Survivin is an essential mitotic gene, and this has been speculated to reflect its primary function in development and cancer. Here, we generated a knock-in transgenic mouse (SVVp-GFP) in which a green fluorescent protein (GFP) reporter gene was placed under the control of the survivin promoter that regulates transcription at mitosis. The expression of endogenous survivin was widespread in mouse tissues during development and shortly after birth. In contrast, GFP reactivity was undetectable in transgenic mouse embryos, and was largely limited postnatally to mitotic cells in the testes. Double transgenic mice generated in the tumor-prone Min/+ background exhibited intestinal adenomas that strongly expressed endogenous survivin, but only isolated GFP-positive cells. Conversely, dysplastic adenomas (16%) stained intensely for GFP, and revealed focal reactivity for mutant, but not wild-type, p53. The expression of GFP was increased by ∼10-fold in p53−/− as opposed to p53+/+ HCT116 colorectal cancer cells, and reintroduction of p53 in p53−/− cells abolished GFP expression. Therefore, the mitotic transcription of the survivin gene is highly restricted in vivo, and unexpectedly negatively regulated by p53. Contrary to a commonly held view, the dominant function(s) of survivin in development and tumor ontogeny are largely cell cycle–independent. (Cancer Res 2006; (66)7: 3392-5)
Survivin is a unique member of the inhibitor of apoptosis (IAP) gene family (1) with dual functions in the control of mitosis and preservation of cell viability (2). It has been difficult, however, to determine which of these two roles is dominant in vivo, and conditional knockout studies in mice produced conflicting phenotypes with only defects of cell division (3), only exaggerated apoptosis (4), or both (5). Insights into the function of survivin may come from analysis of gene expression. Accordingly, the cell cycle–dependent transcription of the survivin gene at mitosis (6), coupled with the localization of the protein to the mitotic apparatus and the multiple mitotic defects ensuing from survivin loss/targeting (2), have been taken to suggest that this pathway operates predominantly, if not exclusively, in the control of mitosis (7, 8). Conversely, cell cycle–independent mechanisms of survivin gene expression have been described, and linked to the inhibition of apoptosis in various cell types (2). Elucidating how the survivin gene is regulated in vivo is important for the indispensable role of this pathway during development (4), and the sharp differential expression of survivin in cancer as opposed to normal tissues, which has been pursued for novel cancer diagnostics and therapeutics (2).
To address these questions, we have generated a knock-in transgenic mouse that expresses a green fluorescence protein (GFP) reporter gene under the control of the proximal survivin promoter (9). This DNA region has been characterized for maintaining cell cycle periodicity at mitosis via CDE/CHR G1-repressor elements and Sp1-type boxes (9). Unexpectedly, we found that the expression of survivin during development and tumor formation is largely cell cycle–independent.
Materials and Methods
Plasmid construction and generation of transgenic mice. A mouse cDNA containing the proximal 830 nucleotide of the survivin promoter (mS) upstream of the translational initiation codon (9) was inserted into BamHI and HindIII sites of pBluescript II KS. The construct was fused to a full-length cDNA encoding an enhanced GFP (EGFP) gene at the 3′ end, plus SV40 splice and polyadenylation sequences to generate pBS-mS-830-GFP. A control pEGFP-C1 plasmid was obtained from BD Biosciences Clontech (Palo Alto, CA). All experiments involving animals were approved by an institutional animal care and use committee. The pBS-mS-830-GFP targeting vector was purified, confirmed by DNA sequencing, and microinjected (200 ng/μL) into C57BL/6 embryos, which were transferred into pseudopregnant females. Littermates were screened by PCR of tail genomic DNA using primers (10 pmol) corresponding to mS (5′-AAGTTGGGCTGCAGGAATTCCCGCT-3′) and GFP (5′-CTTGTCGGCCATGATATAGACGTTGTG-3′). Colonies from three independent transgene-positive founder mice (SVVp-GFP), or control littermates were maintained as described (10). C57BL/6J-APCMin/J breeder mice were purchased from Jackson Laboratory (Bar Harbor, ME) and crossed with SVVp-GFP mice. Littermates were screened by PCR of tail genomic DNA with primers (10 pmol) corresponding to pMin (5′-TTCTCGTTCTGAGAAAGACAGAAGTTA-3′) and pCOM (5′-TTTATATTCCACTTTGGCATAAGGC-3′), and double transgene-positive mice were maintained as described (10).
RT-PCR. Total RNA was prepared from total mouse embryos collected at E12, E14, or E18, from individual tissues isolated from 6- to 12-week-old mice or at days 1, 3, or 5 postnatally using RNeasy Mini Kit (Qiagen, Valencia, CA), and cDNA was synthesized using a First-Strand Synthesis System (Invitrogen, Carlsbad, CA), according to the manufacturer's recommendations. mRNA expression was determined by RT-PCR amplification with the following primers: pGFP-1, 5′-AGCTGACCCTGAAGTTCATCTG-3′; pGFP-2, 5′-GATCTTGAAGTTCACCTTGATGC-3′; pM-Survivin-1, 5′-GCAGCTGTACCTCAAGAACTACC-3; pM-Survivin-2, 5′-TGCAGTCTCTTCAAACTCTTTTTG-3; pGAPDH-1, 5′-ACGGATTTGGTCGTATTGGGCG-3; pGAPDH-2, 5′-CTCCTGGAAGATGGTGATGG-3.
Cell culture, antibodies, and Western blotting. Cervical carcinoma HeLa, and p53+/+ or p53−/− HCT116 colorectal adenocarcinoma cells were maintained in culture as described (11). Transient transfection experiments were carried out using LipofectAMINE (Invitrogen), according to the manufacturer's recommendations. Cell cycle synchronization using thymidine block and release was carried out as described (6). Cells were lysed in 50 mmol/L Tris-HCl (pH 7.5), 150 mmol/L NaCl, 5 mmol/L EDTA, 50 mmol/L NaF, 0.5% Nonidet P40 plus protease inhibitors (Roche Applied Science), and protein-normalized extracts (40 μg) were processed for Western blotting, as described (11). The following antibodies to survivin (Novus Biologicals, Littleton, CO), GFP (BD Biosciences), mutant/wild-type p53 (ab 26, Abcam), wild-type p53 (ab16776, Abcam, Cambridge, MA), or β-actin (Sigma-Aldrich, St. Louis, MO) were used.
Immunohistochemistry. Tissue expression of survivin, p53 or GFP was determined with the Histostain-Plus (Zymed, South San Francisco, CA) system after antigen retrieval by pressure cooking, as described (10).
Results and Discussion
To determine whether cell cycle regulation of survivin accounts for its expression in development and tumor formation, we generated a knock-in transgenic mouse that expresses GFP under the control of an 830 nucleotide genomic fragment encoding the core survivin promoter, and with SV40 regulatory sequences at the 3′ end (Fig. 1A). The proximal region of this promoter (∼300 nucleotides) contains cell cycle elements for transcription at mitosis (Fig. 1B), which were previously confirmed by site-directed mutagenesis and promoter activity (9). Transfection of the pBS-mS-830-GFP targeting construct (Fig. 1A) in synchronized HeLa cells resulted in strict cell cycle–regulated GFP expression (Fig. 1C), with undetectable levels in interphase, and peak activity 8 to 12 hours after thymidine release, which coincides with progression through mitosis, by DNA content analysis and flow cytometry (Fig. 1D). Endogenous survivin was also regulated in a cell cycle–dependent manner peaking at mitosis in synchronized HeLa cells (6), albeit high levels of expression were also found in interphase and asynchronous cultures (Fig. 1C).
Confirmed transgenic mice containing the pBS-mS-830-GFP targeting construct (SVVp-GFP) were next analyzed for the expression of GFP and endogenous survivin during development. Previous data suggested that survivin is abundantly and ubiquitously found in embryonic and fetal development, progressively down-regulated postnatally, and largely undetectable in most adult tissues (12). Consistent with this, endogenous survivin was strongly expressed in 14-day-old embryos of SVVp-GFP transgenic mice or nontransgenic littermates (Fig. 2A), and was present in the thymus, lung, spleen, and kidney at postnatal day 5 as determined by RT-PCR (Fig. 2B). In contrast, GFP expression was undetectable in 14-day mouse embryos of SVVp-GFP transgenic mice (Fig. 2A), and was faintly present in brain and kidney, but not in other tissues at postnatal day 5, as determined by RT-PCR (Fig. 2B). In developed mouse tissues (6-12 weeks of age), GFP or endogenous survivin was undetectable with the exception of brain, in agreement with recent observations (13). Conversely, both GFP and endogenous survivin were highly expressed in the testes of SVVp-GFP transgenic mice as determined by RT-PCR (Fig. 2A) and Western blotting (Fig. 2C). By immunohistochemistry of 8-week-old testes, GFP-positive cells in SVVp-GFP transgenic mice were identified morphologically as mitotic cells, preferentially at the metaphase transition (Fig. 2D,, inset, and E) whereas nontransgenic animals had no GFP reactivity in the testes (Fig. 2F and G). Identical results were obtained with colonies from independent SVVp-GFP founder mice.
We next studied the mitotic regulation of survivin gene expression during tumor development, and we generated double transgenic mice expressing SVVp-GFP on the tumor-prone Min/+ background. In this model, a mutation in the mouse homologue of the adenomatous polyposis (APC) gene results in the formation of multiple intestinal adenomas (14), which are histologically characterized by epithelial hyperplasia and disorganized glandular architecture, as determined by H&E staining (Fig. 3A). Adenomas in double transgenic mice stained intensely for endogenous survivin (Fig. 3B), whereas GFP reactivity was restricted to isolated cells (Fig. 3C), and a control IgG was negative (Fig. 3D). In contrast, 2 out of 12 adenomas (16%) formed in double transgenic mice exhibited morphologic features of dysplasia, which coincided with extensive labeling for GFP (Fig. 3F), and focal reactivity for mutant (Fig. 3G), but not wild-type, p53 (Fig. 3H). A control IgG gave no staining in dysplastic adenomas (Fig. 3E).
To test a role of p53 in negatively regulating mitotic survivin gene expression, we used p53+/+ and p53−/− HCT116 colorectal cancer cells (11). Transfection of control pEGFP resulted in the strong expression of GFP, which was indistinguishable in p53+/+ or p53−/− cells, as determined by fluorescence microscopy (Fig. 4A) and Western blotting (Fig. 4B). In contrast, transfection of the SVVp-GFP targeting construct pBS-mS-830-GFP resulted in a ∼10-fold increased GFP expression in p53−/− as compared with p53+/+ HCT116 cells, as determined by fluorescence microscopy (Fig. 4A) and Western blotting (Fig. 4B). Reintroduction of wild-type p53 in p53−/− HCT116 cells largely abolished the expression of the pBS-mS-830-GFP targeting construct, but did not affect pEGFP reactivity (Fig. 4C).
In summary, the mitotic regulation of the survivin gene is highly restricted in vivo, and does not account for the widespread expression of endogenous survivin in development and tumor formation. Our results counter a commonly held view that the primary function of survivin is at cell division (7, 8), and argue that survivin gene expression in vivo is largely mediated by cell cycle–independent mechanisms. This may reflect the ubiquitous role of survivin in apoptosis inhibition, which contributes to tumorigenesis (15), and, at least in certain tissues, is indispensable for development (4). Unexpectedly, the mitotic component of survivin gene expression is negatively regulated by p53, in a pathway that may involve direct transcriptional silencing via putative p53-responsive elements (16), or indirect promoter squelching (17). Inhibition of mitotic survivin gene expression by p53 may cooperate with checkpoint functions at the G2-M phase, especially p53-dependent apoptosis (18), as acute loss of survivin in dividing cells results in massive cell death (19). Conversely, loss of p53 during tumor progression may cooperate with cell cycle–independent mechanisms to further deregulate survivin gene expression in vivo, ablating mitotic checkpoint mechanisms and promoting resistance to apoptosis of cells traversing mitosis (19).
Grant support: NIH grants CA90917, CA78810, and HL54131.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Drs. Bert Vogelstein for generously providing HCT116 cells, Jianyuan Luo for the p53 construct, Ron Lubet for discussion, and David S. Garlick for reviewing histology.