Off-Target Deletion of Conditional Dbc1 Allele in the Foxp3^YFP-Cre Mouse Line under Specific Setting

Abstract

The Cre-LoxP conditional knockout strategy has been used extensively to study gene function in a specific cell-type. In this study, the authors tried to engineer mice in which the Dbc1 gene is conditionally knocked out in Treg cells. Unexpectedly, the conditional Dbc1 allele was completely deleted with a low frequency in some Foxp3^YFP-Cre mice harboring floxed Dbc1 allele under specific settings. It was found that the germline recombination of floxed Dbc1 allele, which caused Dbc1 knock out mice, occurred in the male Foxp3^YFP-Cre mice harboring floxed Dbc1 allele. Even though the authors documented that Foxp3 is expressed in the testis, the germline recombination was not caused by the germline expression of Cre, which was driven by the Foxp3 promoter. The germline recombination may be caused by the unspecific expression of Cre recombinase in the fetus, in which the floxed Dbc1 allele of some stem cells with development potential to germ cells may be recombined. Additionally, this study found that the floxed Dbc1 allele was recombined in non-T cells of some Foxp3^Cre Dbc1^fl mice, which need to be characterized. Our results also suggest that using male mice with a low frequency of recombined gene allele can reduce the risk of having full knock out mice.

1. Introduction

Cre–LoxP technology has been utilized extensively for conditional knockouts in yeast [1], plants [2], cultured mammalian cells [3], T cells [4], neuronal tissue [5] and so forth. Conditional knockout mice using this technology are constructed by breeding a floxed (i.e., flanked by LoxP) gene of interest with mice expressing Cre-recombinase driven by a tissue or cell-type specific promoter. Some examples include the construction of CD4 positive cell knockouts of Irf4 by breeding Cd4^Cre mice with those containing floxed Irf4 sites [6]; breeding dendritic cell driven CD11c^Cre mice with floxed TGFRβII [7] and the construction of neuronal tissue knockouts of Rem2 by breeding Rgs9^Cre with Rem2^LoxP/LoxP mice [8].

Although the Cre-Lox conditional knockout method is an ideal tool for the study of a gene of interest in a specific cell type, some limitations exist [9]. One limitation is the non-specific expression of the targeted gene resulting in off-target recombination in other cell types [10]. Another limitation is the potential for germline recombination of conditional alleles due to the transient Cre expression in the germline [8,9,11,12,13,14]. Thus, a careful characterization of mice generated with this system is an absolute requirement to confirm proper control of conditional recombination.

Treg cells have been recognized to be crucial in immune tolerance and the prevention of autoimmune responses [15,16]. Foxp3^YFP-Cre (or Foxp3^Cre) mice were constructed by Rubtsov et al. [17] to study genes related to regulatory T (Treg) cell biology. In the Foxp3^YFP-Cre mouse, the Cre gene is located after an internal ribosome entry site (IRES) sequence which is inserted just after the normal termination codon of Foxp3. This mouse was used by several groups to study the function of genes such as Uqcrsf1 [18], Id2 [19], Foxp1 [20], Maf [21] in Treg cells. However, Rubtsov et al. [17] and Franckaert et al. [22] have reported the off-target recombination in some hematopoietic lineage cells of a Foxp3^YFP-Cre mouse.

The authors have found that the DBC1 is involved in regulatory T-cell function under conditions of inflammation [23]. To study the specific function of Dbc1 in regulatory T-cells, Foxp3^YFP-Cre mice were crossed with floxed Dbc1 mice to gain the regulatory T-cell knockouts of Dbc1. During the breeding process, it was unexpectedly found that the floxed Dbc1 allele in some mice was totally deleted and the off-target recombination of floxed Dbc1 allele occurred in tissues other than T cells.

2. Materials and Methods

2.1. Mice

All mouse studies were carried out in strict accordance with the guidelines of the Institutional Animal Care and Use Committee at the Sun Yat-sen University. All mice were maintained under specific pathogen-free conditions. The mouse line carrying the Foxp3^YFP-Cre [17], CD4^Cre [24] or Pkm2^fl/fl [25] allele was purchased from the International Mouse Knockout Consortium. The Dbc1 conditional knockout mice, which contain LoxP sequences flanking exons 4, 5, 6, and 7 were constructed by Shanghai Model Organisms Center, Inc. The breeding strategy is indicated in the text.

2.2. Genomic PCR

The genomic DNA (gDNA) was prepared from the toe clips or tail snips of mice. The tissues were lysed by incubation with lysis buffer [100 mM Tris HCl (pH 7.8), 5 mM EDTA, 0.2% SDS, 200 mM NaCl and 100 µg/mL proteinase K] overnight at 56 °C, followed by centrifugation at a speed of 10,000 g for five minutes to obtain supernatants with genomic DNA. The DNA was precipitated by isopropanol, washed in 70% ethanol and dissolved in deionized water. The wild type Dbc1 allele and the floxed Dbc1 allele were genotyped using the following two primers: 1 (P1, 5’-TCTCACTGTAGCGCAGCCTGAC-3’) and 2 (P2, 5’-ACCCAGCAGAGTACAACAGAAGACAC-3’). The recombined Dbc1 allele was genotyped by using primer 0 (P0, 5’-CCTGCAGCCCAATTCCGATCA-3’) and primer 2. The wild type Foxp3 allele and Foxp3-IRES-YFP-Cre allele (Foxp3^YFP-Cre or Foxp3^Cre) were genotyped using the following primer set: (Wt-fwd: 5’-CTATGGAAACCGGGCGATGA-3’, Wt-rev: 5’-AGTGGCAAGTGAGACGTGGG-3’, Cre-fwd: 5’-AGGATGTGAGGGACTACCTCCTGTA-3’, Cre-rev: 5’- TCCTTCACTCTGATTCTGGCAATTT-3’). The PCR product of the wild type Foxp3 allele is 630 bp, while the Foxp3-IRES-YFP-Cre allele is 346 bp. The wild type Cd4 allele and Cd4-Cre allele (Cd4^Cre) were genotyped using the following primer set: (Cd4-Cre-Co: 5’-AACTTGCACAGCTCAGAATGC-3’, Cd4-Cre-Wt: 5’-ACCTGAGATTCCACCAAACTTGA-3’, Cd4-Cre-Mu: 5’-TTAGGGTGGGGCTCAGAAGG-3’) [26]. The PCR product of the wild type Cd4 allele is 560 bp, while the Cd4-Cre allele is 741 bp. The wild type Pkm2 allele and floxed Pkm2 allele were genotyped using the following primer set: (P-F: 5’- CCAAAGGATTCCCTTGGGCACAG -3’, P-R: 5’- GCTTTGTCAGAGCTTTGTCACAAATGG-3’). The PCR product of the deleted Dbc1 fragment was purified using SanPrep Column DNA Gel Extraction Kit (Sangon Biotech Co., Ltd., Shanghai, China) and sequenced by Beijing TsingKe Biotech Co., Ltd in China.

2.3. Quantitative PCR (qPCR)

To quantify the floxed or deleted Dbc1 alleles, the tissues were dissected and gDNA was purified using the Ezup Column Animal Genomic DNA Purification Kit (Sangon Biotech Co., Ltd., Shanghai, China) according to the instructions provided by the manufacturer [27]. The purified DNA was analyzed by real-time PCR analysis with SYBR Green PCR master mix (Takara Biomedical Technology (Beijing) Co., Ltd., Beijing, China). For the floxed Dbc1 allele, the primer set was Q-mgDbc1-F (5’-TACGAAGTTATTAGGTCCCTC-3’) and Q-mgDbc1-R(5’-CTTGACTTCATCCAAACCG-3’). For the deleted Dbc1 allele, the primer set was Q -mzDbc1-F (5’-CCAATTCCGATCATATTCAATAAC-3’) and Q-mzDbc1-R (5’-ATAAGCAAGGAAGGTCTGA-3’). The fragment amplified by the primer set of Q-miDbc1-F (5’-AGAATACTCCACTTTCCCTG-3’) and Q-miDbc1-R (5’-GTAGACCTCATAAGTGGAGG-3’), which targeted the intron 15 of Dbc1 served as a normalization control. The quantification of the floxed Dbc1 allele was normalized against Dbc1^fl/+ mice, while the deleted Dbc1 allele was normalized against the mouse #4 (Dbc1^-/+). For the Foxp3 and Cre mRNA level, the primer set qmFoxp3-F (5’-CCCATCCCCAGGAGTCTTG-3’) and qmFoxp3-R (5’-ACCATGACTAGGGGCACTGTA-3’) was used to amplify Foxp3 [28]. The primer set qmCre-F (5’-CCTTTGAACGCACTGACTTTG-3’) and qmCre-R (5’-GTCCTTCACTCTGATTCTGGC-3’) was used to amplify Cre. The fragment amplified by the primer set of qmβActin-F (5’-GGCTGTATTCCCCTCCATCG-3’) and qmβActin-R (5’-CCAGTTGGTAACAATGCCATGT-3’), which targeted Actb served as a normalization control. The amplification was quantified using the 2^−ΔΔCT method. The qPCR data was plotted as the mean ± standard deviation and analyzed by two-way ANOVA. For all tests, α was set at 0.05.

2.4. Immunoblot Analysis

The tissues were dissected and the cells were lysed in RIPA buffer containing 20 mM Tris/HCl, pH 7.5, 1% Nonidet P-40, 0.5% Na-deoxycholate, 135 mM NaCl, 1 mM EDTA, 10% glycerol with 1 mM PMSF, 1 mM Na₃VO₄, 1 mM NaF and cocktail protease inhibitor (Sigma, St. Louis, MO, USA). The samples were analyzed by western blotting using the standard procedures [29]. The targeted protein was detected by immunoblot with anti-Foxp3 (14-7979-82; Invitrogen, Carlsbad, CA, USA), anti-GAPDH (#5174; Cell Signaling Technology, Inc., Danvers, MA, USA), and ECL western blotting detection reagents were used (Millipore, Darmstadt, Germany) and visualized using a Tanon 5500 Chemiluminescent Imaging System.

3. Results

3.1. The Conditional Dbc1 Allele Was Completely Knocked Out in Some Progeny of Foxp3^Cre/Y Dbc1^fl/+ × Foxp3^Cre/Cre Dbc1^fl/Fl Mice

It was found that Dbc1 is involved in the maintenance of Treg in an inflammatory environment [23]. To study the function of Dbc1 in a specific cell type, a transgenic mouse strain carrying a LoxP-flanked (floxed) Dbc1 allele was generated, in which the LoxP sites were placed upstream of exon 4 and downstream of exon 7 (Figure 1a). In these mice, the wild type or floxed Dbc1 allele was identified by a PCR assay using primer 1 (P1) and primer 2 (P2) (Figure 1b). The PCR product of the wild type Dbc1 allele is 384 bp, while the floxed Dbc1 allele is 488 bp. The mice harboring wild type and floxed Dbc1 alleles are called Dbc1^fl/+ mice, while the mice containing homozygous floxed Dbc1 alleles are called Dbc1^fl/fl mice in this study.

In order to knockout Dbc1 specifically in Treg cells, the conditional Dbc1 mice were crossed with Foxp3^YFP-Cre mice [17], in which the Cre recombinase was under the control of Foxp3. In this study, the male mice harboring Foxp3^YFP-Cre allele are designated Foxp3^Cre/Y while the female mice are designated as Foxp3^Cre/Cre or Foxp3^Cre/+. When the female mice homozygous for floxed Dbc1 and Foxp3^Cre alleles were crossed with the male mice heterozygous for floxed Dbc1 and hemizygous for Foxp3^Cre (Foxp3^Cre/Y Dbc1^fl/+ × Foxp3^Cre/Cre Dbc1^fl/fl), it was found that some progeny contained no floxed Dbc1 alleles, such as the mouse #4 in Figure 2a. This phenomenon could be explained by the fact that the floxed Dbc1 allele was totally deleted, which could not be detected using primers 1 and 2. To verify this hypothesis, another primer (P0) (Figure 1) was designed to test the deleted Dbc1 allele. The results showed that the floxed Dbc1 allele of mouse #4 was deleted in the tail, muscle, brain, liver, immune organs and the reproductive system (Figure 2b). These results demonstrated that the floxed Dbc1 allele changed to deleted Dbc1 allele in mouse #4. As the deleted Dbc1 allele could be also present in the Dbc1^fl/fl mice (which would not be detected using P1 and P2), the Dbc1 allele of other mice homozygous for floxed Dbc1 allele were also analyzed using P0 and P2. The results showed that two mice (#5 and #6) had a deleted Dbc1 fragment (Figure 2c). Since the deleted Dbc1 fragment may be from Treg cells, qPCR was used to verify how many floxed Dbc1 allele were recombined. The results showed that mouse #5 had an equal frequency of deleted Dbc1 fragment to mouse #4 and floxed Dbc1 fragment to Dbc1^fl/+ mouse, while mouse #6 was not (Figure 2c). This suggests that one of the floxed Dbc1 allele in mouse #5 may have changed to the deleted Dbc1 allele. To further confirm the deleted Dbc1 allele, the two mice (#4 and #5) were back crossed to wild type mice (Wt). Interestingly, the results confirmed that indeed some of the offspring had deleted Dbc1 allele (Figure S1a,b).

To identify whether the deleted Dbc1 allele was derived from the parent mice, the Dbc1 alleles of the parent mice were further analyzed. It was observed that one of the mothers also contained a deleted Dbc1 allele (Figure S1c). Therefore, the deleted Dbc1 alleles in mouse #4 and #5 are derived from the mother 1 mouse. These results suggest that using the Cre/Lox conditional knockout method in the context of Foxp3 may indeed result in off-target deletion effects.

To verify the deleted Dbc1 allele, the deleted Dbc1 fragment (279 bp) was purified and sequenced. The sequence of deleted Dbc1 fragment is just the sequence of deleted Dbc1 allele, which supports that the floxed Dbc1 alleles in mouse #4 or #5 have changed to deleted Dbc1 alleles (Figure S1d).

This study also observed some mice (Figure S2, mouse #P-18 and #P-21) without the floxed Pkm2 allele in the progeny of male Foxp3^Cre/Y Pkm2^fl/+ mice crossed with female Foxp3^Cre/Cre Pkm2^fl/fl mice, in which the Pkm2^fl/fl floxed mutant mice were from the Jackson Laboratory [25]. These results suggest that using the Cre/Lox conditional knockout method in the context of Foxp3 may result in some mice with gene knock out.

3.2. Germline Recombination of Floxed Dbc1 Allele Occurs in the Male Foxp3^Cre/Y Dbc1^fl/+ Mice

The germline recombination of floxed Dbc1 allele can result in the progeny with a deleted Dbc1 allele. To verify this and ascertain whether the germline recombination occurred in male mice or female mice, different genotypes of male or female Foxp3^Cre Dbc1^fl mice were used to cross with Wt mice. To rule out that the deleted Dbc1 allele in the progeny was derived from the maternal/paternal mice with deleted Dbc1 allele, the Foxp3^Cre mice heterozygous for floxed Dbc1 allele were chosen to breed (

Table 1. Frequency of mice with deleted Dbc1 allele in the progeny of different breeding pairs.

Breeding Mice	Mouse No.	Frequency of Recombined Dbc1 Fragment	Litter Size	Number of Dbc1−/+ Mice in Progeny
Foxp3Cre/YDbc1fl/+ × Wt	YN-3, 4, 5	low	89	0
Foxp3Cre/YDbc1fl/+ × Wt	YN-1, 2	high	67	2
Wt × Foxp3Cre/+ Dbc1fl/+	VN-3, 4	low	55	0
Wt × Foxp3Cre/+ Dbc1fl/+	VN-1, 2, 5, 6	high	64	0
Foxp3Cre/YDbc1fl/fl × Wt	YM-1, 2	low	75	0
Wt × Foxp3Cre/Cre Dbc1fl/fl	XM-1, 2	low	79	0

“high” means higher frequency of recombined Dbc1 fragment than Cd4^Cre Dbc1^fl mice tested by qPCR; “low” means lower frequency of recombined Dbc1 fragment than Cd4^Cre Dbc1^fl mice tested by qPCR (Figure S4).

).

After identifying the genotypes of progeny mice using primers P1 and P2 or P0 and P2, this study found 1.28% (2 out of 156) mice (#13 and #26) only contained the deleted Dbc1 allele in the offspring of Foxp3^Cre/Y Dbc1^fl/+ × Wt mice (Figure 3a, Figure S3a). The results of qPCR also were consistent with mouse #13 containing deleted Dbc1 allele (Figure 3b). It was also observed that a deleted Dbc1 allele could be found in the offspring of Foxp3^Cre/+ Dbc1^+/+ × Foxp3^Cre/Y Dbc1^fl/+ mice (Figure S3b). In contrast, the genotypes of 119 mice in the offspring of Wt × Foxp3^Cre/+ Dbc1^fl/+ mice were identified, but no mice were obtained with the deleted Dbc1 allele (Table 1; some results are shown in Figure 3c). Considering that the frequency of the deleted Dbc1 allele might be lower when mice heterozygous for floxed Dbc1 allele were used and to rule out that the deleted Dbc1 allele in the progeny was derived from the maternal mice with deleted Dbc1 allele, the female Foxp3^Cre/Cre Dbc1^fl/fl mice with a lower frequency of deleted Dbc1 fragment than Cd4^Cre/Cre Dbc1^fl/fl mice were also used to cross Wt male mice (Table 1). By genotyping 79 mice in the progeny, not one mouse could be found with deleted Dbc1 allele (Table 1, some results shown in Figure S3c). All the results demonstrated that germline recombination of floxed Dbc1 allele could occur in the male Foxp3^Cre/Y Dbc1^fl/+ mice, which could result in progeny with deleted Dbc1 allele.

3.3. The Floxed Dbc1 Allele Was Recombined in non-T Cells of Some Foxp3^Cre Dbc1^fl Mice

In the process of genotyping, it was observed that some Foxp3^Cre Dbc1^fl mice showed an obvious deleted Dbc1 fragment, which was amplified by primers P0 and P2, while the others were not (Figure S3). This suggests that the recombination of floxed Dbc1 allele in these Foxp3^Cre Dbc1^fl mice may be various. To check that possibility, qPCR was first used to quantify the recombined Dbc1 allele in toe and tail, which were tissues used for mice genotyping. The results demonstrated that mice expressing a recombined Dbc1 fragment obviously as determined by conventional PCR exhibited a greater frequency of recombined Dbc1 allele (Figure 4a). Additionally, some mice showed a higher frequency of a recombined Dbc1 fragment than Cd4^Cre/+ Dbc1^fl/+ mice. This indicates that the floxed Dbc1 allele was recombined in more than T cells in these Foxp3^Cre Dbc1^fl mice.

To investigate whether the recombination of floxed Dbc1 allele occurred in other tissues, two male mice (#B-3 and #G-3) and two female mice (#A-2 and G-1) with a high frequency of deleted Dbc1 allele in the toe and tail were chosen to measure the extent of Dbc1 deletion in the brain, muscle, liver, thymus, lymph node, testis, ovary and sperm. As predicted, the Cd4^Cre/+ Dbc1^fl/+ mice showed the highest incidence of Dbc1 deletion in the thymus and lymph node, while almost no Dbc1 deletions were detected in other tissues. The Foxp3^Cre Dbc1^fl/+ mice expressed a detectable but lower incidence of Dbc1 deletion in the immune organs than Cd4^Cre/+ Dbc1^fl/+ mice. However, the Foxp3^Cre Dbc1^fl/+ mice with an obvious deleted Dbc1 fragment in the toe and tail showed many deleted Dbc1 fragments in the other tissues, while the mice without obvious deleted Dbc1 did not (Figure 4b). All these results suggest that the floxed Dbc1 allele is recombined in non-T cells of some Foxp3^Cre Dbc1^fl mice. These mice with nonspecific recombination should be ruled out when this kind of conditional knockout mice are used.

3.4. Expression of Cre Recombinase in the Fetus Results in Germline Recombination and Dbc1 Knock out Mice

The germline recombination in Foxp3^Cre Dbc1^fl mice can be caused by the germline expression of the Cre recombinase driven by a Foxp3 promoter. Jasurda et al. [30] reported that even though Foxp3 protein was not detected, the Foxp3 transcript could be found in the testis. In the ENCODE database, the Foxp3 mRNA was shown to be present in the testis and ovary (Figure S5a). The Foxp3 protein in the testis and ovary were also detected using western blotting (Figure 5a). However, this study observed that the two mice (#13 and #26 in Figure 3a, Figure S3a, Figure S4 and Table 1) with deleted Dbc1 allele are the progeny of male Foxp3^Cre/Y Dbc1^fl/+ mice with a high frequency of the recombined Dbc1 fragment. There were no mice with deleted Dbc1 allele in the progeny of male Foxp3^Cre/Y Dbc1^fl mice with a low frequency of recombined Dbc1 fragment, even though the floxed Dbc1 allele is homozygous (Figure S5b,c; Table 1). This study also could not obtain any mice with deleted Dbc1 allele in the progeny of female Foxp3^Cre Dbc1^fl crossed with Wt male mice (Figure 3c and Figure S3c; Table 1). This means that the off-target deletion is not dependent on whether the Dbc1 allele in the breeding mouse is heterozygous or homozygous. Conversely, the off-target deletion depends on whether the breeding mice contain higher frequency of the recombined Dbc1 fragment or not. The germline recombination probably occurred at the developmental stage of the mice.

In the EMAGE database (http://www.emouseatlas.org/emap/home.html), Gray et al. [31] and Diez-Roux et al. [32] detected the Foxp3 expression in the fetus by RNA in situ hybridization. This study also detected the Foxp3 protein, Foxp3 and Cre mRNA in the fetus of Foxp3^Cre mouse at embryonic day 14.5 (Figure 5b,c). In addition, by quantifying the frequency of the recombined Dbc1 fragment, it was indeed found that some fetuses of Foxp3^Cre Dbc1^fl/+ mice (mouse #F3-2, # F3-3, # F3-5 and # F3-7) showed an obvious higher incidence of Dbc1 deletion than Cd4^Cre/+ Dbc1^fl/+ mice (Figure 5d). These results suggest that the nonspecific expression of Cre in non-T cells of the fetus could activate and excise the floxed Dbc1 allele. When some cells with deleted Dbc1 allele develop to germ cells, it can result in Dbc1 knock out mice in the offspring. Thereby, to limit the risk of having full knock out mice, the male mice with a high frequency of recombined gene allele should be ruled out as breeders.

Comparing the mice genotype in the progeny of each breeding pair, it seems that the Foxp3^Cre allele derived from a female breeder causes a higher frequency of the recombined Dbc1 fragment in the progeny (Figure 3c and Figure S3c, and Figure S5b,c). Even though the floxed Dbc1 allele was not inherited from the same female mice harboring the Foxp3^Cre allele, the floxed Dbc1 allele still could recombine in some Foxp3^Cre Dbc1^fl mice (#106 and #114) (Figure S6). These support that the Foxp3^Cre allele derived from a female breeder can more easily be expressed and cause off-target recombination in the offspring.

All the results show that there is a risk of high nonspecific recombination in the Foxp3^Cre mice. Thus, a careful characterization of Foxp3^Cre mice with the floxed gene is required.

4. Discussion

The data reported here demonstrate that Dbc1 conditional alleles can be completely deleted when mice are bred with Foxp3^Cre mice. Although the frequency is not high, it remains possible that the Dbc1 allele deletion existed in the Dbc1^fl/fl mice, while the genotype identification using primers flanking only one LoxP site could fail to detect this. Moreover, the deleted Dbc1 allele is inherited by the offspring. Additionally, this study observed the off-target deletion in the Foxp3^Cre Pkm2^fl mice. The off-target deletion may be limited to the conditional Dbc1 and Pkm2 alleles. Otherwise, the off-target deletion may also occur in Foxp3^Cre mice with other conditional gene alleles. If functional inactivation of the targeted gene requires the deletion of both alleles, it will not affect the results. Nonetheless, when expression of a single copy affects the overall gene function in cells, this off-target deletion needs to be carefully characterized.

It was reported that germline recombination occurred in the Rgs9^cre line, and that the germline Cre expression could cause this recombination [8]. This study found that Foxp3 was expressed in the ovary and testis, a finding which is consistent with that of Jasurda et al. [30]. However, the probable Cre expression following the Foxp3 expression in the testis or ovary of Foxp3^cre mice did not cause germline recombination. Schmidt-Supprian et al. [33] reported that some floxed genes are more easily be excised than others. Therefore, the Foxp3 expression in the testis or ovary of Foxp3^cre mice did not cause germline recombination which might be either as a result of Cre’s unequal sensitivity for conditional gene alleles, or as a result of Cre’s inability to cleave the LoxP site in the final haploid gametes.

In this report, the fetal Cre expression caused deletion, an interesting finding which to date has not been previously reported. In the fetal development stage, Cre driven by Foxp3 may function in some embryonic stem cells, resulting in the Dbc1 allele deletion. These stem cells differentiate into the germline, which resulted in the Dbc1 deleted offspring.

It is likely that the Cre expression in fetal cells caused the off-target recombination in the fetus of Foxp3^Cre mice. However, there is a small possibility that the source Cre is from the egg, but less likely from sperm. While the Cre expressed in the egg might not function in the egg itself, it may have functional effects in the early development of the fetus. Although the source of the Cre may indeed have been from the egg, the Cre expression does not occur during the diploid stage of gametogenesis, since all mice with recombined Dbc1 alleles contain Foxp3^Cre allele simultaneously. This study detected the Foxp3 expression in the ovary, but this does not mean that the Foxp3 expression is in mature eggs. The Foxp3 expression could not be identified in mature eggs, limited by the number of eggs, which might be resolved by single cell sequencing.

Even though this study found that Foxp3 drove fetal Cre function, the frequency of deleted Dbc1 was variable in the offspring. This means that different mice may contain different levels of the off-target Dbc1 recombination. Additionally, the off-target recombination possibly occurred in one or both of the two Dbc1 alleles in non-Treg cells, which results in the same type of non-Treg cells containing different levels of the Dbc1 protein. The variation of the off-target Dbc1 recombination implies that the fetal Cre expression is transient and the resultant gene recombination is random. The stochastic Foxp3 expression in the fetus shown by Gray et al. [31] and Diez-Roux et al. [32] also provides support.

When Rubtsov et al. [17] generated Foxp3^Cre mice, they used a ROSA26YFP (R26Y) recombination reporter allele in Foxp3^Cre mice to examine the specificity of Cre-mediated recombination. They observed that Foxp3^Cre mice exhibited varying degrees (2–10%) of R26Y recombination in different hematopoietic lineage cells. Further, Franckaert et al. [22] observed the deletion of CD28 on conventional CD4⁺ T cells in Foxp3^Cre Cd28^fl/fl mice. When considered in the context of our results, the off-target recombination in different hematopoietic lineage cells could indeed be as a result of recombination in the fetus.

In this report, it was found that the off-target deletion and many somatic cells rather than Treg cells may contain the deleted floxed gene in the Foxp3^Cre line. The off-target deletion can be genotyped using primers that flank the entire floxed locus of a conditional gene in combination with primers flanking a single LoxP site as suggested by Song et al. [9]. Furthermore, our results suggested that avoiding the use of male mice with a high frequency of recombined floxed allele to breed can limit the risk of having full knock out mice in the Foxp3^Cre line. However, the deleted gene allele can exist in either of the homologous chromatids of the somatic cells and this cannot be detected by ordinary PCR or qPCR. Our results showed that there is no breeding scheme to limit the risk of having mice with an off-target gene recombination. Therefore, to avoid errors, the mice with a low frequency of deleted gene fragments as detected in ordinary PCR or qPCR should be chosen to do the experiment.