothelial cells. To more thoroughly define the in vivo regulatory properties of this enhancer, we generated a tamoxifen-inducible Cre transgenic mouse line using the Gata2 VE enhancer (Gata2 VECre) and utilized it to temporally direct tissue-specific conditional loss of Gata2. Here, we report that Gata2 VECre-mediated loss of GATA-2 led to anemia, hemorrhage, and eventual death in edematous embryos. We further determined that the etiology of anemia in conditional Gata2 mutant embryos involved HSC loss in the fetal liver, as demonstrated by in vitro colony-forming and immunophenotypic as well as in vivo long-term competitive repopulation experiments. We further documented that the edema and hemorrhage in conditional Gata2 mutant embryos were due to defective lymphatic development. Thus, we unexpectedly discovered that in addition to its contribution to endothelial cell development, the VE enhancer also regulates GATA-2 expression in definitive fetal liver and adult BM HSCs, and that GATA-2 function is required for proper lymphatic vascular development during embryogenesis.

Figure 1. A Gata2 vascular enhancer confers CreERT² and mCh transgene expression in the embryonic vasculature.

(A) Schematic depicting the CreERT² (15) or mCh (19) cDNA driven by the HSV TK promoter (tkP) and the 1.2-kbp Gata2 VE (5) in VECreERT² or VEmCherry expression plasmid, respectively. Each minigene cassette was flanked by tandem repeats of chicken HS4 insulators (ins) (53). Both inserts were excised from the vector and coinjected (1:1) into the pronuclei of mouse oocytes to generate doubly transgenic (Tg^VE) mice. F₂–F₅ progeny were used in subsequent analyses (B–E). (B) Cre transgene copy number (normalized to Actin) was determined by qPCR to range between 5 and 47 copies. Cre mRNA level (normalized to endogenous endothelia-restricted Flk1 mRNA) in the heart (black bars) and kidney (white bars) of neonatal Tg^VE pups (n = 3 to 8) was determined by RT-qPCR. Of the 3 Tg^VE lines that showed significant endothelial mCh staining (see below), Tg^VE56 and Tg^VE62 both robustly expressed Cre mRNA, while Cre transcripts were barely detectable in Tg^VE73. qPCR primer sequences are listed in Supplemental Table 1. Data represent mean ± SD. (C) mCh epifluorescence in representative E10.5 embryos. Of 7 lines that stably transmitted both transgenes, Tg^VE62 expressed mCh most robustly in an endothelia-specific manner, while in some lines mCh was weakly expressed (e.g., Tg^VE60 and Tg^VE473). (D) Coincident expression (merge) of mCh (Tg^VE62) and eGFP (generated from Gata2^+/gfp; ref. 21) in the major and fine cranial vasculature in an E10.5 Tg^VE62:Gata2^+/gfp compound mutant embryo. mCh expression temporally and spatially parallels that of eGFP (Gata2) in the vasculature. The asterisk indicates a GFP-exclusive area of Gata2 expression in the ventral midbrain. (E) Coincidence (merge) of mCh (Tg^VE62) and eGFP (from Tg^Tie2.gfp) epifluorescence in the major and intersomitic vasculature of an E10.5 Tg^VE62:Tg^Tie2.gfp embryo, underscoring the vascular endothelial tissue specificity of the Tg^VE transgene.

Figure 2. ROSA26R-derived β-gal expression in Tx-treated Tg^VE embryos.

(A) X-gal accumulation is evident in the vasculature of a Tx-treated whole-mount E11.5 Tg^VE62:R26R compound mutant embryo (right), but not in a (Tx-treated) R26R littermate that lacks the Cre transgene (left). In this experiment, embryos were administered Tx from E9 to E11 by gavage of pregnant dams. (B–G) The Tg^VE62:R26R compound mutant embryo (A, right) was sectioned to further analyze X-gal localization at the cellular level. (B) The endocardium (e) and the endocardial cushion (ec) in the E11.5 heart were positive for β-gal activity. (C) X-gal staining is also evident in the cardinal vein (cv), from which the lymphatic vascular system develops. (D–F) Transverse cryosections in the vicinity of the dorsal aorta (da) revealed budding single (arrowheads, D and E) and clustered (arrows, enlarged view, E and F) lacZ-positive (rounded) hematopoietic cells. ms, mesonephros. (G) Punctate, single-cell X-gal staining was also present in FL endothelial and hematopoietic cells. Scale bars: 1 mm (A); 100 μm (B, D and G); 50 μm (C); 10 μm (E and F).

Figure 3. Tx-treated Tg^VE:Gata2^–/fl FL cells fail to generate hematopoietic colonies in vitro.

(A–D) FLs recovered from Tx-treated (see Methods) E14.5 embryos were dissociated into single cells. The total FL cell number in Gata2 heterozygous (Gata2^+/–, Tg^VE:Gata2^+/fl [+/–]; n = 13) and Tg^VE:Gata2^–/fl embryos (Tg:–/fl; n = 6) was significantly and correspondingly reduced with increasing loss of GATA-2 activity compared with wild-type embryos (Gata2^+/+, Tg:Gata2^+/+ [+/+]; n = 11). (B) Total FL cells from the embryos shown in A were seeded into methylcellulose medium (M3434). For control embryos, 2 × 10⁴ total FL cells were seeded per plate. For Tg^VE:Gata2^–/fl embryos, cells were seeded at multiple concentrations (2 × 10⁴, 5 × 10⁴, and 1 × 10⁵ per plate) in triplicate. Colonies were assessed after 7–12 days in culture. Results represent mean ± SD of 4 independent experiments. Colony numbers were normalized to 10⁵ FL cells for graphical illustration. Compared with wild-type FL cells, Gata2 heterozygous and Tg^VE:Gata2^–/fl cells generated fewer and no CFU colonies, respectively. (C) Total RNA was prepared from FL cells of Tx-treated embryos (E9–E11; see A) and then quantified (in triplicate) for Gata2 mRNA (normalized to endogenous Hprt mRNA) by RT-qPCR. Gata2 mRNA was reduced by 85% in Tx-treated Tg^VE:Gata2^–/fl (Tg:–/fl; n = 6) compared with wild-type embryos (Gata2^+/+ [+/+]; n = 6). qPCR primer sequences are listed in Supplemental Table 1. (D) Individual hematopoietic colonies (n = 35) isolated from methylcellulose cultures seeded with Tx-treated FL cells from different Tg^VE:Gata2^+/fl embryos from separate experiments (see B) were used for genomic DNA extraction, and then qPCR (in triplicate) was performed using primers that detected Gata2 exon 5 and Actin (normalization control; see Supplemental Table 1). Gata2 wild-type (arbitrarily set at 100%) and heterozygous genomic DNAs from tail snips of adult mice were included to validate the assay. Overall, in Tg^VE:Gata2^+/fl cells, 84% of the single Gata2 floxed allele was successfully deleted by Tx-induced Cre recombinase. Data represent mean ± SD

Figure 4. Reduced HSC numbers in Tx-treated Tg^VE:Gata2^–/fl FLs.

FLs recovered from E14.5 embryos that had been exposed to Tx in utero were mechanically disrupted, individually processed, and stained with various antibodies prior to analysis by flow cytometry. (A and B) Representative contour plots for the LSK fraction of 3 Tx-treated Gata2 control (Tg^VE:Gata2^+/fl, Tg^VE:Gata2^+/–, and Gata2^–/fl in A; Gata2^+/+, Tg^VE:Gata2^+/+, and Gata2^–/+ in B) and 3 Tx-treated compound mutant Tg^VE:Gata2^–/fl embryos, which were administered Tx in utero on E9–E11 (A) or E11–E13 (B). The fraction of LSK cells in the gated areas used to quantify the HSC compartment (boxed) is shown. LSK cells were essentially absent in the Tx-treated Tg^VE:Gata2^–/fl FLs whether Cre was induced during or after the initiation of definitive hematopoiesis. The total number of LSK (C) and LSKS (LSK Slam or LSKCD150⁺CD48^–; D) cells recovered from the livers of E14.5 embryos, which had been exposed to Tx from E9 to E11, of various Gata2 genotypes are represented (the ordinate axis is on log scale). Gata2 wild-type and pseudo-wild-type (Gata2^+/+, Tg^VE:Gata2^+/+, Gata2^+/fl [+/+]; n = 6), Gata2 heterozygous (Gata2^–/fl, Tg^VE:Gata2^+/–, Tg^VE:Gata2^+/fl [+/–]; n = 9), and Tg^VE:Gata2^–/fl (Tg:–/fl; n = 4). Data were compiled from two independent experiments; horizontal black bars represent the mean number of LSK (C) or LSKS (D) cells of each group of embryos. Statistical significance was determined by Student’s t test. (E) Prominent mCh expression in FL LSK and LSKS cells. FL cells from Tx-treated (from E9 to E11) E14.5 embryos were analyzed for mCh expression by flow cytometry. A large fraction of E14.5 FL LSK or LSKS cells express mCh in Tg^VE-positive Gata2 wild-type (Tg:+/+; n = 2) as well as Tg^VE-positive Gata2 heterozygous (Tg^VE:Gata2^+/–, Tg^VE:Gata2^+/fl [Tg:+/–]; n = 5) embryos. Compared with the LSK cells recovered from Tg:+/+ and Tg:+/– FLs (2.2 × 10⁴ to 5.7 × 10⁴ and 1.3 × 10⁴ to 4 × 10⁴, respectively), very few LSK cells (98–477) were recovered from Tg^VE:Gata2^–/fl (Tg:–/fl; n = 4) embryos. Compared with the LSKS cells recovered from Tg:+/+ and Tg:+/– FLs (1.2 × 10³ to 4 × 10³ and 0.5 × 10² to 1.6 × 10², respectively), a minuscule number of LSKS cells were recovered from Tg^VE:Gata2^–/fl FLs. Although there appears to be little difference in mCh⁺ LSK and LSKS percentages in Tx-treated Tg^VE:+/+ and Tg^VE:+/– FLs, the absolute number indeed drops by half.

Figure 5. Severely attenuated HSC recovery from Tx-treated Tg^VE:Gata2^–/fl adult BM.

(A) Adult mice (8–12 weeks of age) were gavaged with Tx (5 mg/d) or sunflower oil for 5 consecutive days. BM cells were isolated 48 hours after the final Tx administration and then stained with various antibodies directed against cell surface antigens prior to analyses. Representative contour plots for the LSKS (LSK Slam; LSKCD150⁺CD48^–) fractions in Tx-treated control (2 Tg^VE:Gata2^+/fl and 1 Tg^VE:Gata2^+/–) mice and in 3 Tx-treated Tg^VE:Gata2^–/fl compound heterozygous adult mice are shown. The percentage of LSKS cells in each gated area (boxed) is shown. (B) Distribution of highly purified HSCs isolated from Tx-treated (white) or mock-treated (black) control or test mice (2 tibias plus 2 femurs). Gata2^+/fl and Gata2^–/fl adult mice without or with Tg^VE were gavaged with Tx (5 mg/d) or sunflower oil for 5 consecutive days. The 4 groups of mice are designated as follows: Gata2^+/fl, +/fl (n = 2); Tg^VE:Gata2^+/fl, Tg:+/fl (n = 8); Gata2^–/fl, –/fl (n = 3); and Tg^VE:Gata2^–/fl, Tg:–/fl (n = 10). Data were compiled from 3 independent experiments. The ordinate axis is on a log scale; the horizontal bars represent the mean number of LSKS cells in Tx-treated mice of each genotype. Statistical significance was determined by Student’s t test. (C) Absolute number of adult BM LSKS cells in each mouse (2 tibias plus 2 femurs) expressing mCh (ordinate axis is log scale). The number of mCh-expressing cells in the Tg^VE:Gata2^+/fl versus Tg^VE:Gata2^–/fl fractions differs by 2–3 orders of magnitude. The horizontal bars represent the mean number of mCh-positive LSKS cells in Tx-treated mice of each genotype. Statistical significance was determined by Student’s t test.

Figure 6. Long-term hematopoietic reconstitution of Tx-treated E14.5 Tg^VE:Gata2^–/fl FL cells is severely compromised.

(A) Strategy for competitive reconstitution of lethally irradiated CD45.1 adult mice. 5 × 10⁵ E14.5 total FL cells from control (Gata2^+/fl [n = 1],Tg^VE:Gata2^+/+ [n = 3], or Tg^VE:Gata2^–/fl [n = 2]) Tx-treated CD45.2 embryos were transplanted together with 5 × 10⁵ CD45.1 adult BM cells. FL cells from each embryo were transplanted into 5 recipients. (B) Donor chimerism in the peripheral blood was assessed by flow cytometry using anti-CD45.1–PECy7 and anti-CD45.2–APC antibodies from 4 to 16 weeks after transplantation. CD45.2 donor cell contribution (mean ± SD) from 2 independent experiments is shown. Unlike the robust peripheral blood reconstitution by control FL cells in transplant recipients, Tg^VE:Gata2^–/fl FL cells barely reconstituted their irradiated hosts. Transplant recipients: Gata2^+/fl (n = 5), Tg^VE:Gata2^+/+ (n = 15), Tg^VE:Gata2^–/fl (n = 10). (C and D) Sixteen weeks after transplantation, BM cells were harvested from each mouse (2 tibias plus 2 femurs), and the ratios of CD45.1 to CD45.2 LSKS cells were determined by flow cytometry. Representative contour plots of irradiated CD45.1 recipients that received Gata2^+/fl (+/fl) or Tg:Gata2^–/fl (Tg:–/fl) FL cells are shown (C). Note the conspicuous absence of CD45.2 FL donor-derived LSKS cells in the latter representative recipient. The average ratio of CD45.1 to CD45.2 LSKS cells in transplant recipients that received Gata2^+/fl (+/fl; n = 5), Tg:Gata2^+/+ (Tg:+/+; n = 5), or Tg:Gata2^–/fl (Tg:–/fl; n = 10) is shown (D). Note that in the latter group of recipients, the majority of LSKS cells are not derived from FL. Data represent mean ± SD.

Figure 7. mCh expression is restricted predominantly to the LSK fraction in adult BM.

Cells were isolated from BM, spleen, and thymus of wild-type adult mice without Tg^VE (n = 2; gray bars) or hemizygous for Tg^VE (n = 4; black bars) and then stained with various antibodies directed against developmental cell surface antigens prior to flow cytometric analyses. The percentages of mCh-positive cells within the immature (LSK and lin^–Sca-1^–c-Kit^hi) and committed erythroid, myeloid, B lymphoid, or T lymphoid cell compartments are summarized. The latter 4 committed lineages were characterized using pairs of antibodies (CD71 and TER-119, Mac1 and Gr-1, B220 and CD19, CD4 and CD8), respectively. Most conspicuously, mCh is expressed at 30- to 1,000-fold lower abundance in all mature hematopoietic lineages, but is robustly expressed exclusively in adult BM LSK cells.

Figure 8. Tx-treated Tg^VE:Gata2^–/fl embryos exhibit anemia, edema, hemorrhage, and late-embryonic lethality.

(A) Gata2^+/fl female mice were intercrossed with Tg^VE:Gata2^+/– compound transgenic males. Embryos were collected from pregnant dams that were gavaged with Tx between E9 and E11. Anemia, edema, and hemorrhage were detected in Tx-treated Tg^VE:Gata2^–/fl (Tg:–/f) embryos, but not in any other genotype control littermates, of which some are shown here (Gata2^–/fl, Tg^VE:Gata2^+/–, Tg^VE:Gata2^+/fl [–/f, Tg:+/–, Tg:+/f, respectively]), beginning around E13.5 and progressively increasing in severity before they succumbed to lethality and necrosis by E15.5–E16.5. Arrows: subcutaneous edema in Tg^VE:Gata2^–/fl embryos. Note the visibly paler FLs (magenta asterisks) in Tx-treated E13.5 and E15.5 Tg^VE:Gata2^–/fl embryos in comparison to Tx-treated controls. (A patch of hemorrhage in the E14.5 Tg^VE:Gata2^–/fl embryo visually obscured its pallid liver). These phenotypes were reproducible using either of two transgenic lines (Tg^VE56: E14.5 embryos; Tg^VE62: E13.5 and E15.5 embryos). Further experiments performed with only a single Tx delivery on E9 (data not shown) or after 3 consecutive doses from E11 to E13 instead of E9 to E11 generated the same phenotypes (anemia, edema, hemorrhaging) in Tg^VE:Gata2^–/fl embryos. Scale bar: 0.5 mm. (B) Genotyping of a (large) representative E13.5 litter collected from a Tx-gavaged Gata2^–/fl female intercrossed with a Tg^VE:Gata2^+/fl adult male. Yolk sac genomic DNAs were used in separate PCR reactions to specifically detect the Cre and mCh transgenes (top), 3 Gata2 alleles (fl, wild-type [+], or knockout [–]; middle) and the Cre-mediated Gata2 exon 5–deleted allele (Δ; bottom). The Gata2^Δ amplicon was only detected in embryos bearing both the Gata2^fl allele and Tg^VE. Since CreERT² expression is restricted to endothelial cells in the yolk sac, the PCR amplicon derived from the unrecombined Gata2^fl allele was detected in total yolk sac genomic DNA of Tx-treated Tg^VE:Gata2^–/fl embryos. To generate the Gata2^Δ allele as a positive control, the Gata2^+/fl mouse was interbred with the ubiquitously expressed AyuI-Cre transgenic mouse (21). Sequences of primers used for PCR genotyping are listed in Supplemental Table 1.

Figure 9. Aberrant lymphatic development in Tg^VE:Gata2^–/fl embryos.

(A and B) Tx-treated E14.5 Gata2^+/fl (A) and Tg^VE:Gata2^–/fl (B) embryos were fixed in 4% paraformaldehyde and then gradually cleared in benzyl benzoate:benzyl alcohol for whole-mount photography. Note the blood-engorged jugulo-axillary lymph sac (arrow) in the Tg^VE:Gata2^–/fl, but in not in the control, embryo. (C–F) Transverse paraffin sections (6 μM) of an E13.5 Tx-treated (E9–E11) Tg^VE:Gata2^+/+ (C and E) or Tg^VE:Gata2^–/fl (D and F–J) embryos were stained with H&E. While in the control embryo, the jugular vein (v) is distinctly separated from the jugular lymph sac (ls) at the axial level of the thyroid gland (th, C; an enlarged view of boxed area is shown in E), only a single vessel lumen is observed in the Tx-treated Tg^VE:Gata2^–/fl embryo (D; enlarged view of boxed area is shown in F). C and D were taken at the same magnification; the tissue edema in Tg^VE:Gata2^–/fl embryo did not permit coverage of the entire area in the microscopic field. By following serial sections anteriorly (G and H) and posteriorly (I and J) of the section shown in F, it appeared that the jugular vein remained abnormally connected to the lymphatic system. Note the interstitial edema (indicated by the arrow; B) and the aberrant presence of erythrocytes in the lymph sacs (indicated by asterisks) before the onset of visible hemorrhage (e.g., see Figure 8) of E13.5 Tx-treated Tg^VE:Gata2^–/fl embryo. a, carotid artery; eo, esophagus; sg, sympathetic ganglion; tr, trachea; va, vagal nerve trunk. Scale bars: 1 mm (A and B), 200 μm (C and D), 100 μm (E–J).