hCD2-iCre and Vav-iCre Mediated Gene Recombination Patterns in Murine Hematopoietic Cells

Cre-recombinase mediated conditional deletion of Lox-P site flanked ("floxed") genes is widely used for functional gene annotation in mice. Many different Cre-transgenic mouse lines have been developed for cell-type specific gene disruption. But often, the precise tissue-patterns of Cre activity remain incompletely characterized. Two widely used transgenes for conditional gene recombination in hematopoietic cells are Vav-iCre driven from the murine Vav1 promotor, and hCD2-iCre driven from the human CD2 promotor. Vav-iCre expresses active Cre in fetal and adult hematopoietic stem cells and all descendants, hCD2-iCre in immature and mature B and T lymphocytes. To better characterize which hematopoietic cells contain hCD2-iCre activity, we compared EYFP fluorescence in hCD2-iCre+/- R26-stop-EYFP+/- and Vav-iCre+/- R26-stop-EYFP+/-mice. R26-stop-EYFP ubiquitously encodes EYFP preceded by a floxed stop cassette. By removing it, Cre activity induces measurable EYFP expression. Our results confirm the known activity patterns for both Cre transgenes and unveil additional hCD2-iCre mediated reporter gene recombination in common lymphoid progenitors, in natural killer cells and their progenitors, and in plasmacytoid and conventional dendritic cells. This supports previously proposed common lymphoid origins for natural killer cells and subsets of dendritic cells, and indicates the need to consider pleiotropic effects when studying hCD2-iCre mediated conditional knockout mice. Vav-iCre+/- R26-stop-EYFP+/-mice did not show the non-hematopoietic recombination in vascular endothelial cells seen in other Vav-Cre mouse lines, but displayed an unexpected Vav-iCre mediated recombination in a bone cell subset lacking hematopoietic markers. This pinpoints the need to consider stromal cell contributions to phenotypes of Vav-iCre mediated conditional knockout mice. Altogether, our data provide the first detailed assessment of hCD2-iCre and Vav-iCre mediated deletion of floxed genes during lymphocyte development from hematopoietic stem cells and open up novel applications for either Cre-transgenic mouse line.

Introduction approved by the Institutional Animal Care and Use Committee (IACUC, Assurance Number: A3194-01) of The Scripps Research Institute (TSRI). All efforts were made to minimize animal suffering. Mice were euthanized by CO 2 /O 2 mixture or halothane volatile anesthetic overdose inhalation.

Cell preparation
BM, thymocyte and splenocyte single cell suspensions were prepared and BM and spleen red blood cells lysed with BD Pharmlyse (BD Biosciences) as previously described [19][20][21]. For isolation of bone cells, whole bones were cleaned from muscle tissue, BM was flushed out, and the bones were then chipped into little pieces and digested with collagenase as described elsewhere [22].
Dendritic cells include cDC with important roles in antigen presentation and T cell activation, and pDC capable of producing large type I IFN amounts upon viral encounter. CLP and common myeloid progenitors (CMP) can each give rise to cDC and pDC, although CLP contributions are thought to play a minor role under steady state conditions in vivo, in particular for cDC development [15][16][17]. Interestingly, we found 98% EYFP + pDC but only~20% EYFP + cDC in hCD2-iCre +/-R26-stop-EYFP +/mice ( Fig 2B). Thus, although most cDC and pDC express no or very low CD2 mRNA (Fig 1A,1F and 1G), hCD2-iCre is active in pDC and some cDC, or in their progenitors.
Altogether, our data indicate that hCD2-iCre leads to gene recombination not only in B and T cells, but also in iNKT cells, NK cells, pDC and some cDC.
As previously noted [13], the EYFP expression pattern is consistent with increasing CD2 mRNA expression in post-DN3 stage thymocytes and mature T cells, but contrasts with the paucity of CD2 mRNA in ETP and DN2 cells (Fig 1E and 1G).
B lymphocytes arise from CLP through a series of developmental stages in the BM [34]. A previous study reported hCD2-iCre induced EYFP expression in all peripheral B cells and BMderived CD19 + IgDimmature B cells [3], but the precise developmental stage where EYFP is first induced remained unclear. To better characterize hCD2-iCre activity during B cell development, we thus analyzed EYFP expression on successive stage A-C pro-B and early pre-B cells, stage D late pre-B cells, stage E newly formed/immature B cells and stage F recirculating mature/follicular B cells [19,34] in the BM of hCD2-iCre +/-R26-stop-EYFP +/or R26-stop-EYFP +/mice ( Fig 3B). Interestingly, we found >95% EYFP + cells in all B cell developmental subpopulations of hCD2-iCre +/-R26-stop-EYFP +/but not control mice. This activity pattern is consistent with the high CD2 mRNA expression in post-fraction C B cell developmental stages and mature B cells, but contrasts with the low CD2 mRNA expression in fractions A-C ( Fig  1A,1D and 1G).
Altogether, hCD2-iCre recombines floxed genes in all B and T cell developmental subsets, including the earliest thymic T cell and BM B cell precursors despite their low CD2 mRNA expression levels.

Discussion
Among several Cre-transgenes that allow conditional disruption of floxed genes in hematopoietic cells in mice, Vav-iCre is commonly used to recombine genes in HSC and all descendants, and hCD2-iCre for gene manipulation in B and T lymphocytes [2,3,12,13]. Here, we show that beyond their reported gene recombination patterns, Vav-iCre and hCD2-iCre also disrupt floxed genes in other murine cells. Besides hematopoietic cells, Vav-iCre recombined a floxed EYFP reporter gene also in a Lin -CD45 -CD31 -CD51 -Sca-1bone cell type that lacks hematopoietic surface markers. Its identity remains to be elucidated. hCD2-iCre recombined a floxed EYFP reporter not only in B and T cells, but also in CLP, pro-B and all other stages of B cell development, ETP and subsequent stages of T cell development, NK cell progenitors, immature and mature NK cells, pDC and subsets of cDC.
Our analysis of Vav-iCre +/-R26-stop-EYFP +/mice confirmed the previously reported Vav-Cre or Vav-iCre activity in all hematopoietic cell subsets [3,[5][6][7][8][9]. In one other Vav-Cre transgenic mouse model, Vav was also expressed in germ cells and endothelial cells (EC) [6]. Similarly, the Vav-iCre transgene used in our study was previously reported to express Vav in testis and ovaries [3]. But in contrast to the other Vav-Cre line, our Vav-iCre transgenic mice showed no reporter gene recombination in vascular EC. We also found no Vav-iCre activity in osteoblasts and MSC (Fig 5), alleviating concerns that Vav-iCre activity in vascular niche cells could indirectly impact early hematopoiesis in the BM. However, we observed an unexpected Vav-iCre mediated EYFP expression in a Lin -CD45 -CD31 -CD51 -Sca-1subset of bone cells. Although these cells have been observed elsewhere [22], their identity is unknown. Vav-iCre activity might suggest a hematopoietic origin, but their lack of hematopoietic markers (Fig 5) opens the possibility that these cells are non-hematopoietic, reminiscent of Vav-iCre expression in germ cells [3]. In the future, it will be important to determine whether these bone cells express endogenous Vav, and to elucidate their identity and function.
EYFP expression in CLP, ETP and all later stages of T cell development in both systems is consistent with a common lymphoid-primed origin for most T cells [13,38]. hCD2-iCre mediated EYFP expression in pro B cells and all later stages of B cell development, and in NKP, iNK and mNK moreover supports previously proposed common lymphoid origins of B cells and NK cells [14,40]. The existence of EYFP -ETP in hCD2-iCre +/-R26-stop-EYFP +/mice could reflect insufficient transgene activation in the specific progenitors that gave rise to these cells, or the existence of distinct hCD2-activating and non-activating T cell progenitor subsets, reminiscent of the hypothesized origin of Il7r-reporter activating and non-activating pro T cells from distinct BM progenitors whose identity remains to be determined [38]. Maximal hCD2-iCre activity in NKP with progressive reduction in iNK and mNK might indicate partial Cre-inactivation and outgrowth of undeleting cells after the NKP stage, or simply EYFP loss during red blood cell lysis as seen in T cells [3]. In an alternative possibility, it will be interesting to study if the emergence of EYFPmature NK cells in hCD2-iCre +/-R26-stop-EYFP +/mice, which was also seen in another study [41], ascribes in vivo relevance to the ability of myeloid progenitors to produce NK cells in human in vitro systems [30].
One of our most interesting findings is that 98% of splenic pDC, and 20% of splenic cDC express EYFP in hCD2-iCre +/-R26-stop-EYFP +/mice (Fig 2). The hematopoietic origin of DC has long been controversial, mainly due to the ability of both lymphoid and myeloid progenitors to generate DC in vitro and after transplantation into lymphopenic mice [17,42]. Recently, IL-7 fate mapping [38] and the identification of a CMP-derived macrophage/DC progenitor (MDP) [43] and a downstream common DC precursor (CDP) capable of producing both pDC and cDC [44][45][46] have suggested a primarily myeloid origin in particular for cDC. On the other hand, lymphoid progenitor contributions would be consistent with the ability of CLPs to generate pDC and cDC, with the expression of Rag1 and detection of D-J rearranged IgH genes in splenic pDC [16,[47][48][49], and with cell-intrinsic requirements for IL-7 signaling for the development of subsets of splenic pDC and cDC [50]. The EYFP reporter expression pattern on hCD2-iCre +/-R26-stop-EYFP +/-DC again strikingly resembles that on Il7r Cre R26-stop-YFP + DC [38] and might support the view that under steady-state in vivo conditions, many pDC and a minor fraction of cDC may develop from lymphoid progenitors. Reminiscent of the situation in ETP, hCD2-iCre induced EYFP expression in pDC and cDC despite their low CD2 mRNA content (Fig 1A,1F and 1G) might possibly be explained by EYFP induction in upstream CLP (Fig 4). To further elucidate the origin of the EYFP expression in cDC from hCD2-iCre +/-R26-stop-EYFP +/mice, it will be interesting to study if these cells emerge from EYFP + CD11c + B220 + CCR9 -cDC progenitors [51]. These are included in our pDC gate, which contains 98% EYFP + cells. In any case, our data are consistent with contributions of both myeloid and lymphoid progenitors to pDC and cDC development [17,42]. They indicate that contributions of altered DC function to any phenotypes of hCD2-iCre based conditional knockout mice need to be considered, in particular when studying T cells which are activated by antigens presented on DC.
Again similar to Il7r Cre R26-stop-YFP + mice [38], we found low-level (<5%) EYFP expression in hCD2-iCre +/-R26-stop-EYFP +/macrophages and granulocytes (Fig 2) but not CMP, GMP and MEP (Fig 4). Although these EYFP + myeloid cells are very rare, they could reflect the ability of CLP and T lineage precursors to generate myeloid cells under appropriate conditions, or the ability of certain neutrophil subsets to express lymphoid markers, discussed in detail elsewhere [38].
Altogether, our data confirm and expand the EYFP expression in developing and mature T cells, B cells and NK cells previously found in the same hCD2-iCre transgenic line [3,41] and the EYFP expression in developing and mature T cells and B cells in other CD2-Cre transgenics [3,12,13]. One study also reported mosaic EYFP expression in the testis of hCD2-iCre transgenic males but not in ovaries in females [3]. A hCD2 transgenic mouse line expressed hCD2 protein in T cells, but not in B cells [52]. In mice harboring multiple hCD2 transgene copies, immunohistochemistry suggested hCD2 protein expression in T cells, megakaryocytes, platelets and bone marrow cells of unknown identity with convoluted nuclei speculated to be myeloid precursors [52]. This contrasts with the lack of EYFP expression in myeloid progenitors and MEP in our mice (Fig 4) but could be consistent with the low level EYFP expression we found in macrophages and granulocytes (Fig 2), and with myeloid contributions to the EYFP + DC in our mice (Fig 2).
The differences in transgene expression between different mouse lines in part result from differences in the precise Vav or hCD2 gene regions and supporting features included in each transgene. Moreover, during the generation of transgenic mice, the transgene integrates randomly and in variable copy numbers into the genome. Chromatin context affects transgene expression levels. During breeding, meiotic recombination might affect transgene structure and copy numbers. This, differential maternal/paternal effects and differences in genetic backgrounds such as the C57BL/6-J129 mixed background of the R26-stop-EYFP +/mice can lead to variability in transgene expression levels between different mouse lines and even between individual mice of the same line [4]. These unavoidable limitations of transgenic mice likely explain the differences between our data and previously published data even using the same Cretransgenic lines.
This notwithstanding, our data provide important novel insight into the Vav-iCre and hCD2-iCre transgene activity patterns in hematopoietic and non-hematopoietic cell populations. They suggest applications for these transgenes for conditional gene modulation in novel cell types, and provide a map of cell types whose extrinsic contributions to phenotypes of Vav-iCre or hCD2-iCre mediated gene modulation in a cell type of interest must be taken into account when interpreting results.
Finally, we propose that in addition to Il7r Cre R26-stop-YFP + mice [38], hCD2-iCre + R26-stop-EYFP + mice are an excellent genetic tool for lineage tracing studies of lymphoid cells in vivo that avoids the often non-physiological differentiation potential of CLP or other lymphoid precursors in vitro or upon engraftment into lymphopenic hosts [38].