Temporal Dissection of K-rasG12D Mutant In Vitro and In Vivo Using a Regulatable K-rasG12D Mouse Allele

Animal models which allow the temporal regulation of gene activities are valuable for dissecting gene function in tumorigenesis. Here we have constructed a conditional inducible estrogen receptor-K-rasG12D (ER-K-rasG12D) knock-in mice allele that allows us to temporally switch on or off the activity of K-ras oncogenic mutant through tamoxifen administration. In vitro studies using mice embryonic fibroblast (MEF) showed that a dose of tamoxifen at 0.05 µM works optimally for activation of ER-K-rasG12D independent of the gender status. Furthermore, tamoxifen-inducible activation of K-rasG12D promotes cell proliferation, anchor-independent growth, transformation as well as invasion, potentially via activation of downstream MAPK pathway and cell cycle progression. Continuous activation of K-rasG12D in vivo by tamoxifen treatment is sufficient to drive the neoplastic transformation of normal lung epithelial cells in mice. Tamoxifen withdrawal after the tumor formation results in apoptosis and tumor regression in mouse lungs. Taken together, these data have convincingly demonstrated that K-ras mutant is essential for neoplastic transformation and this animal model may provide an ideal platform for further detailed characterization of the role of K-ras oncogenic mutant during different stages of lung tumorigenesis.


Introduction
Animal models which allows temporal regulation of gene activity have been proved very valuable for gaining insights into the gene function involved in different tumor formation stages. Previous studies have demonstrated that knock-in of ER region into certain genes including P53 and Myc to generate fusion proteins allow for precise and temporal regulation of gene activity via ectopic provision of 4-hydroxytamoxifen [1,2]. The Trp53ER KI/KI mouse model is the first example of genetic models that allows specific, rapid and reversible perturbation of the biological function of a single endogenous gene in vivo. Studies based on this mouse model have dissected the precise role of P53 engaged in different stages of tumorigenesis as well as tumor regression [2,3,4,5,6,7,8]. Similarly, studies using the MycER mouse allelesignificantly contribute to our understanding about the role of Myc in cell proliferation, differentiation as well as the relationship between stem cells and cancer [1,6,7,9,10,11,12].
Oncogenic mutations including K-RAS G12D is frequently observed in approximately 20% of all types of human cancers including carcinomas of the lung, colon, and pancreas [13,14,15,16,17,18]. K-RAS is a small GTPase which works as a binary molecular switch between a GDP-bound inactive form and a GTP-bound active form. When mutated in codon 12, 13, or 61, the K-RAS mutants remain active and constitutively transduce signals through MAPK pathway and PI3K pathway [19,20,21,22].
Previous studies have demonstrated that proteins such Ras or Raf fused to ER could be activated by tamoxifen treatment [23,24,25,26,27]. However, there is no animal model has been established based on this knowledge yet.
We have here engineered a knock-in allele encoding a 4hydroxy tamoxifen (4-OHT) inducible estrogen receptor-K-ras G12D (ER-K-ras G12D ) with placing the Loxp-Stop-Loxp fragment in front of the exon 1 of K-ras coding region. Using the mouse embryonic fibroblast (MEF), we have shown that ER-K-ras G12D expression is induced by Adeno-Cre treatment and the activity of ER-Kras G12D mutant is regulated by an optimal dose of tamoxifen administration. We further demonstrated that the ER-K-ras G12D mutant is essential for neoplastic transformation as well as tumor maintenance.

Mouse cohorts and Treatment
The K-ras G12D and p53 L/L mice were originally generously provided by T. Jacks (Cambridge, MA) and R. Depinho (Boston, MA), respectively. The LSL-ER-K-ras G12D mice allele was constructed by placing the estrogen receptor cDNA in front of the Kras G12D coding region as shown in Figure 1. The targeting vector carried a negative selection marker for diptheria toxin (DT), a positive selection marker for neomycin acetyltransferase (Neo) and loxP sites (black triangles). The restriction sites were BamH I (B); Kpn I (K); Not I (N); Xba I (Xb); Sph I (S). We electroporated embryonic stem cells and selected transformed cells by 39 arm and 59 arm PCR screening to identify 3 recombinants. Blastocyst injections were carried out with these different targeted clones and germline transmission was achieved. The LSL-ER-K-ras G12D mice were then crossed to p53 L/L mice to obtain LSL-ER-K-ras G12D , p53L/L mice. All mice were housed in a specific pathogen-free environment at Shanghai Institute of Biochemistry and Cell Biology and treated in strict accordance with protocols approved by the Institutional Animal Care and Use Committee of the Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences. These mice were treated with Adeno-Cre through nasal inhalation as described before [28]. After three weeks of nasal inhalation, we treated these mice with 4-hydroxytamoxifen with different doses daily via intraperitoneal injection. Mice were then sacrificed for pathological inspection. Some mice were stopped for tamoxifen treatment for one or two weeks and then sacrificed for pathological inspection. Genotyping primers for p53L/L, LSL-K-ras G12D and LSL-ER-K-ras G12D was listed as following: For p53L/L, forward primer: 59-CACAAAAACAGGT-TAAACCCAG-39; reverse primer: 59-AGCACATAGGAGGCA-GAGAC-39. For LSL-K-ras G12D and LSL-ER-K-ras G12D , same primers were used for genotyping. Forward primer: 59-CTAGC-CACCATGGCTTGAGT-39; reverse primer:59-TCCGAATT-CAGTGACTACAGATG-39.

Cell culture and cell proliferation assay
We generated MEFs from 13.5 post-coitum embryos and grew them in DMEM medium plus 10% fetal bovine serum (Biochrom, AG), penicillin and streptomycin. MEFs (,4610 5 ) at 6-well plate were virally infected with Adeno-Cre (4610 6 CFU) overnight and then changed with fresh medium. The MEFs were then cultured for at least two more passages to get enough cells for various functional assays. Cell number were counted after each passage with or without tamoxifen treatment. Cells were also taken for PI staining and flow cytometry analysis for cell cycle.

Cellular functional assay
MEFs with indicated genotype with or without tamoxifen (Sigma, Cat: 011M1682V) treatment were cultured in DMEM with 10% fetal bovine serum until 100% confluence. MEFs were changed with fresh medium every three days for 3 weeks before staining with 0.005% crystal violet. Foci formation were observed and the number of foci were counted.
As for anchorage-independent cell growth, MEFs were plated in 0.4% top agarose on top of a 1% agarose base supplemented with complete medium. MEFs were changed with fresh medium with or without tamoxifen every 3 days for about 4 weeks before staining with 0.005% crystal violet. The number of colonies (.0.5 mm) were counted under the microscope.
As for invasion assay, MEFs were cultured in matrigel (BD, Cat: 356234) in the presence or absence of tamoxifen treatment for 10 days. Photos with typical morphologic change were shown.

Histopathological Analysis and Immunological Studies
Histopathological analysis were performed as described before [30]. Briefly, mice were sacrificed and lung tissues were inflated and fixed in 10% formalin, embedded in paraffin and sectioned for hematoxylin and eosin (HE) staining. For immunostaining, sections of 5 mm-thickness were cut from paraffin-embedded lung tissues, deparaffinized in xylene and rehydrated in ethanol. Deparaffinized sections were pretreated with 0.3% H 2 O 2 (in absolute methanol) for 15 min at room temperature to block endogenous peroxidase activity. The slides were then heated in a microwave for three periods of 5 min high power, 2 min defrost and 20 min medium low power of microwave treatments. After recooling at room temperature for about 30 min, slides were rinsed twice in H 2 O for 3 min and then were treated for 2 h with normal non-immune goat serum albumin to block nonspecific staining. Sections were incubated overnight with Ki67 antibody (1:100, Santa Cruz, SC-23900) and Cleaved Caspase-3 antibody (1:300, Cell Signaling, Cat: 9661). The secondary biotinylated antibody was incubated for 15 min at 37uC. The slides were then stained using a diaminobenzidine tetrahydrochloride (DAB) detection kit (SP-9001), counterstained with haematoxylin and mounted with permount.

Results
Establishment of the knock-in mice with an endogenous conditional regulatable LSL-ER-K-ras G12D allele We have made the conditional regulatable LSL-ER-K-ras G12D knock-in allele using the homologous recombination with the replacement of endogenous K-ras with a fragment containing of a 1.0 kb estrogen receptor cDNA fused to the K-ras G12D (Fig. 1A). The mice were then crossed with b-actin-Flpe to delete the Neo cassette and the resultant mouse allele was named as LSL-ER-Kras G12D . The homologous recombination were confirmed by the 59arm and 39arm PCR as described before (Fig. 1B) [31]. We obtained the LSL-ER-K-ras G12D mouse embryonic fibroblasts (MEFs) and infected them with Adeno-Cre. As expected we have detected the expression of ER-K-ras G12D at mRNA level (Fig. 1C), which was further confirmed by direct sequencing (Fig. 1D).

Activation of ER-K-rasG12D and downstream signaling by tamoxifen treatment
We next determined if the ER-K-ras G12D MEFs are responsive to tamoxifen treatment. Tamoxifen is an anti-breast cancer agent and potentially toxic. We therefore titrated its dosage for the ER-K-ras G12D MEF treatment. Interestingly, we found that the proliferation rate of MEFs increased with increased tamoxifen dose from 0.0125 mM to 0.05 mM ( Figure S2A-B). However, decreased cell proliferation was observed when an even higher dose of tamoxifen (from 0.1 mM to 0.2 mM) was given to MEFs ( Figure S2A-B). Similar finding was observed in the ER-K-ras G12D MEFs with p53 deficient background ( Fig. 2A and Figure S2C). Furthermore, both ER-K-ras G12D and p532/2, ER-K-ras G12D MEFs (PEK-C) from either male or female has shown consistent response to tamoxifen treatment ( Fig. 2A, Figure S1, Figure S2), suggesting a gender-independent pattern. Hereafter, we chose a dose of tamoxifen at 0.05 mM for further in vitro studies. P53 loss is commonly observed in human cancer [32]. Moreover, p53 deficiency results in the immortalization of MEFs and also significantly contributes to colony formation of K-ras G12D MEFs in soft agar. For these reasons, we have therefore used the p532/ 2, ER-K-ras G12D MEFs (refer to PEK-C) for our further studies. We compared the mRNA level of total K-ras in p53L/L, LSL-ER-K-ras G12D (PEK), p532/2, ER-K-ras G12D (PEK-C) and p532/2, K-ras G12D (PK-C) MEFs (Fig. 2B) as well as LSL-K-ras G12D , LSL-ER-K-ras G12D , ER-K-ras G12D MEFs with or without Adeno-Cre treatment ( Figure S3A-C). No significant change of total K-ras mRNA and protein levels was observed among all MEFs with or without either Adeno-Cre or tamoxifen treatment (Fig. 2C-D). Using K-ras antibody, we have detected the ER-K-ras band with correct molecular weight (Fig. 2D). This was further confirmed by western blot using ER antibody ( Figure S3D). Consistently, we have observed the phosphorylation of several Ras downstream effectors including Raf, Erk, Akt and Gsk after tamoxifen treatment (Fig. 2E).

Inducible activation of ER-K-ras G12D promotes cell proliferation
Abnormal cell proliferation is considered as an early event during tumorigenesis [33]. Tamoxifen administration significantly increased the cell division in MEFs indicated by an increase of cells at G2/M phase (Fig. 3A). Increase of cell cycle related proteins including Cyclin d1, Cdk4, Cdk6, p15, Cyclin b1, and Cdc2 were also observed (Fig. 3B). Conversely, p27 protein level decreased after tamoxifen treatment (Fig. 3B). We further checked the cell proliferation rate of PEK, PK-C, and PEK-C with or without tamoxifen treatment using cell number counting for 6 passages. We found that the tamoxifen treatment significantly increased cell proliferation in PEK-C MEFs. Actually, the growth rate of PEK-C MEFs treated with 0.05 mM tamoxifen is comparable to PK-C MEFs. Taken together, these data have demonstrated that ER-Kras activation by tamoxifen treatment promoted cell proliferation in vitro.

Activation of ER-K-ras G12D is essential for cell transformation and cell invasiveness
We then determined if the inducible ER-K-ras activation promotes cell transformation, anchorage-independent cell growth and cell invasiveness in 3D cell culture. As expected, the PEK MEFs didn't form any foci. Tamoxifen treatment of PEK-C MEFs obviously promoted the foci formation, comparable to PK-C MEFs (Fig. 4A). The anchor-independent growth of PEK-C MEFs was also enhanced by activation of ER-K-ras G12D after tamoxifen treatment (Fig. 4B). These data have demonstrated that the ER-Kras G12D activation is essential for cell transformation. Consistently, we found that tamoxifen administration significantly increased cell invasion in matrigel, suggesting that ER-K-ras G12D activation by tamoxifen treatment may promote cell invasion (Fig. 4C).

Continuous activation of ER-K-ras G12D induced lung hyperplasia
RAS mutant play a pivotal role in neoplastic transformation and tumorigenesis [13,19,34,35,36]. To test the in vivo function of ER-K-ras in lung tumorigenesis, we first tried tamoxifen treatment at different doses in mice. All the p53L/L, LSL-ER-K-ras G12D mice were given Adeno-Cre at 6,8 week-old. After three weeks of Adeno-Cre treatment, these mice were then given tamoxifen daily via intraperitoneal (IP) injection. Our pilot experiment has used a series of different tamoxifen doses at 0, 100, 250, 500, 1000 mg per mouse. All the mice without tamoxifen treatment showed normal lung histology (Fig. 5B and Figure S4). Mice with low doses of tamoxifen treatment at 100 or 250 mg displayed normal lung pathology, similar to those with a higher dose of tamoxifen at 1000 mg ( Figure S4). Interestingly, mice receiving tamoxifen at 500 mg were found to have lung hyperplasia as early as after 5week treatment ( Figure S4). Lung adenomas, positive for Ki-67 staining, arose at 8 weeks post tamoxifen treatment (Fig. 5A). These data suggested that an optimal dose of tamoxifen is sufficient to drive lung epithelial cell proliferation, hyperplasia and early tumor formation.

Continuous activation of ER-K-ras G12D is important for tumor maintenance
To further test the potential role of K-ras mutant in tumor maintenance, we have performed both in vitro and in vivo studies. As we have shown above, tamoxifen treatment significantly promoted PEK-C MEFs proliferation after several passages (Fig. 3C). However, after tamoxifen withdrawal, these MEFs obviously reduced the proliferation rate in comparison with PK-C MEFs (Fig. 6A). Similarly, tamoxifen withdrawal resulted in a dramatic decrease of the colony formation ability and cell invasiveness of these MEFs (Fig. 6B-C).
Our data have shown that continuous tamoxifen treatment at an optimal dose for 12 weeks consistently drove lung hyperplasia formation in p53L/L, LSL-ER-K-ras G12D after Adeno-Cre treatment (Fig. 5). We then stopped tamoxifen treatment afterwards for either 1 week or 2 week. Pathological inspection have shown that the lung hyperplasia significantly decreased in mouse lungs even after one week of tamoxifen withdrawal (Fig. 6D). The tumor regression is accompanied by apoptosis indicated by increased cleaved caspase-3 immunostaining ( Figure 6D). The mouse lungs were largely normal after two weeks of tamoxifen withdrawal (Fig. 6D), suggesting that continuous activation of K-ras mutant is important for tumor maintenance.

Discussion
Temporal controlling K-ras activity in vitro and in vivo is of great importance for dissection of its function in tumorigenesis. We have here constructed a conditional inducible oestrogen receptor-Kras G12D (ER-K-ras G12D ) knock-in mice allele that allows us to temporally switch on or off the K-ras oncogenic mutant through the tamoxifen administration. This allele is able to be regulated by tamoxifen treatment at an optimal dosage in vitro and in vivo. Our data have demonstrated that the deregulation of K-ras activity plays an important role in neoplastic transformation as well as tumor maintenance.
Several animal models have been generated to study the role of K-ras mutant in tumorigenesis [28,33,37,38]. Animal model with oncogenic alleles of K-ras LA was initially constructed via homologous recombination [37]. The K-ras LA mice have developed a variety of tumor types including lung neoplastic transformation [37]. However, an early mortality was caused by an overwhelming number of predominantly early-stage lung lesions which limits the capacity for tumor progression [37]. A transgenic mouse allele with doxycycline-inducible K-ras4B G12D expression was established under the control of rat Clara cell secretory protein CCSP promoter [38]. Although this mouse allele has been proved to be very helpful in the regulation of K-ras expression with time, it's a transgenic line with potentially multiple copies and uncertain genomic insertion sites which may influence the function of K-ras mutant. A significantly improved animal model was generated by placing a Loxp-stop-Loxp cassette fragment in front of K-ras G12D coding region, in which the K-ras mutant doesn't express until the removal of stop codon by either Adeno-Cre treatment or tissuespecific CRE transgene expression [28,33]. This model has been widely used for cancer studies, especially for the study on tumor initiation and progression process [4,7,29,39,40]. With this model, once the K-ras mutant is activated, it can't be turned off, which may limit it's usage in studying tumor maintenance. Our animal model with ER knock-in into K-ras gene allele allows us to temporally control the K-ras activity via tamoxifen administration, which makes it possible to dissect the role of K-ras mutant at different stages of tumorigenesis.
Tamoxifen is an anti-estrogen commonly used in the treatment of breast cancer. Previous studies have showed that a dosage range between 10 nM to 1 mM is able to inhibit cell growth [41,42]. Therefore, the tamoxifen stimulation of ER-fusion protein activation may somehow be dependent on the dosage. Consistently, a 4OHT concentration-dependent induction of fusion protein activity were also observed previously [1,24,25,26,27,43,44]. In our study, we have observed that a dose of tamoxifen at 0.05 mM works effectively for stimulating the activation of ER-K-ras G12D and cell proliferation. However, we are not certain about the optimal dosage of tamoxifen for in vivo treatment. As we know, it usually takes about more than 20 weeks for us to observe tumor nodules on lung surface in K-ras G12D mouse model [45] and our unpublished data). In contrast, simultaneous p53 deficiency with K-ras activation significantly facilitates lung tumor progression. Most K-ras G12D mice with p53 deficiency develop large tumors on lung surface at 10 weeks post Adeno-Cre treatment [46]. To shorten the potential tamoxifen treatment duration and make our in vivo study more practical, we have therefore tested the tamoxifen dosage in P53L/L, ER-K-ras G12D mice after Adeno-Cre treatment. Interestingly, we have found that tamoxifen treatment at 500 mg is sufficiently to drive lung hyperplasia in P53L/L, ER-K-ras G12D mice. However, the tumor latency is much longer than that of the P53L/L, K-ras G12D mouse model. It takes about 12 weeks of tamoxifen treatment for the P53L/L, ER-K-ras G12D mice to develop lung tumors. One possibility for this could be the low level of ER-K-ras G12D activation by tamoxifen treatment at 500 mg. Future study is definitely necessary to optimize the tamoxifen dosage to achieve the full-activation of ER-K-ras G12D .
In sum, we have here constructed a mice allele with regulatable K-ras and demonstrated that the activation of K-ras is important for tumor formation as well as tumor maintenance. This animal model may provide a useful platform for further detailed characterization of the role of K-ras during different stages of tumorigenesis.  Figure S3 The expression of K-ras and ER-K-ras on mRNA and protein level in different MEFs. A) The expression of K-ras has no obvious change on mRNA level in Loxp-Stop-Loxp-K-ras G12D MEFs and Loxp-Stop-Loxp-ER-Kras G12D with or without Adeno-Cre treatment. B-C) The expression of ER-K-ras G12D has no obvious change on mRNA level in ER-K-ras MEFs in the absence (B) or in the presence (C) of tamoxifen treatment. D) Detection of ER and ER-K-Ras protein level in MEFs infected with or without Adeno-Cre in the presence or absence of 0.05 mM tamoxifen treatment in indicated MEFs. bactin serves as internal control. (TIF) Figure S4 500 mg per mouse per day tamoxifen treatment could induce ER-K-ras G12D activation and lead lung hyperplasia (A and B).