The SIL Gene Is Essential for Mitotic Entry and Survival of Cancer Cells Free

SIL was cloned from the most common chromosomal rearrangement in T cell acute lymphoblastic leukemia (1). The SIL gene encodes a cytosolic protein with no homology to known proteins that is expressed mainly in proliferating cells (2–4). Upon entrance to the cell cycle, SIL is induced in an immediate early fashion, reaches peak levels in mitosis and then degrades upon transition into the next cell cycle (3). Recently, we have shown that SIL is phosphorylated during mitosis, a process crucial for its interaction with the mitotic regulator PIN1 (5).

We have observed that SIL is expressed in multiple cancers and that its expression correlates with the expression of mitotic spindle checkpoint genes, with the pathologic mitotic index, and with worse prognosis (6, 7). Here, we show that SIL is essential for mitotic entry and the survival of cancer cells in vitro and in vivo.

Plasmids. pTR, pSuper, and pTER were provided by van de Wetering et al. (8). For primers and oligonucleotide sequences, see Supplementary Table S2. For SIL-short hairpin RNA (shRNA), the construction of pSuper/pTER was done as published (9). The inducible system was generated by the transfection of pTer-plasmid carrying the tetracycline operator with an insert of shRNA specific for the human SIL into LS174T colon cancer cells carrying the tetracycline repressor (8). Two other duplexes of dsRNA oligonucleotides were designed using the BLOCK-iT RNAi designer by Invitrogen (Supplementary Table S1).

Chemical staining. Crystal violet staining was done as published (8). For PAS staining, cells were washed and stained with Schiff's reagent (Merck) according to protocol. For Giemsa staining, cells were incubated with colcemid (Biological Industries) for 16 h, washed with PBS, and subjected to RBC hypotonic lysis buffer (NH₄Cl, 41.5 gm; KNCO₃, 5 gm; EDTA, 1 mL; H₂O, 5 L). Following cytospin, cells were fixed in May-Grunwald stain, rinsed with water, and stained with Giemsa (Sigma-Aldrich Company) for 7 min.

Cell cycle synchronization and analysis. Prometaphase-synchronized cells were obtained by treating exponentially growing cells with 2.5 mmol/L of thymidine for 17 h, washing twice with PBS buffer, growing them in fresh medium for 9 h, and then re-treating the cells with 2.5 mmol/L of thymidine for 16 h. After 4 h of release, cells were treated with paclitaxel for the indicated times. Cells were analyzed by a two-color flow cytometry using the FACScan (BD Biosciences). For each sample, 5,000 events were collected and analyzed. Antibodies used were antimitotic protein monoclonal (MPM2; Upstate; ref. 10) and FITC-conjugated goat anti-mouse IgG (Biosource). A Roche Diagnostics kit was used for Annexin-propidium iodide (PI) analysis.

Fluorescence in situ hybridization probes and procedures. Cells were concentrated by cytospin and fixed with methanol/acetic acid (3:1). LSI BCR/ABL and LSI BCR/ABL extra-signal dual-color DNA probe kits were used (Vysis). Fluorescence in situ hybridization (FISH) was done as developed by Esa et al. (11). Cdc2/cyclin B1 kinase assays were done as previously published (5).

In vivo experiments. All experiments involving mice were approved by the institutional animal care and use committee. Six- to eight-week-old female nonobese diabetic/severe combined immunodeficiency (NOD/SCID) mice (Weizmann Institute) were injected s.c. into the right flanks with 2 × 10⁶ exponentially growing LS174T cells in 200 mL of sterile PBS (pH 7.4). Tumor volume was measured in three dimensions with a caliper. Animals with a tumor burden of >2,000 mm³ were sacrificed. Mice were treated with 2 mg/mL of doxycycline (Sigma) in 2.5% sucrose in their drinking water or with 2.5% sucrose as control, from day 5 after injection of cells. The drinking water was changed twice a week to avoid doxycycline toxic metabolites. After the animals were sacrificed, tumor was removed and stained as described (6).

Inducible shRNA-mediated knockdown of SIL arrests the growth of cancer cells in vitro. To study the role of SIL in cancer cells, we designed an inducible knockdown of SIL in the colon cancer cell line LS174T that stably expresses the tetracycline repressor (8). shRNA is produced after the addition of tetracycline (in the form of doxycycline) to the growth medium. Eight stable clones for the inducible SIL shRNA construct were isolated (Supplementary Fig. S1). To correlate the phenotype with the expression level of SIL, we chose clones with variable levels of SIL after the addition of tetracycline; i.e., clone no. 2 versus no. 6 (Supplementary Fig. S1A). SIL expression was down-regulated >6-fold in clone no. 2, whereas there was almost no change in SIL levels in clone no. 6.

Exposure of the LS174T clones to tetracycline caused a growth arrest that correlated with the magnitude of SIL down-regulation (Fig. 1A). Similar growth suppression was observed in other clones with significant SIL knockdown. These results were confirmed by independent transfection of two short interfering RNA (siRNA) oligos (Supplementary Fig. S2) into seven other cancer cell lines: MCF-7 (breast adenocarcinoma), PANC1 (pancreatic carcinoma), U-87 (glioblastoma), Caki-2 (kidney carcinoma), H1299 (non–small cell lung carcinoma), HeLa (cervix adenocarcinoma), and PC3 (prostate adenocarcinoma). Figure 1B is representative of four such cell lines.

We ruled out nonspecific activation of the IFN pathway by shRNA by reverse transcription-PCR of the IFN-inducible gene, OAS1 (ref. 12; Supplementary Fig. S3). To definitely confirm the specificity of this phenotype, we transfected clone no. 2 cells with a construct encoding the murine Sil (Mu-Sil), which is not affected by human SIL shRNA. After the addition of tetracycline, clones transfected with murine Sil survived and proliferated, showing that murine Sil rescues the cells from the effects of the knockdown of the endogenous human SIL (Fig. 1C). Thus, the inhibition of cell growth is caused by the down-regulation of SIL.

SIL knockdown causes apoptosis coupled with a delay in mitotic entry. The phenotype of growth suppression caused by SIL knockdown could result from different mechanisms: induction of differentiation, cell death, or perturbation of the cell cycle.

Differentiated colon cells show an increase in the expression levels of the GAL4 gene (8). There was no change in GAL4 expression levels after down-regulation of SIL (Supplementary Fig. S4). Thus, the difference in growth could not be attributed to increased differentiation. Instead, there was a marked increase in apoptosis after the induction of SIL knockdown as depicted in Fig. 2A by a flow cytometry assay using PI and Annexin V. The apoptosis was confirmed independently by a caspase 3 cleavage assay (Fig. 2B).

As the SIL expression pattern suggests its involvement in mitosis, we examined the temporal relationship between the apoptosis caused by SIL knockdown and the cell cycle. We synchronized the cells at S-G₂ phase by a double thymidine block and treated them with paclitaxel. At different time points, cells were harvested and analyzed both for mitosis by staining with the MPM2 antibody (a mouse monoclonal antibody specific for phosphorylated mitotic proteins) and for apoptosis by measurement of the sub-G₁ fraction by staining with PI (Fig. 2C). SIL knockdown almost completely prevented mitotic entry coupled with a marked increment of apoptosis upon increasing exposure times to paclitaxel. Thus, it seems that apoptosis was associated with an impediment in G₂ to M transition.

We next studied, in more detail, the effects of SIL knockdown on the mitotic phenotype. Metaphase cell spreads were prepared from LS174T clones exposed to colchicine in the presence of tetracycline (Fig. 3A). There was almost no metaphase in cells from clone no. 2 in which SIL expression was eliminated (panel 3) compared with normal cells or clone no. 6 in which SIL is expressed (panels 1 and 2).

To observe mitoses biochemically, we used the antibody MPM2, which allows the distinction between the G₂ fraction and the M fraction of the cell cycle which are usually combined in the standard flow cytometry analyses of DNA content. As seen in a representative experiment (Fig. 3B), a smaller but still considerable percentage of the SIL knockdown cells had DNA content consistent with either the G₂ or M phases (top, histograms); however, a true mitotic fraction was nearly absent in SIL knockdown cells, as measured by anti-MPM2 antibody (bottom, dot plots). This strongly suggests that in the absence of SIL, the cells do not enter the M phase and undergo apoptosis in the G₂ to M boundary (Fig. 2C).

To further analyze this mitotic phenotype, we did a FISH assay on clone no. 2 with probes targeting the BCR and ABL genes on chromosome 22 and 9, respectively. Cells were exposed to colchicine for 16 h before and after treatment with tetracycline. In correlation with the results above, the majority of SIL knockdown cells were not in mitosis but displayed a clear pattern of doublet signal hybridization, characteristic of the G₂ phase (Fig. 3C).

The major regulator of mitotic entry is the CDK1 (CDC2)-cyclin B complex. To test if the perturbed mitotic entry of SIL knockdown cells is associated with a decreased activity of CDK1, we did a kinase assay for its activity. To arrest cells at metaphase, cells were treated with paclitaxel alone or with paclitaxel and tetracycline. Cell lysates were collected at different time points and the CDK1-cyclin B complex was immunoprecipitated with anti–cyclin B1 antibody. Consistent with the delay in mitotic entry, there was a marked delay in the activation of CDK1 in SIL knockdown cells (Fig. 3D). Thus, cells with low levels of SIL display a marked decrease in mitotic entry associated with the low activity of CDK1.

To further examine the specificity of the shRNA-mediated knockdown of the human SIL, we analyzed the cell cycle and apoptosis patterns of clone no. 2 cells stably expressing the Mu-Sil. The ectopic expression of Mu-Sil rescued these cells from apoptosis and G₂ arrest. Thus, both phenotypes of delayed mitotic entry and the apoptotic phenotype are specifically caused by SIL down-regulation (Supplementary Fig. S5A and B).

SIL is required for cancer growth in vivo. To investigate the requirement of SIL for tumor growth in vivo, we s.c. injected immunodeficient NOD/SCID mice with 2 × 10⁶ cells of LS clone no. 2. Tetracycline was added to the drinking water 5 days after the injections to half of the mice. The growth of the tumors in the tetracycline group was significantly slower that in the untreated group (Fig. 4A and B). As expected, there was no effect of treatment with tetracycline on the growth of tumors derived from clone no. 6, in which SIL shRNA has nearly no effect on SIL levels (Supplementary Fig. S6).

After 3 weeks of continuous exposure to tetracycline, several clone no. 2 tumors began to grow at the same doubling time as those without tetracycline exposure. This escape could have been explained either by selection of cells in which there is no shRNA effect on SIL or by an acquired mutation promoting cell growth despite the continuous reduction of SIL. To distinguish between these possibilities, we analyzed the tumors for SIL mRNA and protein levels. As shown in Fig. 4C and D, SIL was expressed in the growing tumors, suggesting that SIL shRNA was not expressed or was not active in these cells. Thus, the tumors grew only after escaping the RNAi-mediated SIL repression, suggesting that SIL is critical for cell growth and survival.

In this study, we have shown the requirement of SIL for cancer cell survival and mitotic entry. To elucidate the role of SIL in cancer cells, we generated an inducible knockdown of the endogenous human SIL. We show that SIL is necessary for the survival of L174T colon cancer cells in vitro and in vivo, in a manner that correlates with its expression levels. Interestingly, examination of explant tumors in mice that escape the growth suppression of tetracycline revealed normal expression of SIL. These findings suggest that SIL is essential for tumor growth and that the tumor escape is caused by “takeover” of cells in which the shRNA was probably silenced or lost. The apoptosis induced in the absence of SIL does not depend on an intact p53 pathway as it is also observed in p53-deficient cell lines transfected with the SIL-specific siRNA oligos.

In addition to the induction of apoptosis, SIL knockdown resulted in a substantial reduction of mitotic entry. Using morphologic, FISH, and biochemical analyses we showed that in the absence of SIL, cells delay their entry to mitosis and undergo apoptosis. This phenotype was most pronounced upon treatment with chemical agents that cause metaphase arrest (colchicine and paclitaxel). SIL is also necessary for the timely activation of CDK1 (CDC2), the major kinase regulating mitotic entry (13, 14). SIL may either directly regulate the activation of CDK1 or it may work further upstream. The rescue of the knockdown phenotype by a construct encoding murine Sil confirms the specificity of the knockdown and provides a useful system to further characterize the structural elements in SIL mediating these mitotic and apoptotic phenotypes.

We have recently shown that SIL is phosphorylated during mitosis on several conserved serine/threonine residues (5) necessary for its interactions with the mitotic regulator PIN1. Because the phosphorylation occurs during mitosis, it is probably not required for regulation of mitotic entry. This conclusion is strengthened by preliminary experiments which show that SIL mutated in the phosphorylation sites rescues the apoptosis phenotype of LS174T cells with knockdown of the endogenous SIL (Castiel and Izraeli; data not shown). Therefore, we believe that mitotic phosphorylation of SIL is not important for its pro-survival function at mitotic entry.

We have reported here our novel discovery that SIL functions in a “checkpoint” coupling the transition into mitosis with cell survival in variety of cancer cells. SIL is not necessary for the survival of all cells. Mouse embryonic stem cells lacking any functional Sil protein proliferate normally and create teratomas in nude mice (4). Although mitosis is a general physiologic process, cancer cells are highly sensitive to antimitotic drugs. Indeed the therapeutic ratio of drugs such as paclitaxel and vincristine is surprisingly high. Consequently, there is a marked effort to develop new drugs targeting molecules that regulate mitosis and mitotic entry (15, 16). As the SIL protein regulates mitotic entry and cell survival, it may prove to be a target for novel anticancer therapeutics.

Grant support: Israel Science Foundation, the Israel Cancer Research Foundation, and the US-Israel Binational Science Foundation. This work was done in partial fulfillment of the requirements for the Ph.D. degree of Ayelet Erez and Asher Castiel, Sackler Faculty of Medicine, Tel-Aviv University.

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 Hans Clevers and Reuvan Agami for providing reagents. The invaluable technical assistance by Inna Muler, Osnat Ashir-Fabian, Zamir Kohavi, Batia Plotkin, and other members of laboratories in the Sheba Cancer Research Center, Weizmann Institute, and Tel-Aviv University is highly appreciated. We thank Peter Aplan for fruitful discussions.