|Home | About | Journals | Submit | Contact Us | Français|
Selective labeling of specific cell types by expression of green fluorescent protein (GFP) within the hematopoietic system would have great utility in identifying, localizing, and tracking different cell populations in flow cytometry, microscopy, lineage tracing, and transplantation assays. In this report, we describe the generation and characterization of a new transgenic mouse line with specific GFP labeling of all nucleated hematopoietic cells, as well as platelets. This new “Vav-GFP” mouse line labels the vast majority of hematopoietic cells with GFP during both embryonic development and adulthood, with particularly high expression in hematopoietic stem and progenitor cells. With the exception of transient labeling of fetal endothelial cells, GFP expression is highly selective for hematopoietic cells, and persists in donor-derived progeny after transplantation of hematopoietic progenitor cells. Finally, we also demonstrate that the loxP-flanked reporter allows for specific GFP labeling of different hematopoietic cell subsets when crossed to various Cre reporter lines. Notably, by crossing Vav-GFP mice to Flk2-Cre mice, we obtained robust and highly selective GFP expression in HSCs. Together, these data describe a new mouse model capable of directing GFP labeling exclusively of hematopoietic cells or exclusively of HSCs.
There have been many efforts to generate transgenic mice with transgene expression exclusively in the hematopoietic compartment1. The vav1 gene has been the focus of many such studies as it is highly expressed throughout hematopoietic development from the embryonic day 11.5 (e11.5) embryo through adulthood 2 There appears to be very limited expression in other tissues in the adult mouse, with the exception of the developing tooth bud2. Vav1 has been shown to activate the Rac/Jun kinase pathway and gene disruption assays have shown it to be essential for signaling through the antigen receptors of lymphocytes 3–5. Interestingly, even though Vav1 is highly expressed throughout the hematopoietic system, it is not essential for the development of blood cells in general 6.
The unique expression pattern of the vav1 gene has led to generation of several vav-driven cre mouse lines as well as vav-driven direct reporter mouse lines 7–12. These mouse lines have generally had great success in labeling the hematopoietic compartment with minimal off-target expression. Of particular note is Stadtfeld and Graf’s model where Cre recombinase is driven with promoter and enhancer elements of the vav gene 10. When crossed to a stop-lox-YFP reporter line, this model accomplished almost 100% labeling in all nucleated bone marrow (BM) cells and platelets in adult mice. They also found that nearly all KLS (ckit+, lin−, sca+) cells were labeled in the e13.5 fetal liver and approximately half of CD45+ (hematopoietic) cells from the e10.5 fetal liver were reporter positive 10. While this mouse line demonstrated great success in labeling the entire hematopoietic compartment, it does not allow for the resolution of specific cell populations within the hematopoietic lineage needed for experiments such as lineage tracing from hematopoietic stem cells (HSCs) and/or progenitor cells (HSPCs) or localization of HSCs/HSPCs. To enable fluorescent labeling of specific hematopoietic cell populations, we modified Stadtfeld’s construct so that the vav enhancer/promoter elements drive a fluorescent reporter that can be excised in specific hematopoietic cell subsets using Cre-mediated recombination. This new mouse line, called Vav-GFP mice, allows for two levels of specificity: firstly, the fluorescent reporter is under control of vav promoter elements and, secondly, it can be crossed to a multitude of Cre lines to drive excision of the reporter and thereby restricting fluorescence to a desired population of HSCs or HSPCs.
In this study we characterized the fluorescence of the Vav-GFP mouse line in BM and peripheral blood in both adult and fetal mice. In addition, we showed that the Vav-GFP cells can be distinguished from wild type host cells after transplantation as this is a likely application of the new mouse line. Finally, we also crossed the Vav-GFP mice to a Flk2-driven Cre mouse line to achieve targeted labeling exclusively of HSCs within the BM compartment 13,14. These data collectively show that the Vav-GFP mouse line generated here represents a novel tool to interrogate HSC differentiation and trafficking by providing hematopoietic-specific expression of a reporter construct under control of Cre mediated recombination.
Our goal was to generate a dual-purpose transgenic mouse line that allows for pan- hematopoietic or, in combination with selected Cre-expressing mouse lines, labeling of a subset of HSCs/HSPCs. To generate Vav-GFP mice, we used the murine regulatory elements of the vav gene to drive expression of a dual color reporter. A vector consisting of Vav regulatory elements and Loxp-flanked EGFP was linearized and injected into pronuclei of C57/B6l6 mice (Figure 1A). In this model, GFP is expressed until Cre-mediated recombination causes excision of GFP and a stop codon (Figure 1A and and5A5A).
To investigate the ability of the reporter construct to fluorescently label hematopoietic cells, HSPCs and mature cell populations were isolated from BM and peripheral blood (PB) of Vav-GFP mice. Flow cytometry analysis revealed reporter expression in all HSPCs and mature cells, including platelets and erythroid progenitors, but not in mature circulating red blood cells (RBCs) (Figure 1B and 1C). All HSCs, multipotent progenitors (MPPs) and myeloid progenitors displayed strong GFP expression that was clearly distinguishable from WT control cells (Figure 1B and 1D). Mature BM cell populations also expressed GFP, although at lower fluorescent intensities compared to HSPCs (Figure 1B, D, E). In the PB, virtually all hematopoietic cells expressed GFP at levels readily distinguishable from WT control cells (Figure 1C, F, G). Notably, GFP expression was robust in platelets, but not detected in circulating RBCs. Also, the reporter expression in circulating GM was significantly higher than in bone marrow resident GM (Supplemental Figure 1). This may be due to differences in Vav expression between these populations. Together, these data show strong reporter expression in all BM and circulating hematopoietic cells, with the exception of RBCs, making the Vav-GFP transgenic mouse a valuable tool for interrogating the blood system.
The Vav-GFP model was designed to only fluoresce in hematopoietic cells and no other tissues or cell types. To test the reporter specificity, we investigated GFP expression in non-hematopoietic cells of brain, liver, heart, and lungs. Whole organs were isolated, prepared into single cell suspensions, and stained with pan-hematopoietic (CD45) and the erythroid (Ter119) markers to exclude hematopoietic cells. Each organ was then analyzed by flow cytometry and CD45-Ter119− cells were tested for GFP expression. While GFP expression was readily detected in co-isolated CD45+ cells, there was no detectable off-target GFP expression found in non-hematopoietic cells in any of the organs surveyed (Figure 2A,B).
The most likely population of cells to exhibit off-target expression in the Vav-GFP model are endothelial cells (ECs), as other studies using vav regulatory elements have reported mixed results of off-target expression in ECs. For example, Georgiades et al showed that all CD31+ cells were labeled with b-galactosidase with their vav-cre line, whereas Ogilvy et al reported no vav-driven hCD4 in non-hematopoietic tissues by immunohistochemical analysis 7,8. To test endothelial GFP expression in our model, we isolated CD45-Ter119−CD31+Sca1+ ECs from the BM of Vav-GFP mice. In all (n=6) but one mouse surveyed, GFP expression was undetectable in BM ECs (Figure 2C,D). One mouse had approximately 5% of its ECs labeled with GFP, but no other mouse surveyed, including littermates, showed similar expression patterns.
Given that vav expression has been detected in the embryo as early as embryonic day 11.5 2, we wanted to test GFP expression in embryonic hematopoietic cells in Vav-GFP mice. Most, but not all (~85%), of CD45+ckit+ hematopoietic cells isolated from the caudal half of e11.5 Vav-GFP embryos displayed robust GFP expression (Figure 3A). The partial labeling may be due to inadequate accumulation of GFP as the transgene is just beginning to be expressed at this time point, or due to heterogeneity of the cell types included in the CD45+ckit+ phenotypic compartment.
As both progenitor and mature cell populations have been well characterized in fetal livers at e14.5, we investigated the GFP expression of hematopoietic cell subsets at this stage. We observed strong labeling in all hematopoietic cells surveyed, including HSPCs and mature cells (Figure 3B). Compared to the adult counterpart, the level of GFP expression varied in some cases, with fetal HSCs displaying lower GFP intensity than adult HSCs, whereas embryonic B cells appeared brighter than their adult equivalents (Figure 3B).
To test the specificity of GFP expression in fetal Vav-GFP mice, we isolated ECs from e14.5 fetal livers. In contrast to the lack of GFP expression in adult ECs, a large proportion (~50%) of ECs in the e14.5 fetal liver exhibited GFP expression (Figure 3C). Their EC identity was confirmed by costaining with anti-Tie2 and Vcam1 antibodies (data not shown). GFP expression in ECs is clearly transient as it was not detected in the adult and may reflect a brief period of vav expression by early endothelial cells or progenitors. This off-target expression is unlikely to affect adult studies with our Vav-GFP mouse line, but it may require exclusion of ECs to study developmental hematopoiesis in these mice. For example, the ECs can be excluded from flow cytometry assays by either MECA 32 expression or lack of CD45 expression. Conversely, endothelial GFP expression may be used to investigate fetal EC populations, or the relationship between endothelial and hematopoietic development.
To test whether GFP expression remains robust during reconstitution following transplantation, we performed transplants with KLS cells isolated from Vav-GFP mice. All mature cell and progenitor lineages surveyed in recipient mice showed clear separation by GFP expression between host and donor populations (Figure 4A,B). In addition, analysis of live, Ter119− cells in chimeric animals showed that all donor-derived cells (CD45.2+) were also GFP+. This result verifies that the transgene clearly labels donor cells upon transplantation and supports the utility of the Vav-GFP mice to track transplanted hematopoietic cells in reconstitution assays.
The flanking of GFP with loxP sites enables excision of the reporter gene in desired cell populations. To test whether HSC-specific labeling could be accomplished, we crossed the Vav-GFP mice to Flk2-Cre transgenic mice 15. We and others have previously shown that Flk2 is expressed by MPPs 16–19 and that Flk2-Cre labeling combined with a color switch model is capable of labeling the entire hematopoietic system downstream of the HSC 13,14. By crossing Vav-GFP mice to Flk2-Cre line we anticipated that all HSCs would be labeled with GFP, and that all downstream progenitors and mature cells would lack GFP expression due to excision of the reporter gene in Flk2+ progenitor cells20. In addition, non-hematopoietic cells would lack GFP expression due to the hematopoietic-specific expression of Vav-driven transgenes (Figures 1 and and2).2). Indeed, consistent with our previous findings, we did observe strong and highly selective GFP expression in HSCs, whereas the vast majority of hematopoietic progenitors and mature cells were unlabeled (Figure 5). Expression of GFP in a small proportion of downstream populations is a result of incomplete floxing from the specific Cre mouse used. In the case of Flk2-Cre mice, the recombination efficiency varies between individual Flk2-Cre mice 13,14. Thus, the combination of Vav-GFP and Flk2-Cre enables very bright, highly specific GFP labeling of HSCs.
In conclusion, the Vav-GFP mouse line directs GFP expression exclusively in the hematopoietic system, with the option of differentiation stage specificity by Cre-mediated recombination. We detected strong GFP labeling of all hematopoietic cell types assayed, except for RBCs, with minimal off-target expression in adult tissues. We also detected strong labeling of hematopoietic cells in the developing embryo, with limited off-target GFP expression in ECs. Transplantation experiments demonstrated GFP expression can be used to distinguish donor-derived cells, from host cells in hematopoietic reconstitution assays. We showed that the loxP elements are functional for Cre-mediated recombination by crossing the Vav-Cre mice to a Flk2-Cre line, which lead to highly selective GFP labeling of HSCs in the BM compartment. The Vav-GFP x Flk2-Cre cross provides an excellent example of how the Vav-GFP line can be used for lineage tracing studies, as well as direct visualization of HSC in adult BM for in situ assays, and for tracking HSC migration upon transplantation or mobilization. Our mice can also be crossed to commercially available floxed-stop-reporter mice to achieve labeling of HSC progeny with a second color. For example, a cross to a Rosa26-lox-stop-tomato mouse would result in mice with GFP+ HSCs, Tomato+ hematopoietic progenitor and mature cells, and unlabeled non-hematopoietic cells. Collectively, our data show that the Vav-GFP mouse line, alone, or by breeding to specific Cre lines to obtain selective GFP labeling in hematopoietic subpopulations, represents a novel tool for interrogating the hematopoietic system without fluorescence interference from non-hematopoietic cells. Lastly, our strategy to achieve HSC-specific reporter expression can be utilized for expression of any other transgene specifically in HSCs to assess transgene function exclusively in HSCs.
EGFP and HcRed were cloned into the pZ/EG plasmid 21, replacing lacZ and EGPF, respectively, to generate an EGFP-stop-loxP-HcRed-loxP reporter construct. The reporter fragment was then migrated into the vavINS-Cre-IRES-YFP plasmid 10 after SfiI and Not I digestion and 3-piece ligation to replace Cre-IRES-YFP with the EGFP-stop-loxP-HcRed-loxP reporter fragment. The vector was linearized and injected into pronuclei of C57/Bl6 mice at the UCSC transgenic facility. Multiple founders were used to establish a colony, but founder lines were not analyzed separately. Characterization of this line revealed no HcRed fluorescence after floxing. All mice were maintained and investigated in the UCSC vivarium according to IACUC-approved protocols.
Bone marrow and peripheral blood cells were isolated, processed and analyzed using a four-laser FACSAria or LSRII (BD Biosciences, San Jose, CA) as described previously 19,22,23. Flowjo Software (Ashland, OR) was used for data analysis and display. Mean fluorescent intensity (MFI) was determined for each cell population by calculating the average intensity for the entire population of both experimental and wild type controls, then subtracting the average wild type intensity from each experimental replicate. Cell populations were defined by the following cell surface phenotypes: HSC (Lin−Sca1+c-kit+CD48−Slamf1+Flk2−), MPP (Lin−Sca1+ c-kit+CD48+Slamf1−Flk2+), CMP (Lin−Flk2−Sca1− c-kit+FcgRmidCD34mid), GMP (Lin−Flk2−Sca1− c-kit+FcgRhiCD34hi), MEP (Lin−Flk2−Sca1− c-kit+FcgRloCD34lo), CLP (Lin−Sca1mid c-kitmidFlk2+IL7Ra+), GM (Ter119−Mac1+Gr1+CD3−B220−), Plts (FSCloTer119−CD61+), B cell (Ter119−Mac1−Gr1−CD3−B220+), T cell (Ter119−Mac1−Gr1−CD3+B220−), and erythroid progenitors (B220−, CD3−, Mac1−, Gr1−, Ter119+, Cd71+). The lineage cocktail consisted of antibodies recognizing CD3, CD4, CD5, CD8, B220, Gr1, Mac1, and Ter119. Mac1 antibodies were excluded from the lineage cocktail when analyzing fetal progenitors 24.
CD117-enriched bone marrow cells were isolated and double-sorted from Vav-GFP mice using a FACSAriaIII then transplanted into sublethally irradiated (500 rads) WT mice (C57BL6) 13. 1000 KLS (ckit+, lin−, Sca+) cells were transplanted into each recipient and peripheral blood or bone marrow were analyzed up to 24 weeks after transplantation.
Whole organs were dissected from Vav-GFP and wild type C57B6 mice and homogenized using a mortar and pestle for all organs. Cell suspensions were passed through 70 micron filters and stained with anti-CD45 and Ter119 antibodies to exclude hematopoietic cells from analysis. Endothelial cells were isolated as previously described 22. Briefly, bone marrow cells were isolated from Vav-GFP and wild type mice and were digested in a 3mg/mL collagenase solution for one hour. Similarly, e14.5 fetal livers were isolated from Vav-GFP and wild type embryos and digested with a 1mg/mL collagenase solution. Samples were then filtered and stained with anti-CD45, CD31, Sca1, Tie-2, and Vcam1 antibodies and analyzed using flow cytometry.
We thank Drs. Thomas Graf and Matthias Stadtfeld for the vavINS-Cre-IRES-YFP plasmid, Corrine Lobe for the Z/EG plasmid, Armen Shamamian at the UCSC vivarium for generation of transgenic mice, Stephanie Smith-Berdan and other Forsberg lab members for technical assistance and helpful discussions. This work was supported by an NIH/NIAID award (R21AI103656), an NIH/NIDDK award (R01DK100917) and UCSC startup funds to ECF; by NIH Training Grant 2T32GM008646 and an HHMI Gilliam Fellow Award to JP-C; CIRM Training grant TG2-01157 to SWB; and by CIRM Shared Stem Cell Facilities (CL1-00506) and CIRM Major Facilities (FA1-00617-1) awards to UCSC. ECF is the recipient of a California Institute for Regenerative Medicine (CIRM) New Faculty Award (RN1-00540) and an American Cancer Society Research Scholar Award (RSG-13-193-01-DDC).
Conflicts of interest: The authors have no conflicts to declare.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.