Characterization of a BAC transgenic mouse expressing Krt19-driven iCre recombinase in its digestive organs

Cytokeratin 19 (KRT19) protein is highly expressed in the epithelium of the gastrointestinal (GI) tract, hepatobiliary tissues, and pancreas of humans and mice. In the present study, we used an improved Cre (iCre) gene to enhance the efficiency of Cre expression in mammalian cells. We established a new transgenic Krt19-iCre bacterial artificial chromosome (BAC) mouse model using the BAC recombineering strategy. Site-specific iCre expression pattern was examined in embryos, adults, and elderly Krt19-iCre mice crossed with Tomato or LacZ reporter mice. Both iCre and reporter protein expressions in adult Krt19-iCre;Tomatoflox/+ (Krt19-iCre Tomato reporter) mice were observed mainly in the epithelial cells of the GI tract, hepatobiliary tissues, and pancreas. However, the expression in the intrahepatic and small pancreatic duct were lower than those in the common bile and large pancreatic duct. In the Krt19-iCre; LacZ reporter embryos, β-galactosidase for the LacZ reporter was expressed in the glandular epithelial cells of the GI tract in 9.5-day embryos, 12-day embryos, and newborn mice. The reporter protein expression in Krt19-iCre-Tomato reporter mice was consistent with the KRT19 expression in human GI tissues. In conclusion, Krt19-iCre BAC transgenic mice can be used to investigate developmental and pathological conditions using the iCre-loxP system.


Introduction
KRT19, an epithelial-specific typical type I keratin (~40 kDa) protein, is found in a broad range of epithelial tissues [1]. In mice, it is expressed in the epithelium of the stomach, small intestine, colon, pancreatic ducts of the adult pancreas, and the hepatobiliary ducts [2,3]. In humans, KRT19 immunohistochemistry (IHC) is frequently used in diagnostic pathology to a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 confirm the epithelial immunophenotype of undifferentiated tumors or to identify whether tumors of unknown primary origin are biliary, pancreatic, or renal ductular in origin [4,5].
Cre/loxP-mediated DNA recombination allows for the functional analyses of essential genes in specific tissues. Thus, transgenic mice expressing Cre DNA recombinase in tissuerestricted patterns serve as practical tools to examine developmental and pathological conditions. A Krt19-Cre knock-in mouse line has been used to demonstrate that the mutated β-catenin gene induces intestinal polyposis [6]. Another Krt19-Cre mouse, in which Cre recombinase was inserted in the Krt19-promoter region, was crossed with a transgenic mouse strain with PTEN conditional alleles and PTEN protein expression was abrogated in the stomach and intestine [7]. Furthermore, Krt19-CreERT mice have demonstrated conditional DNA recombination in epithelial cells of multiple endodermal organs [8]. These reports demonstrate that Cre-mediated DNA recombination in Krt19-positive tissues is useful for studying normal developmental and pathological conditions. However, in mice, it is still unclear whether Krt19 is specifically expressed in the epithelium of the GI tract, hepatobiliary tissues, and pancreas; the human equivalents of which have been shown to be sites of KRT19 expression. Previous reports [6,8] using Krt19-Cre mice have predominantly focused only on the stomach and parts of the intestine; therefore, a comprehensive comparison of Krt19 expression between mice and humans is warranted using Cre-mediated DNA recombinant mouse models.
The Cre gene comprises an undesirably high frequency of CpG dinucleotides, which can induce epigenetic silencing during mammalian development. Shimshek et al. [9] designed an improved Cre (iCre) gene, which is more efficiently expressed than the traditional one. Thus, several useful iCre transgenic mice lines have been recently generated [10][11][12][13].
In this study, we established transgenic Krt19-iCre mice using the bacterial artificial chromosome (BAC) recombineering strategy. On crossing with reporter mice, excellent iCre expression was observed in the progeny mice. This expression mimicked the KRT19 expression in the GI tract, hepatobiliary tissues, and pancreas of humans.

Ethics statement
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the Gifu University. The protocol was approved by the Committee on the Ethics of Animal Experiments of Gifu University (30-21). All surgery was performed under isoflurane anesthesia, and all efforts were made to minimize suffering.
This use of human tissues in this study was approved by the Institutional Review Board of Gifu University and all human samples were obtained from Gifu University Hospital. Full written informed consent was obtained for all samples used in this study.
Twenty pups were produced, of which 12 contained one copy of the Krt19-iCre vector in their genome. The BAC DNA is randomly integrated into the mouse genome and can destroy the expression of other inherent mouse genes. However, the BAC transgene integrity has not always been elucidated in published studies [21]. Therefore, we used only heterogeneous mice, carefully checked for the endogenous Krt19 expression, observed the health condition at least for one year, and analyzed the macroscopic and microscopic findings of the founders.
Further, five founders had little or no LacZ expression when crossed with Rosa26-LSL-LacZ reporter mice. Three mice had no pregnancies and two mice had a low body weight, even as adults. Thus, we euthanized the 10 founders humanely.
Two of these 12 Krt19-iCre positive mice, Krt19-iCre#2 and Krt19-iCre#5, showed no health issues or major fertility defects and were chosen for the follow-up characterization of iCre expression. Thus, we used the Krt19-iCre#2 mouse line, with heterogeneous but no homogeneous alleles, in this study. All the experiments in this study were examined using two embryos, six adults, and two elderly mice and the expression patterns were the same in each mouse line. All mice used in this study remained healthy throughout the study.

Human tissue samples
Twelve human samples, which included normal tissues, were collected from nine patients who underwent surgical resection or biopsy at Gifu University Hospital. In each sample, normal tissue surrounding the lesions was identified by two certificated pathologists (T.K. and H.T.). More detailed case information is shown in S1 Table.

Immunohistochemistry and immunofluorescence
The mice were perfused with 10% neutral buffered formalin and dissected. The adult and embryo tissues were fixed in 10% neutral buffered formalin overnight. The tissues were dehydrated with 70% ethanol and embedded in paraffin blocks. These blocks were cut into 3-μm thick sections. For IHC, the slides were deparaffinized in xylene and rehydrated using a graded alcohol series. The slides were heated in 10 mM citrate buffer (pH 6.0) at 120˚C for 1 min for antigen retrieval in a Pascal pressure chamber (DAKO, Glostrup, Denmark). Following immersion in 0.3% H 2 O 2 in methanol to quench endogenous peroxidase activity, the slides were blocked with 2% bovine serum albumin for 40 minutes. Each section was incubated with a primary antibody(described below) at 4˚C overnight. The slides were then incubated with the secondary antibody (Vectastain Elite ABC Kit; Vector Laboratories, Burlingame, CA, USA) at room temperature for 30 min and subsequently treated with ABC reagent (Vectastain Elite ABC Kit) at room temperature for 30 minutes. The reaction products were visualized using ImmPACT DAB (Vector Laboratories). The slides were counterstained with hematoxylin.
To investigate the expression of human endogenous KRT19,anti-KRT19 antibody (rabbit monoclonal IgG, 1:800, Abcam, Cambridge, MA, USA) was used as a primary antibody. To determine the expression level of β-gal and iCre, anti-β-galactosidase (chicken polyclonal IgG, 1:1000, Abcam) antibody and anti-Cre antibody (1:200, Cell Signaling Technology, Danvers, MA, USA) were used as primary antibodies, respectively. Rabbit or chicken IgG were used as negative controls.
For immunofluorescence (IF), rabbit anti-RFP primary antibody (1:100, ROCK LAND, Gilbertsville, PA, USA) and anti-Krt19 antibody were used in a method similar to IHC. Rabbit IgG antibody was used for the negative control (1:100, DAKO, Japan). The slides were incubated with a secondary antibody (Goat polyclonal Secondary Antibody to Rabbit IgG, Alexa Fluor 594, pre-adsorbed, 1:200, Abcam and goat polyclonal Secondary Antibody to chicken IgG, Alexa Fluor 594, pre-adsorbed, 1:200, Abcam) at room temperature for 30 minutes. Then, the slides were incubated with DAPI solution (1:1000, WAKO, Japan) under light-shielded conditions at room temperature for 5 minutes. Finally, the slides were covered with cover slips.

In situ hybridization and evaluation
We used RNAscope (Advanced Cell Diagnostics, Inc., California, USA) (Wanget al., 2012), an RNA in situ hybridization (ISH) method that can detect RNA in paraffin-embedded tissues. Three-μm section slides were deparaffinized in xylene and rehydrated using alcohols, then air-dried. The slides were incubated with an iCre probe (RNAscope Probe-iCre, #423321) or negative control probe (RNAscope Probe-negative control, #310043), and the signal was amplified using the RNAscope 2.0 HD Detection kit (Advanced Cell Diagnosis, Inc.) according to the manufacturer's instructions and a previous report [24]. The expression was scored on a five-grade scale, as described in the manufacturer's protocol (https://acdbio.com/services/ quantitative-analysis). Briefly, the dots per five high-power fields were counted, the average number was calculated, and the score was defined as follows: 0, no staining or less than 1 dot/10 cells; 1+, 1-3 dots/cell; 2+, 4-9 dot/cell and very few dot clusters; 3+, 10-15 dots/cell and the clusters were less than 10% dots; 4+, >15 dots/cell and the clusters were more than 10% dots.
To confirm the co-localization of iCre mRNA expression and Krt19 protein expression, we performed the combined ISH and IHC technique according to a previous report [25] Quantitative RT-PCR. RNA was extracted from P56 Krt19-iCre mice and wild-type mice using an RNeasy Mini kit (Qiagen, Tokyo, Japan). mRNA (500 ng) was reverse transcribed with SuperScript III (Invitrogen) at 55˚C for 1 hour into total complementary DNA (cDNA), which was used as the template for subsequent PCR reactions and analysis. The expression of iCre RNA was examined by real-time polymerase chain reaction (RT-PCR) using SYBR Green Detection Chemistry (Applied Biosystem). cDNA (500 ng) was used as the template for qPCR amplification in the presence of 10 μM specific primers, in a total volume of 20 μl. All reactions were run in triplicate and each reaction was also run without cDNA as negative controls. Ste-pOne software v2.3 (Applied Biosystem) was used to convert the fluorescent data into cycle threshold (delta-Ct) measurements and the relative amount of specific transcript was calculated by the comparative cycle threshold method. All values of RNA accumulation of the specific genes were normalized to the signal of β-actin.

X-gal staining
To detect the localization of iCre expression in mouse embryos, we used Krt19-iCre; lacZ transgenic mouse embryos to detect β-gal activity by 5-bromo-4-chloro-3-indoxyl-beta-β-Dgalactopyranoside (X-gal) staining. Briefly, E9.5 embryos were dissected away from the uterus and fixed 1 hour in 0.2% glutaraldehyde. After the embryos were washed 3 times with buffer for 15min, they were stained with X-gal (Roche, Indianapolis, IN, USA) for six hours under light-shielded conditions. Whole-mount-stained embryos were fixed in 10% neutral buffered formalin overnight. The tissues were dehydrated with 70% ethanol and embedded in paraffin blocks. These blocks were cut into 3-μm thick sections. After counterstaining using nuclear fast red stain, the slides were covered with cover slips.

Results and discussion
iCre mRNA expression in the GI tract of Krt19-iCre mice To investigate mRNA expression in the GI tract, hepatobiliary tissues, and the pancreas of Krt19-iCre mice, we performed in situ hybridization (ISH) using the RNAscope assay for three adult mice. iCre mRNA expression was detected in the epithelial cells of the entire GI tract, i.e., the esophagus, fundus and pylorus of the stomach, duodenum, ileum, jejunum, and colon (Fig  2a-2e). iCre mRNA expression was also observed in the epithelial cells of the biliary tree, including the common bile duct, intrahepatic bile duct, and pancreatic duct (Fig 2f-2h). These expression patterns were consistent among each tissue from the three mice. The expression of iCre mRNA was also confirmed by quantitative RT-PCR. While tissues from the gastrointestinal tract, hepatobiliary tract, and pancreas of the transgenic mice showed amplification of iCre mRNA, tissues from wild-type mice did not express iCre mRNA(S1 Fig). The expression patterns of iCre mRNA were consistent with the expression of endogenous Krt19 protein in the digestive organs, liver, and pancreas. However, in the small intestine, iCre mRNA expression in the three samples was found mainly in transient amplifying cells and it was restricted compared to the endogenous Krt19 expression (S2 Fig). Furthermore, some hepatocytes in the liver and the acinar exocrine cells in the pancreas showed iCre mRNA expression (Fig 2f and  2h). Scoring of the epithelial expression levels revealed that iCre mRNA expression in the intrahepatic bile duct was lower than that in the GI tract, the common bile duct, and the pancreatic duct (Fig 2i).  f and g, respectively), and epithelial cells of the pancreatic duct (arrowheads) and in some exocrine cells(arrows) (h). Scale bars: 50μm. (i) Scores for the epithelial cells of the GI tract, hepatobiliary tissues, and the pancreas. The expression in the epithelial cells was scored on a five-grade scale as described in the manufacturer's protocol (https://acdbio.com/services/quantitative-analysis) and was read as follows: 0, no staining or less than 1 dot/10 cells; 1+, 1-3 dots/cell; 2+, 4-9 dot/cell and very few dot clusters; 3+, 10-15 dots/cell and the clusters were less than 10% dots; 4+, >15 dots/cell and the clusters were more than 10% dots. These data are consistent with those of a previous study [8] which demonstrated that in inducible Krt19-CreERT knock-in mice, the EYFP reporter protein expression level was lower in cells of the intercalated ducts, such as the intrahepatic and small pancreatic ducts, compared to that in cuboidal epithelial cells, such as the common bile duct and main pancreatic duct cells. Unfortunately, even though iCre recombinase was used instead of Cre recombinase, we were unable to improve low reporter protein expression levels in the intercalated ducts of the liver and pancreas.
In addition, there were signals in connective tissues of the esophagus, and these signals were specific because no hybridization signal was detected in other GI tract tissues using a negative control probe (Fig 2a and S2 Fig). Reporter expression in the GI tract of Krt19-iCre mice. To determine whether the reporter protein expression correlated with iCre mRNA expression in the GI tract, hepatobiliary tissues, and the pancreas, IF staining was performed with RFP antibody in tissue sections from the Krt19-iCre;Tomato flox/+ (Krt19-iCre Tomato reporter) mice (n = 3). Of all the samples, those of the GI tract, hepatobiliary tissues, and pancreas had RFP expression (Fig 3a-3h) corresponding to iCre mRNA expression determined by ISH (Fig 2). Further, biliary tree duct cells showed lower RFP expression levels than the epithelial cells of the GI tract (Fig 3a-3h), which is generally consistent with iCre mRNA expression in these cells (Fig 2). Esophageal epithelia and hepatocytes were labeled with anti-RFP antibody, however, these signals were also found in the negative control slides using rabbit IgG (Fig 3a and 3f, and S4 Fig). Moreover, some RFP-positive cells in the esophageal connective tissue were also labeled in the negative control. Although the unexpected mRNA expression of iCre among connective tissue cells of the esophagus was detected (Fig 2), the protein expression of iCre was uncommon. We confirmed the co-localization of RFP and endogenous Krt19 protein in the Krt19-iCre Tomato reporter mice (S5 Fig). To verify that iCre expression pattern corresponded to endogenous Krt19 expression, we performed immunohistochemistry for Krt19 and iCre (using anti-Cre antibody) on tissues from the GI tract, hepatobiliary organs, and pancreas using serial sections from Krt19-iCre-Tomato reporter mice (n = 3). The KRT19 and iCre protein expression patterns were consistent with each other in the epithelial cells of the stomach and colon (Fig 4b, 4c and 4e). Similarly, expression of endogenous KRT19 protein corresponded to that of iCre protein in the hepatobiliary and pancreatic ducts, as well as the stomach and colon (Fig 4f-4h). In the small intestine, iCre protein expression in the three samples was found mainly in transient amplifying cells and it was restricted compared to the endogenous Krt19 expression (Fig 4d). Though iCre expression was somewhat inconsistently correlated with endogenous Krt19 expression, reporter RFP protein expression patterns were observed over the entire intestinal epithelium, including the epithelial cells of the villi (Fig 3d and S5 Fig). Thus, we consider the Krt19-iCre mouse line to be a useful model to investigate GI tract epithelium, including the small intestine.
iCre expression in developmental stages of Krt19-iCre mice. In order to characterize Krt19-iCre reporter expression during development, embryos resulting from Krt19-iCre×homozygous LacZ flox/flox mating were analyzed at different developmental stages. Embryos at E9.5, E12, and E17 stages, and newborn pups (P0) were obtained. The whole-mount X-gal staining of E9.5 embryos showed β-gal activity in the epithelium of the foregut and part of the skin (Fig 5a-5d). The β-gal activity in Krt19-iCre; lacZ embryos (E9.5) was equivalent to that of endogenous Krt19 previously reported in Krt19 knock-in LacZ reporter mice [26]. In contrast, a previous report demonstrated that some E9.5 embryos of Krt19-Cre; lacZ mice expressed β-gal activity in the whole body, a wider tissue range than the expression of endogenous Krt19 [27]. The results indicate that Krt19-iCremicemight be a better model than the previous Krt19-Cre strain for studying early fetal development.
An anti-β-galactosidase antibody for IHC was used to measure the expression level of the β-gal reporter at the cellular level. At E12, β-galactosidase was expressed in the glandular epithelial cells of the GI tract (Fig 6a). This expression was also shown at E17 and P0 (Fig 6b  and 6d). Unlike in adult mice, iCre reporter expression was undetectable in the fetal liver at E17 (Fig 6c). At P0, the expression pattern of β-galactosidase in cholangiocytes was similar to that of Krt19 in the intrahepatic duct of adult reporter mice (Fig 6e-6g). The expression of βgalactosidase was also detected in the epithelial cells of the biliary tree and in some epithelial cells of the exocrine glands in the pancreas in P0 mice (Fig 6h). The expression patterns of βgalactosidase at various stages of fetal development indicate that the Krt19-iCre mouse model is capable of tracing endogenous Krt19 expression in the GI tract, hepatobiliary tissues, and the pancreas during mouse development.
We also confirmed iCre expression in aged Krt19-iCre mice by IHC. The expression level of iCre protein was maintained in the stomach, small intestine, and colon. However, iCre expression in the hepatobiliary duct and pancreatic duct of Krt19-iCre mice at 18 months of age was weak compared to that in mice at eight weeks of age (S6 Fig). KRT19 expression in the GI tract of human tissue. To evaluate KRT19 expression in human tissues, we performed IHC with an anti-KRT19 antibody in normal tissue sections from the GI tract, hepatobiliary tissues, and the pancreas. KRT19 was expressed in the epithelial cells of the GI tract and the biliary tree, i.e., oral mucosa, esophagus (epithelial cells and glands), fundus and pylorus of the stomach, duodenum, ileum, jejunum, and colon (Fig 7). These results demonstrate that the expression of KRT19 in humans closely agreed with that in Krt19-iCre Tomato reporter mice (Fig 3). In human tissues, epithelial cells of the biliary tree had similar KRT19 expression levels as those of the GI tract (Fig 7c-7l). However, in Krt19-iCre Tomato reporter mice, KRT19 expression levels were lower in the epithelial cells of Generations of Krt19 iCre mouse the biliary tree than in the GI tract (Fig 3g and 3h). In a previous report, Krt19-CreERT mice showed similar Krt19 reporter expression patterns in the biliary tree [8]. Taken together, the lower reporter expression in the epithelial cells of the biliary tree may represent mouse Krt19 promoter-dependence rather than an actual difference between Cre and iCre-dependent expressions.

Conclusion
The expression of iCre recombinase from the Krt19 locus in adult mice was apparent in the epithelial cells in the GI tract, intrahepatic and common bile duct, and the pancreatic duct.
We suggest that the Krt19-iCre mouse line we established might be a better model than the current models (Krt19-Cre and Krt19-CreERT lines) for the following two reasons: First, the βgalactosidase activity in the Krt19-iCre; lacZ embryos (E9.5) was equivalent to that of endogenous Krt19previously reported in Krt19 knock-in LacZ reporter mice. Second, this is the first model to demonstrate the expression patterns of Cre recombinase using β-gal at various fetal and neonatal stages. The Cre expression patterns during later fetal and neonatal stages in the current mouse line models for Krt19, Krt19-Cre and Krt19-CreERT were not analyzed. The Krt19-iCre line we established is more suitable for use in Cre-mediated DNA recombination mouse models for embryological study.
Together, we conclude that the generated Krt19-iCre transgenic mouse line is a valid and useful tool for studying most stages, including fetal and neonatal, of epithelial cell development. In the comparison with human samples in this study, the expression of iCre in this transgenic mouse line generally agreed with the expression pattern of KRT19 in human organs. Thus, genomic editing using the iCre-loxP system, more efficient than the traditional Cre-loxP system, may enable the analysis of pathological conditions.