Tamoxifen-independent Cre-activity in SMMHC-CreERT2 mice

Background and aims Recent technological advances have established vascular smooth muscle cells (SMCs) as central players in atherosclerosis. Increasingly complex genetic mouse models have unveiled that 30–70% of cells in experimentally induced atherosclerotic lesions derive from a handful of medial SMCs, and that these can adopt a broad range of plaque cell phenotypes. Most of these models are based on the SMMHC-CreERT2 mouse line as Cre-driver. Importantly, Cre-activation can be controlled in time (by administration of tamoxifen, TAM), which is critical to avoid unwanted effects of premature recombination events. The aim of this study was to scrutinize an unexpected observation of TAM-independent Cre-activity in this mouse line. Methods Cre-activity was assessed by PCR in tissues from SMMHC-CreERT2 mice crossed with mice homozygous for loxP-flanked (floxed) exon 4 of Ccn2 (our gene-of-interest), and Ccn2 protein was measured in aortas by targeted mass spectrometry. Results We observed spontaneous near-complete excision of floxed Ccn2 in aortas from adult mice that were not treated with TAM. As a result, Ccn2 protein was significantly reduced in aortas from these mice, but not to the same extent as TAM-treated littermates. Remarkably, most of the excision was completed in 4-week-old mice. Excision was Cre-dependent, as knockout bands were negligible in heart and liver (dominated by non-SMCs) of these mice, and undetectable in the aorta in the absence of Cre. Conclusion Our observations warrant caution, and we advocate inclusion of appropriate controls (i.e., TAM-untreated mice) in future studies.


Introduction
In recent years, the use of increasingly complex transgenic mouse models has provided unprecedented insight into the role of vascular smooth muscle cells (SMCs) in atherosclerosis. Mono-and multicolor lineage tracing studies have revealed that 30e70% of cells within advanced plaques derive from a handful of clonally expanded medial SMCs, and that SMC progeny can adopt a broad range of plaque cell phenotypes (recently reviewed [1e3]). In addition, genetic intervention studies based on SMC-specific gene deletion have elucidated SMC-dependent mechanisms that candidate future therapeutic targeting [4,5].
To enable temporal and SMC-specific activation of transgenes (e.g., a reporter fluorochrome) or inactivation of a gene-of-interest (GOI), several labs rely on the SMMHC-CreER T2 mouse line, developed in the Offermanns lab in collaboration with Pierre Chambon (GIE-CERBM) more than a decade ago [6] and today available from the Jackson Laboratories (B6.FVB-Tg(Myh11-cre/ERT2)1Soff/J, Stock No: 019079). In this mouse line, SMC-specific Cre recombinase (Cre) activity is ensured by placing Cre downstream the promoter of smooth muscle myosin heavy chain (SMMHC), which is active in fully differentiated contractile SMCs [1]. Moreover, Cre-activity is inducible since Cre is fused to mutated human estrogen receptor 1 (ER T2 ), which sequesters Cre from the nucleus until treating mice with tamoxifen (TAM, an exogenous ER agonist) (Fig. 1a). When entering the nucleus, Cre catalyzes site-specific recombination events between loxP sites. Each loxP site spans two 13 bp palindromic sequences separated by an 8 bp region, which orientation determines whether Cre will excise or invert the loxP-flanked (or "floxed") DNA segment [7,8].
Cross-breeding the SMMHC-CreER T2 Cre-driver line with mice harboring a floxed GOI enables inducible SMC-specific gene knockout by Cre-mediated excision (Fig. 1a, I). Alternatively, SMMHC-CreER T2 mice can be used for inducible SMC-specific activation of a transgene either by excision of a STOP cassette (i.e., a transcriptional roadblock) (Fig. 1a, II) or by inverting the transgene into the active orientation (Fig. 1a, III). These options (and combinations hereof) have opened a plethora of possibilities to investigate SMCs with high specificity and temporal control.
Herein, we report unexpected near-complete excision of a floxed GOI in SMMHC-CreER T2 mice that were not treated with TAM.
Upon initial characterization of knockout efficiency by PCR (allowing assessment of both wildtype (wt) Ccn2, intact (unexcised) flCcn2, and excised flCcn2 (878 bp, 1003 bp and 587 bp  Fig. 1). ntc ¼ no template control. c. Targeted mass spectrometry of Ccn2 (normalized to Gapdh) in aortas from the same experiment as in Fig. 1b. Data represent mean ± SEM. * ¼ p 0.05, * ¼ p 0.001 (Tukey's multiple comparisons test following one-way ANOVA). The peptides targeted in the assay and their position in Ccn2 and Gapdh, respectively, are shown to the right. d. Ccn2 locus genotyping of aorta, heart, and liver from TAM-untreated Ccn2 fl/fl /SMMHC-CreER T2 males at the ages indicated. e. Quantitative analysis of Ccn2 locus PCRs (same samples as in Fig. 1d) separated by capillary electrophoresis. The percentages of intact and excised flCcn2 (moles of each amplicon relative to moles of both amplicons), respectively, are shown for aorta, heart and liver. Data are shown as means. f. Ccn2 locus genotyping of tail biopsies from the time of unweaning (3 weeks of age), and tail biopsies and aortas from the time of sacrifice (10 weeks of age) from four mice. amplicons, respectively [9]), we noticed complete excision of flCcn2 in aortas of 14-week-old mice that had not been treated with TAM (Fig. 1b). TAM-independent flCcn2 excision led to a significant reduction in aorta Ccn2 at the protein level, but not to the extend observed in TAM-treated littermates (Fig. 1c).
To investigate TAM-independent flCcn2 excision further, we analyzed aortas from 1-to 10-week-old mice that had not been exposed to TAM. We observed progressive flCcn2 excision resulting in near-complete Ccn2 knockout in 10-week-old mice (Fig. 1dee). Genotyping of heart and liver samples from the same mice showed negligible flCcn2 excision (faint bands likely represent flCcn2 excision in the relatively small SMC population of these tissues) (Fig. 1d, two lower panels). Moreover, CreER T2 expression was required for flCcn2 excision, as no excision bands were observed in tail and aorta samples from a SMMHC-CreER T2 -negative Ccn2 fl/fl female littermate (the SMMHC-CreER T2 transgene is located on the Y-chromosome) (Fig. 1f).

Discussion
To our knowledge, this is the first report of spontaneous Creactivity ("Cre-leakage") in the widely used SMMHC-CreER T2 line, and the first observation of spontaneous near-complete excision of a floxed GOI in any CreER T2 line. Our observation highlights a potential limitation to the SMMHC-CreER T2 line, which can be problematic depending on experimental design and the research question at hand.
The ability to induce SMC-specific gene knockout in adult mice at the onset of disease is advantageous as it minimizes the risk of affecting disease-phenotype by effects preceding disease initiation. A GOI may have known or unanticipated roles during embryology or tissue maturation in adolescent mice that could affect subsequent disease development. Likewise, untimely activation of reporter transgenes in lineage tracing studies could influence experiment outcomes leading to inaccurate conclusions.
Whether our observation is a rare case, or a widespread problem is unknown, and whether potential premature Cre-leakage has affected previous observations based on this Cre-driver is difficult to evaluate, since TAM-untreated control mice are often omitted. Our measurements of Ccn2 protein in aortas showed that spontaneous flCcn2 excision significantly reduced Ccn2 expression, however, not to the same extend as in TAM-treated littermates (possibly due to residual intact flCcn2 in TAM-untreated mice during the sixweek period prior to sacrifice). Importantly, both TAM-treated and -untreated Ccn2 fl/fl /SMMHC-CreER T2 mice had lower aorta Ccn2 compared to wt mice. This observation highlights that evaluating Cre-leakage by comparing the level of GOI expression between TAM-treated and -untreated groups can be misleading. Optimally, Cre-leakage should be evaluated either by assessing transgene activation (e.g., checking if a reporter fluorochrome is detected in TAM-untreated mice) [4,10] or recombination events at the DNA level by PCR (as in this report).
The second generation CreER T2 system was developed to reduce leakage observed in the original CreER system [8], but leakage (less severe than reported here) has been reported for several CreER T2 lines and may be an inherent limitation of the system [11e14]. An important determinant of recombination susceptibility is the distance between loxP sites [15,16]. A recent study demonstrated that a distance of 0.9 kb between loxP sites (similar to that in flCcn2 mice [9]) in the Rosa26 locus lead to 60% TAM-independent recombination, while loxP sites distanced by > 2.45 kb were not recombined [17]. Also, the genomic position of loxP sites likely affects susceptibility to Cre recombination since open chromatin is more accessible to Cre than densely packed chromatin [15]. As such, genes critical for SMC differentiation/function, which tend to be positioned in regions of open chromatin are more prone to premature excision when floxed.
Although the mechanism by which CreER T2 translocates to the nucleus in the absence of TAM is elusive, our observation shows that nucleus sequestration is not 100% efficient in the SMMHC-CreER T2 line. We notice that Cre-activity accelerates in aortas of mice between the age of 2 and 4 weeks. Whether this represents increased activity of the SMMHC promoter at this timepoint due to maturation (i.e., final stages of differentiation) of SMCs at this time is speculative.
Going forward, we will use SMMHC-CreER T2 males without flCcn2 as experimental controls and acknowledge that progressing lack of Ccn2 starting from early adolescence may contribute to any vascular phenotype observed in Ccn2 fl/fl /SMMHC-CreER T2 adults.
In conclusion, we raise a potential concern to the use of the SMMHC-CreER T2 system. Although the severe leakage observed here may be due to a particularly susceptible flox-design, we advocate that fellow researchers perform and report appropriate control groups, i.e., include TAM-untreated mice -especially during initial characterization of novel mouse lines.
Data shown in Fig. 1dee are from all male offspring from seven litters (n ¼ 19) (mice were not exposed to TAM).
Data shown in Fig. 1f are from and experiment including three Ccn2 fl/fl /SMMHC-CreER T2 males and one Ccn2 fl/fl female littermate. Tail biopsies were harvested from 3-week-old mice (at the time of unweaning). At the age of 8 weeks, two males were treated with TAM, while one male and one female were untreated. All four mice were sacrificed at the age of 10 weeks.
Mice were euthanized by CO 2 /O 2 inhalation, and tissues (descending thoracic aorta, heart, and liver) were snap frozen in liquid nitrogen. Tissues were lysed using DirectPCR Lysis Reagent (250-102-T, Nordic Biosite) supplemented with 0.5 mg/ml proteinase K (28229, Nordic Biosite), and used for genotyping of the Ccn2 locus. PCR reactions (using either Taq DNA polymerase, Ampliqon, or Q5 High-Fidelity DNA Polymerase, New England Biolabs) were analyzed by agarose gel electrophoresis and capillary electrophoresis (QIAxcel Advanced System, Qiagen). Genotyping of the Ccn2 locus was performed using primers 5 0 -AATACCAATG-CACTTGCCTGGATGG and 5 0 -GAAACAGCAATTACTA-CAACGGGAGTGG. SMMHC-CreER T2 was genotyped using the standard protocol from Jackson Laboratories for this line.
Processing of snap frozen mouse aorta and quantitation of target proteins by multiple reaction monitoring (MRM) was performed as previously described [18]. Data analysis was based on peak area ratios between endogenous and heavy isotope labeled spiked peptides. Ccn2 (target peptide: TTTLPVEFK) was normalized to Gapdh (target peptides: GAAQNIIPASTGAAK and LISWYDNEYGYSNR).

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.