Homologous recombination into the eosinophil peroxidase locus generates a strain of mice expressing Cre recombinase exclusively in eosinophils

Cre-recombinase expression in eoCRE mice generates a model system facilitating eosinophil-specific knockouts, and/or, ectopic overexpression of heterologous genes.

Abstract

Eosinophils are generally linked to innate host defense against helminths, as well as the pathologies associated with allergic diseases, such as asthma. Nonetheless, the activities of eosinophils remain poorly understood, which in turn, has prevented detailed definitions of their role(s) in health and disease. Homologous recombination in embryonic stem cells was used to insert a mammalianized Cre recombinase in the ORF encoding Epx. This knock-in strategy overcame previous inefficiencies associated with eosinophil-specific transgenic approaches and led to the development of a knock-in strain of mice (eoCRE), capable of mediating recombination of “floxed” reporter cassettes in >95% of peripheral blood eosinophils. We also showed that this Cre expression was limited exclusively to eosinophil-lineage committed cells with no evidence of Cre-mediated toxicity. The efficiency and specificity of Cre expression in eoCRE mice were demonstrated further in a cross with a knock-in mouse containing a “(flox-stop-flox)” DTA cassette at the ROSA26 locus, generating yet another novel, eosinophil-less strain of mice. The development of eoCRE mice represents a milestone in studies of eosinophil biology, permitting eosinophil-specific gene targeting and overexpression in the mouse as part of next-generation studies attempting to define eosinophil effector functions.

Introduction

Recent studies have suggested that eosinophils play an underappreciated and necessary role(s) in the maintenance of homeostatic baseline (i.e., health), as well as roles in a widening group of inflammatory diseases (reviewed in refs. [1, 2]). This greater understanding of eosinophils and their associated effector functions has been achieved, in part, through investigations of mouse models of ever-increasing complexity. For example, transgenic mice overexpressing eosinophil agonist cytokines and chemokines have been useful reductionist models to study eosinophil-mediated activities [3–5]. Strains of mice congenitally deficient of eosinophils have also provided definitive evidence for the roles of eosinophils in baseline homeostatic mechanisms (e.g., marrow plasma cell survival and accumulation [6], defined roles in parasite infection [7, 8], and roles in immune responses associated with allergen provocation [9, 10]). In addition, eosinophils recovered from the peripheral blood of transgenic mice (NJ1638) [11] or derived from ex vivo bone marrow cultures [12] have been an invaluable source of cells for adoptive transfer strategies mechanistically exploring eosinophil effector functions in vivo [13–15].

Although the methodological advances possible with these mouse model approaches have been formidable, the lack of an ability to target eosinophil gene expression (positively or negatively) in the context of mouse models of health and disease has represented a singularly significant obstacle to a greater understanding of eosinophil-mediated activities. To date, the only available strategy has been partial bone marrow engraftment into strains of mice congenitally deficient of eosinophils [5]. Unfortunately, the complexity and the logistical limitations of such studies (e.g., the costs and difficulties associated with breeding the necessary animals) have all but prevented the wider use of this strategy. Thus, a need exists for an experimental strategy and in particular, a mouse model that would permit the generation of eosinophil lineage-specific knock-outs and the ectopic overexpression of heterologous genes exclusively in eosinophils.

Our previous demonstrations of the eosinophil-specific character of expression occurring from Epx [9, 16] and the subsequent realization that the location of Epx within the genome may be critical to appropriate gene expression (our unpublished observations) suggested a practical experimental approach to achieve eosinophil-specific expression of heterologous genes. Homologous recombination in embryonic stem cells was used to create a knock-in strain of mice (eoCRE) with a chimeric locus that encodes a mammalianized Cre recombinase from the endogenous AUG start codon of the Epx ORF. The subsequent characterization of these mice reported here provides unambiguous evidence for the availability of a model that in conjunction with strains of mice carrying floxed genes of interest now permits eosinophil lineage-specific targeting of gene expression.

MATERIALS AND METHODS

Mice

Knock-in mice (eoCRE) with an ORF encoding a mammalianized Cre recombinase [17] inserted at the AUG start codon of Epx were generated by Ozgene (Bentley, Australia) on a C57BL/6J background. B6.Cg-Gt(ROSA)26Sor^{tm6(CAG-ZsGreen1)Hze}/J [18], B6.129P2-Gt(ROSA)26Sor^tm1(DTA)Lky/J [19], and WT C57BL/6J mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). Eosinophil-less PHIL mice [9], maintained by continual crossing with C57BL/6J animals (n>20 generations), were bred in the Mayo Clinic Arizona Small Animal Facility. Hemizygous experimental eoCRE mice were generated by crossing of homozygous or hemizygous eoCRE mice with WT C57BL/6J animals. All mice were maintained in ventilated microisolator cages and housed in the Mayo Clinic Arizona Small Animal Facility (a specific pathogen-free animal facility). Protocols and studies involving animals were performed in accordance with National Institutes of Health and Mayo Foundation institutional guidelines.

Hematologic assays: cytospins/smear preparations and cell counts/differentials

Summaries and full descriptions of all of the hematological methods used are outlined in detail in our recently published laboratory manual (i.e., Mouse Hematology [20]). It is noteworthy that blood films and cytospins were stained with a Diff-Quick stain set (Siemens Healthcare Diagnostics, Newark, DE, USA) for differential analyses, and WBC numbers were generally quantified manually using a hemocytometer.

Flow cytometric analysis and cell sorting

Single-cell suspensions of peripheral blood, bone marrow, and spleen were depleted of erythrocytes (Pharmlyse; BD Biosciences, San Jose, CA, USA) prior to use; in the case of peritoneal lavage cells, this depletion was not required. All recovered cell suspensions were stained for 25 min on ice with cell type-specific antibodies after blockade of FcRs using 1 μg/μl Fc blocker (CD16/32; BD Biosciences). The cell-surface definitions of various leukocyte populations described here include: (1) differentiated peripheral eosinophils—IL-5Rα-FITC⁺ or PE⁺ (T21; BD Biosciences), CCR3-Alexa 647⁺ or FITC⁺ (83101; R&D Systems, Minneapolis, MN, USA), and PI⁻ (BD Biosciences); (2) EoPs [21]—a lineage cocktail was used to exclude other cells, as well as to identify these progenitors initially, and this lineage cocktail (eBioscience, San Diego, CA, USA) included antibodies specific for CD3 (145-2C11), CD4 (GK1.5), CD8α (53-6.7), B220 (RA3-6B2), CD19 (1D3), Gr1 (RB6-8C5), and Ter119 (Ly-7d). Additional antibodies/markers used in the identification of EoPs included ScaI-PE-Cy7⁻ (D7; eBioscience), CD34-eFluor 450⁺ (RAM34; eBioscience), c-Kit-APC⁺ (2B8; eBioscience), IL-5Rα-PE⁺ (T21; BD Biosciences), and PI⁻; (3) basophils, as described previously [22]—B220-PE-Cy7⁻ (RA3-6B2; BD Biosciences), CD4-PE-Cy7⁻ (GK1.5; BD Biosciences), CD8-PE-Cy7⁻ (53-6.7; eBioscience), GR1-PE-Cy7⁻ (RB6-8C5; eBioscience), c-Kit-V450⁻ (2B8; BD Biosciences), CD49b-APC⁺ (DX5; eBioscience), FcϵRIα-PE⁺ (MAR-1; eBioscience), PI⁻ (BD Biosciences); (4) B cells—B220-PE⁺ (RA3-6B2; eBioscience), PI⁻; (5) DCs—CD11c-APC⁺ (HL3; BD PharMingen, San Diego, CA, USA), F4/80-efluor450⁻ (BM8; eBioscience), PI⁻; (6) mast cells—FcϵRIα-PE⁺, c-Kit-V450⁺, PI⁻; (7) macrophages—F4/80-efluor450⁺, CD11b-APC⁺ (M1/70; eBioscience), PI⁻; (8) neutrophils—Gr1-PECy7⁺, Siglec-F-PE⁻ (E50-4220; BD Biosciences); (9) NK cells—Gr1-PECy7⁻, NK1.1-PE⁺ (PK136; BD PharMingen), CD49b-APC⁺ (HMα2; BD PharMingen), PI⁻; (10) T cells—TCRβ-PE⁺ (H57-597; BD PharMingen), PI⁻.

Flow cytometry was performed with a cytofluorimeter (CyAn; Dako, Carpinteria, CA, USA; or LSR Fortessa; BD Biosciences). Data acquisition and analysis were performed using Summit (version 4.3; Dako) or FACSDiva version 6.2 software (BD Biosciences). Sorting of GFP⁺ peripheral blood cells was performed using a cytofluorimeter (FACSAria; BD Biosciences) using FACSDiva version 6.1.1 software (BD Biosciences).

EPX ELISA

EPX ELISA was performed as described previously [23].

Ex vivo generation of eosinophils by targeted proliferation/differentiation of bone marrow progenitors

Whole marrow (femoral and tibial cells) was cultured under conditions optimal for eosinophil development using a modified version of a methodology described previously [12]. Briefly, one femur and tibia were recovered and cleaned of attached musculature, and the epiphyses were removed, essentially turning each bone into a hollow tube [24]. Marrow was recovered from each bone with 1 ml supplemented RPMI culture media, as described previously [12], with the substitution of embryonic stem cell-qualified FBS (Invitrogen, Carlsbad, CA, USA). Clumps of marrow were disrupted to single cells by repeated pipetting, red blood cells contaminating this preparation of leukocytes were removed by hypotonic shock using ice cold distilled water, and the remaining marrow leukocytes were passed through a 40-μm cell strainer. The recovered marrow leukocytes were resuspended at a density of 10⁶ cells/ml cell concentration, with RPMI culture media supplemented further with 100 ng/ml stem cell factor (PeproTech, Rocky Hill, NJ, USA) and 100 ng/mL Fms-related tyrosine kinase 3 ligand (PeproTech). Marrow leukocytes (9×10⁶) were transferred to 25 cm² culture flasks (Corning, Corning, NY, USA) and were left undisturbed in culture at 37°C–5% CO₂ during Protocol Days 0–4. On Protocol Day 4, unattached cells were recovered by centrifugation [4°C for 10 min (170 g)] and resuspended in RPMI culture media, supplemented with 10 ng/ml mouse rIL-5 (R&D Systems), before returning to culture (again undisturbed) at 37°C–5% CO₂. On Protocol Days 8, 10, and 12, unattached cells were counted and once again, recovered by centrifugation [4°C for 10 min (170 g)]. One-half of the new volume required to adjust the culture to 10⁶ cells/ml was discarded and replaced with fresh RPMI culture media, supplemented with 20 ng/ml mouse rIL-5. A new culture flask was used with each counting of cells and changing of media (i.e., Protocol Days 8, 10, and 12), ensuring removal of any attached stromal cells.

Statistical analysis

Data were analyzed and graphed using GraphPad Prism statistics program (GraphPad Software, San Diego, CA, USA). Results are presented as means ± sem. Statistical analysis was performed using t-tests, with differences between means considered significant when P < 0.05.

RESULTS AND DISCUSSION

Generation of the eoCRE mouse

A direct gene knock-in approach (reviewed in ref. [25]), inserting an ORF encoding a mammalianized Cre recombinase into Epx, was chosen to exploit the high level and eosinophil-specific expression occurring at this endogenous locus on chromosome 11. As shown in Fig. 1, a complete ORF encoding a mammalianized Cre-recombinase protein was inserted by homologous recombination in embryonic stem cells at the endogenous AUG start codon of Epx. This locus was further modified by the insertion of an IRES preceding the exons encoding the complete EPX, generating eoCRE mice expressing a dicistronic mRNA that in theory, would encode the mammalianized Cre and EPX. RT-PCR assessments of bone marrow gene expression provided evidence of Cre-EPX transcripts derived from this modified locus (data not shown). However, enzymatic and ELISA data derived from hemizygous or homozygous eoCRE mice showed that the IRES element was insufficient to rescue EPX expression from the knock-in locus (Supplemental Fig. 1). Thus, all subsequent studies described here were performed with eoCRE hemizygous mice that displayed EPX expression levels within a factor of two of WT C57BL/6J animals.

Figure 1. Homologous recombination in embryonic stem cells was used to generate gene knock-in mice targeting a sequence encoding a mammalianized Cre in-frame with the AUG start codon of Epx.

Restriction maps of the WT Epx locus, eoCRE knock-in targeting construct, and the resulting chimeric Cre-Epx locus provide a schematic outline of the genetic manipulations that were performed in embryonic stem cells. These stem-cell manipulations led to first-generation knock-in mice with a recombinant Epx allele. Subsequent breeding of these first generation animals with germline flippase “deleter” mice removed the flippase recognition target (FRT)-flanked drug selection cassette incorporated into the locus generating eoCRE mice with its characteristic Cre-Epx chimeric locus. PGK-Neo, phosphoglucokinase gene promoter driven expression of neomycin; TGA, opal stop codon.

Expression of Cre recombinase in eoCRE mice occurs in nearly all eosinophils and is absolutely eosinophil-specific

Hematological examinations (i.e., cell counts and differentials) of eoCRE mice demonstrated that the baseline levels of circulating leukocytes in these knock-in animals were unaffected relative to WT (Fig. 2A). More specifically, flow cytometric assessments, focused on the eosinophil lineage, revealed that the insertion of Cre recombinase had no cytotoxic effects on these cells, with eoCRE mice again displaying equivalent levels of eosinophils in peripheral blood (Fig. 2B and C) and bone marrow (Fig. 2D and E) relative to WT mice; eosinophil-less PHIL mice served as a negative control.

Figure 2. eoCRE mice displayed WT levels of eosinophils with no evidence of toxicity associated with the genetic manipulation of the Epx locus.

(A) Cell counts and differentials of peripheral blood leukocytes showed that WBC counts and the composition of various leukocyte subtypes of eoCRE mice were essentially identical to the levels observed in WT mice (n=4 animals/group). Flow cytometric assessments of peripheral blood leukocytes (B, C) or bone marrow progenitors/mature leukocytes (D, E) derived from individual mice showed that unlike eosinophil-less PHIL mice [9], the number of eosinophils (i.e., CCR3⁺/IL-5Rα⁺) in eoCRE animals was not significantly different from the number observed in WT mice. *P < 0.001.

eoCRE mice were crossed with a (flox-stop-flox)-GFP reporter strain (B6.Cg-Gt(ROSA)26Sor^{tm6(CAG-ZsGreen1)Hze}/J) [18], generating animals that were hemizygous at both loci (i.e., eoCRE^+/−/GFP^+/−) as a means of testing the functionality of the genetic manipulations linked with inserting the Cre recombinase into the Epx locus. Peripheral blood eosinophil levels, as a percentage of WBC by flow cytometry, were determined using an eosinophil-specific cell-surface marker profile (CCR3⁺, IL-5Rα⁺) and/or via GFP gene expression (Fig. 3A). These data showed that eosinophil levels in the blood of eoCRE^+/−/GFP^+/− mice (using antibodies recognizing an eosinophil-specific cell-surface marker profile) were the same as the levels observed in various control cohorts, including WT, eoCRE^+/−, and the (flox-stop-flox)-GFP^+/− reporter mice. In contrast, when using GFP expression as the defining flow marker, eosinophils were identified only in eoCRE^+/−/GFP^+/− mice. More detailed flow cytometric assessments of leukocytes in eoCRE^+/−/GFP^+/− mice revealed that 96.5±1.4% of cells within a gate of the putative eosinophil population (CCR3⁺/IL-5Rα⁺) had undergone successful Cre-mediated recombination and were positive for GFP (Fig. 3B). Alternatively, assessments of peripheral blood leukocytes from eoCRE^+/−/GFP^+/− mice, based on an initial GFP-positive gate, showed that 99.5 ± 0.3% of all GFP-positive cells were eosinophils, as determined by their eosinophil-specific cell-surface marker profile (Fig. 3C). Moreover, assessments of the much larger, GFP-negative cell gate showed that CCR3⁺/IL-5Rα⁺ cells (i.e., eosinophils) were absent in this population (Fig. 3C). Indeed, flow cytometric studies of individual noneosinophil populations in eoCRE mice (e.g., basophils, B cells, DCs, mast cells, macrophages, neutrophils, NK cells, and T cells) showed that each of these leukocyte populations contained no GFP-positive cells, further demonstrating the eosinophil specificity of Cre expression (Supplemental Fig. 2). It is significant that cytospin preparations of GFP-positive peripheral blood eosinophils from eoCRE^+/−/GFP^+/− mice showed that these cells stain recognizably as eosinophils using traditional dye sets (e.g., Diff-Quick or Leukostat), and the expression of GFP is so high that they are visualized easily by compound microscopy (Fig. 3D). Indeed, as expected, our preliminary microscopy studies (data not shown) identify GFP-positive, resident eosinophils in tissue compartments, identified previously in earlier studies (e.g., gastrointestinal tract) [26]. The importance and use of this observation for future, more complex in vivo microscopy studies examining eosinophil trafficking and cell–cell interactions are difficult to overestimate.

Figure 3. Cre-recombinase expression in eoCRE mice was restricted to eosinophil-lineage-committed cells and occurred in nearly all eosinophils.

(A) Flow cytometric assessments of peripheral blood leukocytes of mice, resulting from a cross of eoCRE^+/− and (flox-stop-flox)-GFP^+/− mice (eoCRE^+/−/GFP^+/−), as well as assessments of the parental single knock-in strains of mice [i.e., eoCRE^+/− and (flox-stop-flox)-GFP^+/−], showed that all genotypes displayed WT levels of eosinophils using a defined eosinophil-specific cell-surface marker profile (CCR3⁺/IL-5Rα⁺). However, if the leukocytes are stratified using GFP expression as the defining gate, only eoCRE^+/−/GFP^+/− mice displayed evidence of equivalent levels of eosinophils. (B) Flow cytometric assessments of GFP in peripheral blood leukocytes from eoCRE^+/−/GFP^+/− mice following an initial gate using an established eosinophil-specific cell-surface marker profile (CCR3⁺/IL-5Rα⁺). FSC, Forward-scatter. (C) Flow cytometric assessments of peripheral blood leukocytes from eoCRE^+/−/GFP^+/− mice following an initial gate based on GFP expression (GFP-positive) were characterized as eosinophils versus noneosinophils using an eosinophil-specific cell-surface marker profile (CCR3⁺/IL-5Rα⁺). (D) Cytospin preparations of sorted GFP-positive peripheral blood eosinophils from eoCRE^+/−/GFP^+/− mice stained recognizably as eosinophils using traditional dye sets, such as Diff-Quick and Leukostat, were easily visualized by microscopy based on their extraordinarily intense fluorescence. Original scale bars = 10 μm.

eoCRE mice represent a novel model system to target eosinophil-specific gene expression

Flow cytometric assessment of marrow-derived leukocytes from eoCRE^+/−/GFP^+/− mice showed that GFP expression occurs in the myeloid EoP population, the earliest identifiable eosinophil-lineage committed cells (Fig. 4A). Moreover, the expansion and differentiation of eosinophils from hematopoietic progenitor cultures of eoCRE^+/−/GFP^+/− bone marrow also displayed no kinetic or steady-state population differences relative to cultures of WT marrow (Fig. 4B). Specifically, on the final day of culture (i.e., Day 14), 98.6 ± 0.6% of these eoCRE^+/−/GFP^+/− cells were identified as eosinophils following Diff-Quick staining and 93.8±0.1% were GFP-positive by flow cytometry. Collectively, these data suggest no lineage toxicity as a consequence of Cre or GFP expression. The usefulness of this eosinophil lineage-restricted expression of Cre recombinase was demonstrated by crossing eoCRE with a ROSA26 knock-in strain of mice containing a (flox-stop-flox)-DTA gene [19]; successful Cre-mediated recombination of this locus in eoCRE^+/−/DTA^+/− mice should elicit DTA gene expression and in turn, eosinophil lineage-specific cell death. Indeed, flow cytometric assessments revealed that the peripheral blood of eoCRE^+/−/DTA^+/− mice (Fig. 4C) were demonstrably deficient of eosinophils relative to WT animals (0.19±0.04% vs. 1.85±0.22%, respectively); cell differential assessments of blood films prepared from these mice also showed no differences in circulating cell numbers and relative composition of other (i.e., noneosinophil) leukocyte populations, including lymphocytes, neutrophils, and monocytes. The extent of this eosinophil deficiency (Fig. 4D) was comparable to that observed in PHIL transgenic mice [9], a previously established strain of mice congenitally deficient of eosinophils (0.19±0.04% vs. 0.06±0.00%, respectively). This observation is instructive, as it demonstrated, with yet another floxed gene cassette, the near-complete efficiency with which eoCRE mice target eosinophils. Moreover, these studies had the added benefit of generating an independent “gold-standard” line of mice congenitally deficient of eosinophils.

Figure 4. Cre-recombinase expression in eoCRE mice occurred in EoPs, and Cre-targeted expression of a cytotoxin leads to the generation of mice devoid of eosinophils.

(A) Flow cytometric assessment of bone marrow-derived progenitors demonstrated that GFP expression in eoCRE mice extends to the earliest EoP population. (B) The IL-5-dependent ex vivo differentiation/expansion of eosinophils from eoCRE marrow-derived progenitors was equivalent to the marrow culture growth kinetics observed with WT marrow progenitors. (C) Flow cytometric assessments of peripheral blood leukocytes identifying eosinophils from mice resulting from a cross of eoCRE^+/− and the (flox-stop-flox)-DTA^+/− mice (eoCRE^+/−/DTA^+/−); WT animals and eosinophil-less PHIL mice were used as positive and negative controls, respectively. (D) Quantitative analyses of the flow cytometric assessments of peripheral blood demonstrated that Cre-mediated expression of the DTA gene in eoCRE^+/−/DTA^+/− mice resulted in the ablation of eosinophils. *P < 0.001.

Epilogue

Several observations and/or conclusions are noteworthy, as they provide a larger context for the use of this novel strain of mice. (1) eoCRE mice express sufficient levels of Cre recombinase to mediate loxP site-dependent recombination exclusively in eosinophils with >95% of all eosinophils displaying this Cre-mediated recombination. (2) The functionality of Cre-mediated recombination in eosinophils occurs even in hemizygous eoCRE mice (i.e., eoCRE^+/−). Thus, despite the unexpected loss of EPX expression occurring as a consequence of our knock-in strategy at this locus, the generation of animals carrying floxed alleles of a given gene of interest (i.e., eoCRE^+/−/gene-of-interest^flox/flox mice) will allow the production of knock-out animals with targeted gene deletions in all eosinophils, while still maintaining EPX expression. (3) The ability to tag eosinophils exclusively in vivo with a fluorescent marker now provides an unparalleled opportunity to assess eosinophil recruitment and localized tissue accumulation in studies using mouse models of disease. In addition, the availability of these mice will almost certainly expand investigations into cell–cell interactions surrounding eosinophil effector functions. (4) Our demonstration of the use of eoCRE mice to mediate eosinophil-specific expression following crosses with (flox-stop-flox)-GFP or (flox-stop-flox)-DTA mice provides evidence of a general mouse model-based strategy to ectopically overexpress heterologous genes exclusively in eosinophil lineage-committed cells. Namely, the availability of (flox-stop-flox)-ROSA26 knock-in mice [27] or potentially (flox-stop-flox)-GAG, promoter-driven transgenic mice [28] will allow the targeted expression of any candidate gene in eosinophils.

AUTHORSHIP

J.J.L., R.S.P., and N.A.L. conceived of and developed the eoCRE strain of mice. A.D.D., N.A.L., and J.J.L. wrote this manuscript. A.D.D. and E.A.J. were the primary investigators who collected and analyzed data. S.I.O., L.W., K.S., J.N., J.K., and W.E.L. provided technical support to perform studies and assisted with the analysis of data. A.D.D., E.A.J., S.I.O., P.L., R.M., N.A.L., and J.J.L. revised various drafts of this manuscript, leading to the submitted paper for publication.

DISCLOSURES

The authors declare no competing financial interests. Funding sources had no involvement in study design, data collection (including analysis and interpretation), the writing of the manuscript, or the decision to submit for publication. Copyright transfer is subject to applicable Mayo terms (located at http://www.mayo.edu/copyright/).