STAT1 Promotes KRAS Colon Tumor Growth and Susceptibility to Pharmacological Inhibition of Translation Initiation Factor eIF4A

STAT1 plays a key role in innate immunity by protecting the host from infections with viruses and other pathogens (1). STAT1 acts downstream of type I (α/β) and II (γ) IFN receptors to mediate the expression of antiviral and immunoregulatory genes (1). DNA binding of STAT1 is induced by phosphorylation at tyrosine (Y) 701, whereas its phosphorylation at serine (S) 727 within the C-terminus domain promotes gene transactivation in response to IFN (1). Increased STAT1 Y701 phosphorylation has been observed in several forms of cancer, including multiple myeloma, erythroleukemia, and acute myelogenous leukemia as well as in breast and head and neck cancers (2, 3). In mice, STAT1 exhibits antitumor effects in response to carcinogens (4), which is explained by its ability to increase immune surveillance of tumors through IFNγ and increased natural killer (NK) activity (5). STAT1 mediates the inhibition of cell proliferation of both mouse and human tumor cells in culture treated with IFNγ via the increased expression of cyclin-dependent kinase (CDK) inhibitor P21^Cip1 or decreased c-myc expression (6, 7). Experiments in mouse cancer models have shown that STAT1 can function as a tumor suppressor via mechanisms that depend on the induction of antitumor immune responses as well as the inhibition of oncogenic signaling in a cell-autonomous (cell-specific) manner (8, 9). The antitumor properties of STAT1 have been best characterized in mouse models of breast cancer, in which STAT1 assumes both immunoregulatory and cell-autonomous functions to suppress tumor formation by either ErbB2/HER2 or estrogen receptor α (8, 10). Contrary to its antitumor effects in breast cancer, STAT1 promotes leukemogenesis in mice expressing v-ABL or TEL-JAK2 by increasing MHC class I expression and inhibiting tumor clearance by the immune system (11). Also, STAT1 protects different forms of solid tumors from death caused by chemotherapeutic drugs, an effect that is associated with increased expression of a specific subset of interferon-inducible genes that favor tumor survival in response to DNA damage (12, 13).

We previously demonstrated that STAT1 stimulates PI3K class IB signaling, leading to the induction of the AKT/protein kinase B–mTOR pathway in primary and immortalized mouse embryonic fibroblasts (MEF; ref. 14). Increased PI3K signaling by STAT1 was responsible for the degradation of programmed cell death protein 4 (PDCD4), which resulted in increased activity of the translation initiation factor eIF4A and efficient translation of mRNAs encoding for the prosurvival X-linked inhibitor of apoptosis (XIAP) and B-cell lymphoma xl (BCL-XL; ref. 14). We showed that the ability of STAT1 to promote XIAP and BCL-XL expression at the translational level significantly contributed to the survival of immortalized and HRAS G12V-transformed MEFs to treatments with PI3K and mTOR inhibitors or genotoxic drugs like doxorubicin (14).

Considering that the translational effects of STAT1 on cell survival were characterized in MEFs transformed by HRAS G12V overexpression (14), we investigated STAT1′s function in human colorectal carcinoma HCT116 cells bearing an endogenous activated KRAS allele. We demonstrate that the translational and prosurvival effects of STAT1 are especially relevant for human colon tumor cells with an activated KRAS allele and not for isogenic tumors with normal KRAS. STAT1 acts in a cell-autonomous fashion to promote the expression of XIAP and BCL-XL at the translational level and facilitate the growth of human KRAS colon tumors in immunodeficient mice. The stimulatory effects of STAT1 on XIAP and BCL-XL expression are linked to PDCD4 downregulation in a manner that depends on activated KRAS. Moreover, we demonstrate that STAT1 is responsible for the sensitization of KRAS colon tumor cells to death from the pharmacologic inhibition of eIF4A in culture and mice. Our work reveals a previously unidentified property of STAT1 to promote KRAS colon tumor growth and the therapeutic implications of targeting the prosurvival function of STAT1 with inhibitors of mRNA translation initiation.

HCT116 and HK2-8 cells were obtained from Dr. S. Shirasawa (15). The KRAS status in these cells was authenticated by examining their differences in the activation of the MAPK pathway as described previously by our group (16). The cells were maintained in DMEM (Wisent) supplemented with 10% FCS (Wisent) and 100 U/mL penicillin–streptomycin (Wisent). Downregulation of STAT1 was performed by infection with pGIPZ lentiviruses bearing a shRNA for human STAT1 of the following sequence: 5′-TGCTGTTGACAGTGAGCGCAAGGAAGTAGTTCACAAAATATAGTGAAGCCACAGATGTATATTTTGTGAACTACTTCCTTTTGCCTACTGCCTCGGA-3′ (Open Biosystems). CRISPR/CAS9-mediated depletion of STAT1 was performed by the coexpression of gRNA (5′-GAGGTCATGAAAACGGATGGTGG-3′) and CAS9 from pST1374-NLS-flag-linker-Cas9 vector (17). The PDCD4 shRNA was expressed from psiRNA-hH1hygro G2 vector (18). Human IFNα and IFNγ was obtained from Biosource (Invitrogen), whereas treatment with FL3 was performed as described previously (19).

Polysome profile analysis was performed as described previously (20). Total RNA or RNA from polysomal fractions was isolated by TRIzol (Invitrogen), and 1 μg of it was subjected to reverse transcription with 100 μmol/L oligo (dT) primer using the SuperScript III Reverse Transcriptase Kit (Invitrogen) according to the manufacturer's instructions. The synthesized cDNAs from the polysome fractions were amplified with 2.5 units of Taq DNA polymerase (GenScript) by PCR as follows: 94°C 1 minute; 57°C 1 minute; 70°C 1 minute × 28 cycles. The sequences of the primers were as follows: human XIAP forward primer 5′-TGGCAGATTATGAAGCACGGA-3′, reverse primer 5′-GGTCTTCACTGGGCTTCCAA-3′; human BCL-XL forward primer 5′-GATGGGGTAAACTGGGGTCG-3′, reverse primer CACAAAAGTATCCCAGCCGC-3′; human GAPDH forward primer 5′-CCTCCCGCTTCGCTCTCT-3′, reverse primer 5′-CCGTTGACTCCGACCTTCAC-3′.

Transfections of di-cistronic reporter gene constructs containing the XIAP or BCL-Xl 5′ untranslated region (UTR) were described previously (14). Quantification of β-galactosidase and chloramphenicol acetyltransferase (CAT) activity was performed as described previously (21).

Protein extraction and immunoblot analysis were performed as described previously (14). Anti-Stat1α p91(C-111) and Stat1(M-23) antibodies were purchased from Santa Cruz Biotechnology: pT308-AKT, pS473-AKT, AKT, pT389 S6K, S6K, PDCD4, P85, P110α, P110γ, P101, P87 antibodies were from Cell Signaling Technology; anti-actin (C4) was from Biosource International; and horseradish peroxidase (HRP)–conjugated anti-mouse IgG antibody and HRP anti-rabbit IgG antibody were from Amersham Pharmacia Biotech. Quantification of bands in linear range of exposure was performed by densitometry using Scion image software.

Detection of cell death by propidium iodide and flow cytometry analysis was performed as described previously (14).

Injection of 1 × 10⁶ cells per site in 8-week-old female athymic nude mice (Charles River Inc.) and tumor monitoring were performed as described previously (14, 22). Each mouse received two subcutaneous injections of 1 × 10⁶ cells. Tumor growth was measured with digital calipers three times per week for the indicated periods, and the volume was calculated by the formula: tumor volume (mm³) = 1/2 × [length (mm)] × [width (mm)]². Mice were treated with 15 mg/kg body weight of FL3 delivered by intraperitoneal injections every day for 2 weeks based on the established protocol (19). The animal studies were performed in accordance with approved protocols and regulations by the Animal Welfare Committee of McGill University (Montreal, Quebec, Canada; protocol #5754).

Experiments were performed in triplicates, and data represent the average of three independent experiments. Error bars represent SE as indicated, and significance in differences between arrays of data tested was determined using two-tailed Student t test.

We investigated the effects of STAT1 on the expression of the catalytic and regulatory subunits of PI3K in the human colorectal carcinoma HCT116 cells containing a KRAS G13D mutant allele (15). Downregulation of STAT1 by shRNA decreased the expression of the catalytic P110γ subunit but had no inhibitory effects on the expression of the regulatory P101 and P87 subunits of PI3K class IB (Fig. 1A). On the other hand, STAT1 downregulation did not elicit significant effects on the expression of the catalytic P110α and regulatory P85 subunit of PI3K class IA (Fig. 1A). These data showed that STAT1 maintains its ability to increase P110γ expression in human KRAS tumor cells as previously shown by our group for HRAS-transformed MEFs (14).

The ability of STAT1 to increase P110γ expression contributes to PI3K class IB signaling, leading to the activation of AKT and mTOR in HRAS-transformed MEFs (14). We observed that STAT1 deficiency by shRNA expression in HCT116 cells was associated with decreased phosphorylation of AKT at T308 and S473 in response to serum stimulation (Fig. 1B). Moreover, ribosomal S6 kinase 1 (S6K1) phosphorylation at T389, which is mediated by mTOR complex 1 (mTORC1), was decreased in STAT1-deficient compared with proficient HCT116 cells in response to serum treatment (Fig. 1B). These findings supported a stimulatory role of STAT1 in PI3K–AKT–mTOR signaling in HCT116 cells.

We further observed that STAT1 downregulation resulted in increased expression of PDCD4 in HCT116 cells (Fig. 1B). PDCD4 is negatively regulated by AKT and mTOR (23), and as such, decreased PI3K class IB signaling by the loss of STAT1 contributes to PDCD4 stabilization (14). Also, PDCD4 exhibits tumor-suppressive properties by functioning as an inhibitor of the RNA helicase activity of the eukaryotic translation initiation factor eIF4A (24–26). Our previous work demonstrated that STAT1 promotes the degradation of PDCD4 in immortalized and HRAS-transformed MEFs, which in turn increases eIF4A activity and enhances the expression of BCL-XL and XIAP at the translational level (14). In line with those findings, serum stimulation of HCT116 cells led to PDCD4 downregulation, which was prevented by the shRNA-mediated inhibition of STAT1 (Fig. 1B). In addition, serum treatment resulted in the upregulation of XIAP in STAT1-proficient but not in STAT1-deficient HCT116 cells (Fig. 1B). On the other hand, BCL-XL expression was significantly increased in STAT1-proficient compared with deficient cells in spite of BCL-XL downregulation in the former cells by serum treatment (Fig. 1B). These findings were consistent with the stimulatory effects of STAT1 on BCL-XL and XIAP expression in HRAS-transformed MEFs (14) and provided a link between PDCD4 downregulation and increased expression of the prosurvival proteins in KRAS colon tumor cells (Fig. 1B).

We next examined the signaling properties of STAT1 in HK2-8 cells derived from HCT116 cells in which the KRAS G13D allele was replaced with normal KRAS by homologous recombination (15). Downregulation of STAT1 in HK2-8 cells by shRNA expression had no effect on mTORC1 activity as indicated by the induction of S6K1 T389 phosphorylation in response to serum stimulation (Fig. 1C). Also, STAT1 downregulation in HK2-8 cells did not yield significant differences in PDCD4, XIAP, or BCL-XL levels compared with STAT1-proficient HK2-8 cells (Fig. 1C). These data suggested that the signaling effects of STAT1 on the expression of PDCD4 and prosurvival proteins are rather specific for colon tumor cells with activated KRAS.

Considering that PDCD4 downregulation by STAT1 stimulates the translation of XIAP and BCL-XL mRNAs in HRAS-transformed MEFs (14), we determined the translational effects of STAT1 on XIAP and BCL-XL mRNAs in HCT116 cells in polysome profile assays. This methodology relies on sucrose gradient centrifugation to distinguish between efficiently translated mRNAs associated with polyribosomes from poorly translated mRNAs bound to monosomes (i.e., 40S) or disomes (i.e., 80S; Fig. 2A). We observed that XIAP and BCL-Xl mRNAs were more efficiently associated with polyribosomes in STAT1-proficient than deficient HCT116 cells (Fig. 2A). On the other hand, mRNAs encoding for the antiapoptotic myeloid cell leukemia 1 (MCL-1) or the housekeeping GAPDH were better bound to polyribosomes of STAT1-deficient than proficient HCT116 cells (Fig. 2A). These data implied a preferential effect of STAT1 on the translation of XIAP and BCL-XL mRNAs in HCT116 cells (Fig. 2A).

To further support the translational effects of STAT1 in HCT116 cells, we assessed the translational properties of the 5′ UTR of XIAP and BCL-XL mRNAs, both of which contain an internal ribosome entry site (IRES; ref. 27). To this end, we performed transient transfection assays of di-cistronic constructs bearing the 5′ UTR of either BCL-XL or XIAP mRNA between the β-galactosidase and CAT reporter genes. These di-cistronic constructs were previously demonstrated to be devoid of splice-acceptor sites involved in spurious IRES activity (21). The di-cistronic reporter assays revealed that IRES activity of each mRNA was more highly increased in STAT1-proficient than deficient cells (Fig. 2B and C). These data supported a role of STAT1 in the stimulation of cap-independent translation of the antiapoptotic mRNAs in HCT116 cells.

Downregulation of PDCD4 in STAT1-proficient HCT116 cells provided the rationale of the hypothesis that STAT1 sensitizes HCT116 cells to the antiproliferative effects of eIF4A inhibition. Consistent with this hypothesis, we observed that PDCD4-deficient HCT116 cells by shRNA expression were more susceptible than PDCD4-proficient cells to death after treatment with the flavagline FL3 (Fig. 3), which is a specific and potent eIF4A inhibitor (19, 28). We also found that treatment with FL3 substantially increased the susceptibility of STAT1-proficient HCT116 cells to death compared with STAT1-deficient HCT116 cells (Fig. 4A). Moreover, FL3 treatment resulted in a higher downregulation of XIAP and BCL-XL proteins in STAT1-proficient than deficient HCT116 cells, consistent with the increased susceptibility of the former cells to death (Fig. 4B). The antiapoptotic MCL1 was partially downregulated in STAT1-proficient compared with STAT1-deficient HCT116 cells after 24 hours of FL3 treatment, whereas such an effect was not observed after 48 hours of treatment (Fig. 4B). MCL-1 mRNA is translated via eIF4E-sensitive and cap-dependent mechanisms (29), whereas its translation was not facilitated by STAT1 signaling in HCT116 cells (Fig. 2A). As such, the inability of FL3 to exhibit a strong and lasting inhibition of MCL-1 expression in the presence of STAT1 implied a possible higher specificity of the drug for cap-independent rather than cap-dependent mRNA translation in HCT116 cells (Fig. 4B). Di-cistronic reporter assays supported that the inhibitory effects of FL3 are exerted on ISRE-driven translation within the 5′ UTR of either XIAP or BCL-XL mRNA in STAT1-proficient but not STAT1-deficient HCT116 cells (Fig. 4C and D).

Contrary to HCT116 cells, STAT1 deficiency increased the susceptibility of HK2-8 cells to death in response to FL3 treatment (Fig. 4E). Also, FL3 treatment caused a similar inhibitory effect on XIAP expression in STAT1-proficient and deficient HK2-8 cells but displayed no effect on BCL-XL and MCL-1 expression in both cell types (Fig. 4F). The inability of STAT1 to sensitize HK2-8 cells to eIF4A inhibition was in line with the initial observation that STAT1 cannot signal to PDCD4 downregulation and increased XIAP and BCL-XL expression in these cells (Fig. 1C).

We next examined whether the prosurvival properties of STAT1 are implicated in HCT116 tumor growth in immunodeficient mice. To this end, we employed HCT116 cells impaired for STAT1 by the clustered regularly interspaced short palindromic repeats (CRISPR)/CAS9 (herein referred to as STAT1^C/C cells). Immunoblot analyses confirmed the depletion of STAT1 in HCT116 cells in the absence as well as presence of either IFNα or IFNγ, both of which upregulate STAT1 (Fig. 5A). We noticed that downregulation of PDCD4 was further enhanced by IFN treatment in STAT1-proficient but not in STAT1-deficient HCT116 cells, establishing a causative relationship between STAT1 and PDCD4 downregulation in the absence as well as presence of IFN signaling (Fig. 5A).

When the tumorigenic properties of HCT116 cells were tested in athymic nude mice (Nu/Nu) that were deficient for T cells, we observed that STAT1^C/C cells exhibited substantially decreased rates of tumor growth compared with STAT1-proficient cells (Fig. 5B). Identical results were obtained when STAT1^C/C and STAT1-proficient HCT116 cells were transplanted in SCID congenic mice that were deficient in T and B cells as well as in SCID beige mice deficient in T, B, and NK cells (Fig. 5C and D). These findings suggested that STAT1 acts in a cell-autonomous manner to promote the tumorigenic potency of HCT116 cells. The effects of STAT1 on the tumorigenic properties of HK2-8 cells could not be tested due to the inability of HK2-8 cells to form tumors in immunodeficient mice (15). When the antitumor effects of FL3 were tested in nude, we observed that the growth of STAT1-proficient tumors was significantly decreased as opposed to the growth of STAT1-deficient tumors, which was refractory to FL3 treatment (Fig. 6A and B). These data supported the ability of STAT1 to sensitize HCT116 cells to the antitumor effects of the pharmacologic inhibition of eIF4A in vivo.

Our study demonstrates the ability of STAT1 to increase PI3K–AKT–mTOR signaling and promote the survival of KRAS colon tumors (Fig. 6C). The prosurvival properties of STAT1 are mediated, at least in part, by the downregulation of PDCD4 and increased eIF4A activity, leading to increased translation of XIAP and BCL-XL mRNAs. This is in line with other studies demonstrating the important role of PDCD4 in the translational repression of XIAP and BCL-XL mRNAs in different human tumors via cap-independent mechanisms (30, 31). The translational effects of STAT1 occur in HCT116 cells but not in HK2-8 cells, suggesting their dependency on activated KRAS. Also, the translational effects of STAT1 contribute to tumor survival inasmuch restoration of PDCD4 activity by the pharmacologic inhibition of eIF4A induces KRAS colon tumor cell death in culture and inhibits KRAS tumor growth in immunodeficient mice.

STAT1 can exhibit antitumor effects via cell-autonomous and immunoregulatory pathways, as has been best demonstrated for breast cancer (8). However, in human KRAS colon tumor cells, STAT1 can act in a cell-autonomous manner to promote tumor growth. Such differences in STAT1 function may be context dependent and determined by variations in oncogenic lesions and pathways in different forms of cancer. Previous work from our group showed that STAT1 inhibits the tumorigenic potency of HRAS G12V-overexpressing MEFs in nude mice (32), an effect that is different from the ability of STAT1 to promote KRAS-mediated colon tumor growth. Such disparities in STAT1 function can be attributed to differences in the functional specificities of HRAS and KRAS proteins in different tissues (33). Another conceivable explanation is variations in the expression of activated RAS proteins given that the antitumor effects of STAT1 in MEFs were associated with HRAS overexpression, whereas the protumorigenic effects of STAT1 in colon tumors are linked to endogenous expression of activated KRAS. In this regard, overexpression of activated HRAS and hyperactivation of the MAPK pathway has been implicated in the induction of tumor suppressor pathways, leading to increased senescence or apoptosis (34, 35).

Considering the therapeutic implications of our findings, targeting the prosurvival pathways of STAT1 may be a suitable means to treat KRAS tumors. In addition to FL3, potent inhibitors of eIF4A activity, such as silvestrol or hippuristanol, have been considered for the treatment of various forms of cancer (36). However, strategies aimed at the inactivation of the cell-autonomous functions of STAT1 may elicit undesirable effects by compromising immune-mediated responses in the tumor bed that require STAT1 activation (37). The immunoregulatory effects of eIF4A inhibition on STAT1 signaling could not be tested in our study given the limitations from monitoring human tumor growth in immunocompromised mice. Thus, to better understand the therapeutic benefits of the pharmacologic inhibition of eIF4A, it is important to determine whether the translational effects of STAT1 are implicated in the induction of immune-mediated antitumor responses in the microenvironment of KRAS tumors and possibly other tumor types that are sensitive to eIF4A inhibition.

No potential conflicts of interest were disclosed.

Conception and design: S. Wang, A.E. Koromilas

Development of methodology: S. Wang, C. Darini

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Wang, C. Darini, L. Désaubry

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Wang, C. Darini, A.E. Koromilas

Writing, review, and/or revision of the manuscript: S. Wang, A.E. Koromilas

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Wang

Study supervision: A.E. Koromilas

We thank S. Shirasawa for HCT116 and HK2-8 cells, M. Holcik for XIAP and BCL-XL IRES di-cistronic constructs, and B. Lankat-Buttgereit for human PDCD4 shRNA vector.

The work was supported by funds from the Cancer Research Society Inc., the Marjorie Sheridan Innovation grant of the Canadian Cancer Society (CCSRI #701631), and Quebec Breast Cancer Foundation (QBCF; to A.E. Koromilas).

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.