|Home | About | Journals | Submit | Contact Us | Français|
Type I and type II classes of interferons (IFNs) signal through the JAK/STAT1 pathway and are known to be important in adaptive and innate immune responses and in protection against tumors. Although STAT1 is widely considered a tumor suppressor, it remains unclear, however, if this function occurs in tumor cells (cell autonomous) or if STAT1 acts primarily through immune cells. Here, the question of whether STAT1 has a cell autonomous role in mammary tumor formation was addressed in a mouse model of ERBB2/neu-induced breast cancer in the absence and presence of STAT1. For this purpose, mice that carry floxed Stat1 alleles, which permit cell-specific removal of STAT1, were generated. To induce tumors only in mammary cells lacking STAT1, Stat1 floxed mice were crossed with transgenic mice that express cre recombinase and the neu oncogene under the mouse mammary tumor virus LTR (Stat1fl/fl NIC). Stat1 was effectively deleted in mammary epithelium of virgin Stat1fl/fl NIC females. Time-to-tumor onset was significantly shorter in Stat1fl/fl NIC females than in WT NIC (Wilcoxon rank sum test, P = .02). The median time-to-tumor onset in the Stat1fl/fl NIC mice was 49.4 weeks, whereas it was 62.4 weeks in the WT NIC mice. These results suggest that STAT1 in mammary epithelial cells may play a role in suppressing tumorigenesis. The Stat1 floxed allele described in this study is also a unique resource to determine the cellular targets of IFNs and STAT1 action, which should aid our understanding and appreciation of these pathways.
Protection against tumors is critically dependent on the host immune response . Interferons (IFNs) are cytokines classified as type I (IFNα, β, and others) and type II (IFNγ) that play an integral role in the immune response to pathogens and tumors. IFNs are known to be of central importance in tumor clearance; however, the mechanism of their action is unclear . STAT1 is a member of the signal transducers and activators of transcription (STAT) family of cytokine-inducible transcription factors. STATs become active transcription factors after cytokine activation of a cytokine receptor complex. They are tyrosine phosphorylated by receptor-associated janus kinases (JAKs), leading to their dimerization and translocation to the nucleus. STAT1 homodimers exclusively mediate IFNγ signaling, and STAT1 forms a complex with STAT2 and IRF9 (the ISGF3 complex) to mediate type I IFN signaling . As a necessary mediator of IFN signaling, STAT1 has distinct roles in both the adaptive and innate host immune response . These functions were demonstrated in whole animal Stat1 knockout mouse studies, which display hypersensitivity to viral and bacterial infections [4,5]. Humans with loss-of-function mutations in STAT1 have been diagnosed with Mendelian predisposition to mycobacterial diseases [6,7].
Evidence that STAT1 has tumor suppressor properties includes studies in mouse models of STAT1 and IFNγ receptor deficiency. Stat1 and IFNγ receptor null mice show increased incidence of spontaneous tumors when exposed to methylcholanthrene or bred into a p53-deficient background . Administered systemically, IFNα has antitumor activity and it is used clinically as part of a chemotherapy regimen to treat several forms of cancer [9–12]. In vitro, IFNs can inhibit proliferation of tumor cells through STAT1 . In addition, STAT1 has been suggested to mediate the tumorigenic properties of osteopontin in mammary epithelial tumor cells .
Whereas some studies have demonstrated a role of STAT1 as a tumor suppressor, others have linked constitutive activation of STAT1 and/or an overexpression of IFN-related genes to breast cancers associated with poor prognosis, and they have suggested that activated STAT1 may confer resistance to radiation and adjuvant cancer therapy [15,16]. In fact, a role for STAT1 in promotion of leukemia development has been reported . The difference between these studies and those that suggest a tumor suppressor role for STAT1 may reflect the cell autonomous versus cell nonautonomous roles of STAT1.
A complete understanding of the cellular targets of IFNs and the cell autonomous and nonautonomous roles of STAT1 is lacking in part because of the limitations of whole animal knockout models. To determine IFN/STAT1 functions in individual cell types, transplantation of STAT1 null cells into irradiated recipients, NOD severe combined immunodeficient mice or RAG2-/- mice, may be done; however, because of the inherent manipulation of the immune system in the hosts, this approach does not allow for a complete appreciation of the role of STAT1 in the immune system. To address this issue, we have developed a floxed Stat1 allele to enable tissue-specific disruption of STAT1 expression. Here we report the use of this tool to identify the cell autonomous role of STAT1 in mammary tumor formation using a mouse model of neu/ERBB2-induced breast cancer. ERBB2 is a receptor tyrosine kinase that can contribute to breast cancer when mutated or abnormally expressed. Approximately 20% to 30% of primary human breast cancers have an amplification of the ERBB2 gene or overexpression of ERBB2 . Our results indicate that tissue-specific removal of STAT1 reduces the average latency of Neu/ERBB2 tumors. The mouse Stat1 conditional allele introduced in this study serves as a tool that enables the identification of the role of STAT1 in mammary tumor formation independent of the role of STAT1 in the immune system. This model should be valuable for fully appreciating the role of STAT1 in different cell types and in discriminating between immune and nonimmune functions of IFN/STAT1 signaling.
Loxp sites were inserted into Stat1 locus between -750 from the transcriptional start site and flanking the first three exons including exon 3, which is the first translated exon of Stat1. To execute this targeting strategy, a plasmid target vector containing the pLoxp3neoTK backbone with three polymerase chain reaction-amplified homology regions from the Stat1 locus of 129SvEv mouse genomic DNA was constructed. Electroporation of the targeting vector in 129SvEv ES cells was completed by Ingenious Targeting Laboratories (ITL, Stony Brook, NY). Screening for successfully targeted ES clones was done by Southern blot analysis using SpeI digested DNA and a 500-bp probe matching an outside region upstream of the 5′ homology arm. Of 44 clones screened, 2 were positive for the 7.7-kb fragment, indicating a successfully targeted Stat1 locus. Microinjection of blastocysts and breeding of chimeras was performed at ITL.
Stat1 floxed mice were crossed with mice containing the mouse mammary tumor virus (MMTV)-neu-IRES-cre transgene described previously . Animals were housed on a 12-hour light-dark cycle and maintained on standard rodent chow. All procedures were approved by the National Institute of Diabetes and Digestive and Kidney Diseases animal care and use committee under National Institutes of Health guidelines.
Mammary glands were dissected from animals after killing. Glands were fixed in 4% formaldehyde, and paraffin sections were prepared (Histoserv, Germantown, MD). Antigen unmasking was carried out using a decloaking chamber (Biocare Medical, Concord, CA). For staining, primary antibodies against E-cadherin (BD Biosciences, San Jose, CA), STAT1 (no. SC346; Santa Cruz Biotechnology, Santa Cruz, CA), and phospho-STAT1 (no. 9171; Cell Signaling Technology, Danvers, MA) were used. Detection was performed using secondary antimouse and antirabbit immunoglobulin G fluorescent antibodies (Invitrogen, Carlsbad, CA).
Protein was extracted from mammary tumors or cells using lysis buffer containing 50 mM Tris pH 7.6, 150 mM NaCl, 1% Triton X-100, 10% glycerol, 2 mM EDTA, 10 mM tetrasodium pyrophosphate, 50 mM sodium fluoride, and Complete Mini protease inhibitor cocktail (Roche, Indianapolis, IN). Equal amounts of protein were separated on a 4% to 12%NuPage gradient gel (Invitrogen) and transferred to a polyvinylidine fluoride membrane (Invitrogen) and incubated with antibodies against STAT1 (no. SC346; Santa Cruz Biotechnology) and phospho-STAT1 Tyr. 701 (no. 9171; Cell Signaling Technology). Membranes were probed overnight with primary antibodies and were incubated the following day with for 1 hour with HRP-conjugated secondary antibodies (GE Healthcare, Chalfont St Giles, UK). HRP was detected using ECL from Thermo Fisher Scientific (Waltham, MA) and was exposed on radiographs (GE Healthcare).
Tumors (approximately 300 mg) were minced into a paste with a razor blade. The paste was transferred to a 50-ml tube and tumor pieces were washed three times with 25 ml of Dulbecco's modified Eagle medium (DMEM)/F12 + 5% FBS by allowing cells to settle 5 minutes between washes. Pieces were transferred to a T75 flask standing upright and disaggregated by using 8000 U of collagenase (I-05516; Sigma, St Louis, MO)/40 ml of DMEM/F12 + 10% FBS overnight at 37% without shaking and were pipetted repeatedly the next morning. The mixture was washed five times with DMEM/F12 + 5% FBS as before. Cells were plated in DMEM/F12 + 10% FBS + l-glutamine + antibiotic/antimycotic (Invitrogen) + 0.3 µM hydrocortisone (H-0135; Sigma) + 10 ng/ml apotransferrin (T-1428; Sigma) + 5 µg/ml insulin (C-9881; Sigma) + 5 nM β-estradiol (E-2758; Sigma) + heparin 4% x 10-4 (Stem Cell Technologies) + 4 ng/ml epidermal growth factor (Stem Cell Technologies).
We compared time-to-tumor onset among WT NIC and Stat1fl/fl NIC mice using Kaplan-Meier curves, Wilcoxon rank sum, and log-rank tests as calculated by PROC LIFETEST, SAS 9.1 (SAS, Cary, NC).
To determine cell-specific roles of IFN/STAT1 signaling in breast cancer, Stat1 floxed mice were generated, which permitted the specific abrogation of STAT1 in mammary epithelium. Loxp sites were inserted by homologous recombination of a targeting vector in 129SvEv ES cells. One loxp site is located 750 bp upstream of the transcriptional start site and another flanks the first translated exon, a total distance of 4.0 kb (Figure 1A). Southern blot analysis using SpeI digestion and an external probe was used to confirm successful targeting of the Stat1 locus. The expected 7.7-kb fragment of a successfully targeted Stat1 locus was found in 2 of 44 neomycin-resistant mouse ES cell clones (Figure 1B). Recombination of the loxp sites by cre-mediated recombination was expected to remove 4.0 kb of the Stat1 locus and abrogate Stat1 expression (Figure 1C). To test the ability of this Stat1 floxed allele to be recombined by cre recombinase, targeted Stat1 floxed mice were crossed with transgenic MMTV-neu-IRES-cre (NIC)-expressing mice to generate Stat1 floxed NIC mice. Stat1 floxed NIC females expressed cre recombinase and mediated recombination in the germ line, similar to previously studied MMTV-cre-expressing mouse lines , producing offspring containing a combination of alleles we termed Stat1fl (5′ and middle loxp sites recombined), Stat1flneo (original unrecombined locus containing the neomycin resistance gene and containing three separate loxp sites), and null (3′ and 5′ loxp sites recombined; Figure 1C).
To test the ability of the flneo and fl allele to produce STAT1 protein in the absence of cre recombinase, liver, thymus, spleen, and mammary gland tissues were removed from Stat1+/+, Stat1 flneo/fl, and Stat1 flneo/flneo. Unexpectedly, in each tissue, the Stat1 flneo/fl mice produced only approximately 50% of the STAT1 protein. In addition, Stat1 flneo/flneo tissues were nearly devoid of STAT1 protein (Figure 1D). This result indicates that theneomycin resistance gene present in the Stat1flneo allele disrupted the critical transcriptional regulatory elements in the Stat1 gene. Because the Stat1fl allele missing neomycin expressed STAT1 at expected levels, this allele was used for all subsequent conditional knockout studies.
To test the ability of the recombined null allele to abrogate STAT1 protein expression, we crossed animals with a single copy of the null allele to generate embryos homozygous for the null allele (Stat1-/-). From these embryos, we isolated mouse embryonic fibroblasts (MEFs) to assess the completeness of the STAT1 protein abrogation. Using MEFs with one Stat1 floxed allele and one WT allele (Stat1 fl/+) as a control, Western blot analysis demonstrated that Stat1-/- MEFs did not express any detectable STAT1 protein (Figure 2). Antibodies directed against the STAT1 C-terminus were used to control for the possibility that only an N-terminal-truncated product was produced.
To test if STAT1 signaling within mammary epithelium affects the establishment of neu-induced mammary tumors, tumor occurrence in Stat1fl/fl NIC mice and WT NIC mice was followed up for a period of up to 97 weeks (Figure 3). The earliest onset of tumors was similar in both groups, with mice from both groups developing tumors by approximately 36 weeks. However, there was evidence that time-to-tumor development was shorter for Stat1fl/fl NIC versus WT NIC mice. The evidence was stronger for the early period than for the later one (Wilcoxon rank sum test, P =.02) and somewhat weaker for the overall period (log-rank test, P =.08). The median time-to-tumor onset in the Stat1fl/fl NIC mice was 49.4 weeks, whereas it was 62.4 weeks in the WT NIC mice (Figure 3). To monitor STAT1 levels in WT NIC mammary tumors and the efficiency of the Stat1 fl/fl NIC model to remove Stat1 in mammary tumors, immunohistochemical (IHC) staining of STAT1 was performed. STAT1 was readily detectable in WT NIC tumors but not in tumors from Stat1fl/fl NIC mice. These results indicate that STAT1 was efficiently abrogated in mammary tumors from Stat1fl/fl NIC mice and that STAT1 in the mammary epithelium may have a suppressive role in the development of mammary tumors. MMTV-NIC expression has previously been shown to be mainly restricted to the mammary compartment, with very weak expression in immune cells . However, because STAT1 expression remains intact in immune cells, our results do not rule out the importance of STAT1 in the immune system in protecting the host against tumor formation.
To confirm that STAT1 is deleted efficiently in mammary epithelium, mammary tissues from WT NIC and Stat1fl/fl NIC mice were subjected to IHC analysis using the C-terminal STAT1 antibody or phospho-STAT1 antibody (Figure 4). After an intraperitoneal IFNγ injection before harvesting tissues, nuclear STAT1 was visible in WT NIC but not Stat1fl/fl NIC mammary epithelium. These results confirmed that Stat1 can be efficiently deleted from cells expressing cre recombinase and provide a means to test the tissue-specific role of STAT1 in mammary tumor formation.
STAT1 expression was assessed in mammary tumors from WT NIC and Stat1fl/fl NIC mice to confirm that tumors were not the result of a selection of unrecombined cells. Tumors from IFNγ-injected WT NIC and Stat1fl/fl NIC mice were analyzed by IHC staining for the presence of STAT1. WT NIC (Figure 5A) but not Stat1fl/fl NIC tumors (Figure 5B) showed abundant STAT1 staining. Because the IFNγ did not penetrate the tumor in the period (30 minutes) of treatment, pSTAT1 staining was only seen near blood vessels of WT NIC tumors (Figure 5C). No pSTAT1 staining was observed in Stat1fl/fl NIC tumors (Figure 5D). In addition, Western blot analysis revealed that STAT1 was undetectable in tumor cells isolated from Stat1fl/fl NIC tumors (Figure 5E).
To confirm that the loss of STAT1 in mammary epithelium did not result in abnormal mammary development or cause tumors, Stat1fl/fl mice were crossed with MMTV-cre and WAP-cre mice . Parous Stat1fl/fl MMTV-cre females lactated normally, indicating that loss of STAT1 did not impair mammary function (not shown). Whole mounts of parous 15-month-old Stat1fl/fl WAP-cre females demonstrated that the lack of STAT1 did not affect the normal mammary architecture or lead to hyperplasia (Figure 6).
The goal of the current study was to generate a mouse model that will be useful in determining cell autonomous STAT1 actions in mammary tumor formation. In pursuit of this goal, we have generated a conditional Stat1 allele in mice that can efficiently remove Stat1 in various cell types in the presence of cre recombinase controlled by a tissue-specific promoter. Generation of the Stat1 floxed allele was accomplished by addition of loxp sites flanking a 4-kb region of the Stat1 gene approximately 750 bp upstream of the transcription start site and just 3′ of the first translated exon. Surprisingly, addition of the 1.9-kb neo cassette resulted in near-complete disruption of STAT1 expression, indicating a previously unreported vital promoter region at -750 bp. For conditional studies, this challenge was overcome by obtaining mice with the neomycin cassette removed, which restored expression of STAT1. In these mice, only after cre-mediated recombination and disruption of this 4-kb region did loss of STAT1 protein expression occur as detected by a C-terminus-specific STAT1 antibody.
Of particular interest in this study was identifying if STAT1 signaling independent of the immune system has a role in tumor suppression. STAT1 has been primarily considered a tumor suppressor protein because of the substantial increase in tumor susceptibility of STAT1 null mice to various tumor-causing insults . STAT1-deficient mice are more prone to chemically induced tumors, as well as spontaneous tumors induced as a result of a p53 mutation. However, lack of STAT1 activity does not always enhance tumor formation in the presence of an oncogenic stimulus [15,17].
Using mice with an oncogene (neu) expressed under the same promoter as cre recombinase, we were able to test the hypothesis that STAT1 has a tumor suppressor role in mammary epithelium. The lack of STAT1 did reduce the median time onset of mammary tumors in neu-expressing mammary epithelium, suggesting a tumor-suppressing action of STAT1 in these cells. Although this study does not exclude a tumor-suppressing action of STAT1 through immunosurveillance, it suggests that STAT1 can play a role in mammary epithelium and delays tumor occurrence.
In summary, we have developed a floxed Stat1 allele that should be a valuable tool for discriminating the cell-specific roles of STAT1 in mice. The current study suggests that STAT1 plays a suppressive role in ERBB2/neu induced breast cancer. We anticipate the Stat1 floxed mice will provide many other insights into the cell-specific roles of IFN and STAT1 signaling not previously appreciated.
1This work was supported by the National Institutes of Health intramural research programs of National Institute of Diabetes and Digestive and Kidney Diseases and National Cancer Institute.