Elucidating the cross‐talk between inflammatory and cell proliferation pathways might provide important insights into the pathogenesis of inflammation‐induced cancer. Here, we show that the receptor‐interacting protein 1 (RIP1)—an essential mediator of inflammation‐induced nuclear factor‐κB (NF‐κB) activation—regulates p27Kip1 levels and cell‐cycle progression. RIP1 regulates p27Kip1 levels by an NF‐κB‐independent signal that involves activation of the phosphatidylinositol 3‐kinase (PI3K)–Akt–forkhead pathway. Mouse embryonic fibroblasts (MEFs) from RIP1‐knockout mice express high levels of p27Kip1. Reconstitution of MEFs with RIP1 downregulates p27Kip1 levels in a PI3K‐dependent manner. RIP1 regulates p27Kip1 at the messenger RNA level by regulating the p27Kip1 promoter through the forkhead transcription factors. RIP1 expression blocks accumulation of cells in G1 in response to serum starvation and favours cell‐cycle progression. Finally, we show that overexpression of p27Kip1 blocks the effects of RIP1 on the cell cycle. Thus, our study provides a new insight into how components of inflammatory and immune signalling pathways regulate cell‐cycle progression.
Inflammatory and immune responses induce cellular proliferation by various mechanisms, which include the release of cytokines and growth factors (Karin & Greten, 2005). Chronic inflammation has an important role in the development and progression of cancer, affecting various organs such as lung, colon, stomach and liver (Balkwill & Mantovani, 2001). The transcription factor nuclear factor‐κB (NF‐κB) is known to be a crucial component of the inflammation and might also have an important role in the pathogenesis of inflammation‐induced cancer (Karin, 2006). Cross‐talk between inflammatory and oncogenic signalling pathways might be important in cancer (Albanese et al, 2003; Hu et al, 2004; Chien et al, 2006; Schneider et al, 2006; Perkins, 2007).
The death domain‐containing kinase receptor‐interacting protein 1 (RIP1, RIPK1) is an essential component of the signalling cascade that leads to NF‐κB activation in response to cellular stress (Festjens et al, 2007). RIP1 is composed of a kinase, intermediate and death domain (Stanger et al, 1995), and is expressed both constitutively and inducibly in response to cytokines (Meylan & Tschopp, 2005).
The phosphatidylinositol 3‐kinase (PI3K)–Akt pathway has a crucial role in cell‐cycle progression and the regulation of p27Kip1. PI3K–Akt‐mediated regulation of p27Kip1 seems to occur at various levels and includes control of p27Kip1 expression as well as a decrease in protein stability (Liang & Slingerland, 2003). The forkhead/FoxO transcription factors regulate p27Kip1 transcription (Tran et al, 2003) and Akt has been reported to inhibit the transcription of p27Kip1 by inactivating the forkhead transcription factors (Medema et al, 2000).
Here, we show that RIP1 is important in the regulation of p27Kip1 and cell‐cycle progression. We show that RIP1 expression is sufficient to downregulate p27Kip1, phosphorylate retinoblastoma protein (Rb) and inhibit the accumulation of cells in G1 by activation of the PI3K–Akt–forkhead pathway. As RIP1 is known to have an important role in inflammatory responses, our findings suggest that RIP1 might participate in mechanisms involved in cellular proliferation during immune and inflammatory states.
RIP1 regulates expression of p27Kip1
We observed a higher level of p27Kip1 in RIP1−/− mouse embryonic fibroblasts (MEFs) compared with RIP1+/+ MEFs, both in cells growing in DMEM with 10% FBS and in those growing under serum‐starved conditions (Fig 1A,B). Next, by using a tetracycline‐inducible adenoviral vector (tet‐off), we introduced RIP1 into RIP1−/− MEFs that had been serum starved, either at the onset of serum starvation or 48 h after serum starvation. Introduction of RIP1 resulted in downregulation of p27Kip1 under both conditions (Fig 1C). Similarly, the introduction of RIP1 into many other cell lines such as U87MG (Fig 1D), T98G and Rat‐1 cells (data not shown) downregulated p27Kip1 levels, showing that RIP1 is a potent negative regulator of p27Kip1 levels. The effect of RIP1 on p27Kip1 was also detected in the presence of serum (supplementary Fig 5C online); however, RIP1 expression did not change the levels of two other cell‐cycle proteins, p21Cip1 and cyclin‐dependent kinase 4 (CDK4; Fig 1E,F).
p27Kip1 is an inhibitor of several cyclin‐dependent kinases, particularly cyclin E/CDK2, and one of the outcomes of p27Kip1 downregulation is phosphorylation of the Rb. Introduction of RIP1 into cells resulted in Rb phosphorylation (Fig 1G), as shown by a mobility shift in the Rb blot and by an increased signal on the pRb blot.
Next, we compared the levels of p27Kip1 messenger RNA in RIP1−/− MEFs with those in RIP−/− MEFs reconstituted with RIP (RIP1−/− MEFs+RIP1) by using northern blot and found that RIP1 downregulated p27Kip1 mRNA levels (Fig 1H) but not its protein stability (supplementary Fig 1 online).
RIP1 regulates p27Kip1 through PI3K
To investigate the signalling pathway used by RIP1 to downregulate p27Kip1, we introduced RIP1 into RIP1−/− MEFs and studied the effect of inhibition of either PI3K or MAP‐kinase kinase on RIP1‐induced p27Kip1 downregulation. Cells were serum starved for 72 h to induce p27Kip1. RIP1 was introduced into cells along with the inhibitors at the onset of serum starvation. RIP1 efficiently downregulated p27Kip1 levels under all conditions except when PI3K was inhibited with LY294002 (Fig 2A). PI3K inhibition did not have a significant effect on RIP1 expression. This suggests that RIP1 downregulates p27Kip1 through the PI3K pathway, whereas MEK inhibition (U0126) has no effect.
RIP1 activates the PI3K–Akt pathway
A previous study has shown that RIP1 is essential for lipopolysaccharide‐induced Akt activation by toll‐like receptor 4 (TLR4; Vivarelli et al, 2004); however, whether RIP1 has a direct effect on PI3K–Akt activation is not known. It was found that introduction of adenoviral RIP1 into RIP1−/− MEFs resulted in a robust phosphorylation of Akt as tested by phospho‐Akt (S473 and T308) antibodies (Fig 2B,C). As a control, we infected RIP−/− MEFs with green fluorescent protein or p27Kip1 in adenoviral vectors, but did not detect any Akt activation with either (Fig 2B). The level of Akt was not altered by RIP1. Ectopic expression of RIP1 into a variety of other cell types also resulted in Akt phosphorylation (supplementary Fig 2 online). We also confirmed RIP1‐induced PI3K–Akt activation by performing immune complex kinase assays (Fig 2D). The kinase domain of RIP1 was not required for Akt activation or p27Kip1 downregulation, as deletion of the kinase domain (DKD) did not impair either (Fig 2E). Use of a PI3K inhibitor (LY294002) resulted in a block of RIP1‐mediated Akt phosphorylation (Fig 2F). Finally, we found that introduction of RIP1 into cells also resulted in phosphorylation of cellular glycogen synthase kinase‐3 (GSK3; Fig 2G) and of p70 S6K, which is a widely used indicator of mTOR (mammalian target of rapamycin) activation (Fig 2H).
RIP1 influences the activity of the p27Kip1 promoter
From the observation that RIP1 regulates p27Kip1 through a PI3K‐dependent pathway and regulates p27Kip1 at the mRNA level, we examined the effect of RIP1 expression on the p27Kip1 promoter by using reporter assays. RIP1 strongly inhibited the activity of the p27Kip1 promoter linked to luciferase in a dose‐dependent manner (P<0.05; Fig 3A). Deletion of the death domain of RIP1 led to a significant loss in the ability of RIP1 to suppress the p27Kip1 promoter (P<0.001), whereas deletion of the kinase domain had the least effect (Fig 3B). Expression of the Flag‐tagged RIP1 mutants was detected by western blotting (supplementary Fig 3 online). In general, death domains act as protein interaction domains with other death domains, and a loss of the inhibitory effect of RIP1 on p27Kip1 in the death domain deletion mutant might reflect the need for interaction with another death domain protein.
RIP1 inhibits the forkhead transcription factors
We tested the ability of RIP1 to inhibit the activity of a forkhead‐Luc reporter (pGL3‐E4‐DBEx6) in 293 cells. As shown in Fig 3C, RIP1 produced a significant repression of forkhead activity (P<0.024). Furthermore, expression of a wild‐type FoxO3a or a constitutively active mutant FoxO3a rescued RIP1‐mediated repression of the p27Kip1 promoter (P<0.05; Fig 3D). Expression of the Myc‐tagged wild‐type and constitutively active mutant FoxO3a was confirmed by western blotting (supplementary Fig 4 online). Finally, mutating a FoxO response element in the p27Kip1 promoter resulted in a loss of the ability of RIP1 to inhibit the p27Kip1 promoter (P<0.05; Fig 3E). These findings suggest that RIP1 downregulates p27Kip1 by inhibition of the forkhead transcription factors.
NF‐κB is not required for downregulation of p27Kip1
LY294002 has only a minor inhibitory effect on RIP1‐induced activation of NF‐κB‐Luc and fails to block RIP1‐mediated downregulation of IκBα (supplementary Fig 5A,B online), arguing against a requirement for PI3K–Akt activation in RIP1‐induced NF‐κB activation. Next, we investigated whether RIP1 downregulates p27Kip1 by activation of NF‐κB. The introduction of RIP1 adenovirus into AT3 cells stably expressing the IκBα super‐repressor mutant that does not allow NF‐κB activation (Lin et al, 1998) failed to block RIP1‐mediated downregulation of p27Kip1 (supplementary Fig 5C online).
RIP1 regulates cell‐cycle progression
RIP1−/− MEFs were serum starved for 72 h followed by the addition of serum for 6 or 24 h, which was followed by cell‐cycle analysis. RIP1 was introduced into cells at the onset of serum starvation. As can be seen in Fig 4A (lower panel), introduction of RIP1 into cells prevented the reconstituted cells from arresting in G1 in response to serum starvation. Next, cells were serum starved for 48 h and then adenoviral RIP1 was introduced into cells. We found that RIP1 expression was sufficient to induce cell‐cycle progression in serum‐starved and growth‐arrested cells (Fig 4B). Furthermore, RIP1 expression increased 5‐bromodeoxyuridine (BrdU) incorporation in serum‐starved RIP−/− MEFs, confirming the effect of RIP1 on S‐phase progression (Fig 4F). Next, we showed that p27Kip1 overexpression inhibited the effect of RIP1 on the cell cycle (Fig 4C), suggesting that RIP1‐mediated cell‐cycle progression might be mediated by downregulation of p27Kip1. Expression of p27Kip1 and RIP1 is shown in Fig 4D. Finally, the addition of nocodazole, which arrests cells in the G2/M phase of the cell cycle, to cells grown under serum starvation and then exposed to RIP1, resulted in a greater accumulation of RIP1‐expressing cells in the G2/M phase of the cell cycle (Fig 4E) compared with control cells, again indicating the effect of RIP1 on the cell cycle.
JNK is not required for downregulation of p27Kip1
As RIP1 is known to activate Jun amino‐terminal kinases (JNKs), we investigated whether RIP1 can downregulate p27kip1 in JNK−/− fibroblasts. We confirmed that RIP1 activated JNK in RIP1−/− cells (Fig 5A); however, RIP1 efficiently downregulated p27Kip1 in both JNK1−/− and JNK2−/− fibroblasts (Fig 5B,D). Furthermore, inhibiting JNK with SP600125 failed to block both RIP1‐mediated downregulation of p27Kip1 and inhibition of FoxO activity (Fig 5E,G).
The main finding of this study is that RIP1, an essential component of NF‐κB activation pathways, regulates expression of p27Kip1 and cell‐cycle progression through an NF‐κB‐independent pathway involving PI3K–Akt–forkhead. We identified a signalling pathway triggered by RIP1 that led to cell‐cycle progression, as shown in the schematic diagram in supplementary Fig 6 online. We have presented genetic evidence that RIP1 regulates p27Kip1 levels. RIP1‐knockout MEFs express high levels of p27Kip1, and reconstitution of RIP1−/− cells with RIP1 results in a lowering of p27Kip1 levels. In addition, phosphorylation of Rb is increased in response to RIP1 expression. Thus, RIP1 influences crucial regulators of G1‐to‐S transition and blocks accumulation of cells in G1.
RIP1 regulates p27Kip1 mRNA levels by repressing the p27Kip1 promoter, and regulation of p27Kip1 by RIP1 is blocked by inhibition of PI3K. Expression of RIP1 is sufficient to induce a potent activation of the PI3K–Akt pathway; however, the kinase activity of RIP1 is not required for activation of PI3K–Akt, as an RIP1 mutant lacking the kinase domain activates Akt and downregulates p27Kip1. This is analogous to the lack of a requirement for the kinase activity of RIP1 in NF‐κB activation. It has been proposed that RIP1 acts as an adaptor in NF‐κB activation (Meylan & Tschopp, 2005), and we propose a similar mechanism for RIP1 in the activation of PI3K, with the death domain having a crucial role. It is known that Akt negatively regulates the expression of p27Kip1 by inactivation of forkhead transcription factors. We found that RIP1 suppressed the activity of forkhead transcription factors in reporter assays. Overexpression of a wild‐type FoxO3a or a constitutively active mutant FoxO3a inhibits the RIP‐mediated suppression of p27Kip1 transcription. Finally, mutation of a forkhead‐binding site in the p27Kip1 promoter abolishes the ability of RIP1 to downregulate p27Kip1. These experiments show that RIP1 negatively regulates p27Kip1 expression by activating a PI3K–Akt–forkhead pathway, whereas RIP1‐mediated JNK activation does not seem to be important for RIP1‐mediated p27Kip1 or FoxO regulation (Fig 5).
As RIP1 expression favours cell‐cycle progression, RIP1 could contribute to cellular proliferation during states of inflammation. From our recent work, we propose that RIP1 is an important node in the cross‐talk between inflammatory and growth factor signalling and cell‐cycle progression pathways. Here, we have shown that RIP1 activates the PI3K–Akt pathway and promotes cell‐cycle progression. However, consistent with the known complexity of inflammatory and NF‐κB signalling pathways, RIP1 might have antiproliferative/apoptotic or proliferative effects depending on the cellular context. For example, we have recently shown that RIP1 negatively regulates the expression of EGFR in fibroblasts (Ramnarain et al, 2008). Future studies will investigate the role of RIP1 in cancer.
Plasmids and cell lines. Wild‐type RIP and deletion mutants of RIP domains were cloned into pcDNA3.1 with a carboxy‐terminal Flag tag. Immortalized wild‐type and rip1−/− 3T3 cell lines were prepared according to established protocols. A forkhead response element linked to Luc (pGL3‐E4‐BBEx6), FoxO3a wild type and FoxO3aA3 were provided by Dr Diego Castrillon. A putative FoxO transcription factor‐binding site (GGTTTGTTG) around −683 relative to the translation start site was mutated with primer 5′‐GCAAGCCAGAGCAGGTGGGTTGGCAGCAGTACC‐3′ according to the manufacturer's protocol (QuikChangeMultiSite‐Directed Mutagenesis Kit, Stratagene, La Jolla, CA, USA).
Antibodies. The antibodies used in this study are listed in the supplementary information online.
Northern blot. RIP−/− MEFs or RIP−/− MEFs reconstituted with RIP for 16 h were used in this experiment and northern blot was conducted using standard protocols (supplementary information online).
Transfection. 293 cells were transfected using the calcium phosphate technique and expression of transfected genes was confirmed by western blot. For transient transfection experiments, cells were collected 24–48 h after transfection.
Luciferase assays. A total of 2 × 105 293 cells were transfected with either p27Kip1‐Luc or FHRE‐Luc (pGL3‐E4‐DBEx6) along with RIP plasmids or empty vector, or FoxO plasmids. A dual‐luciferase reporter assay system was used according to the instructions of the manufacturer (Promega, Madison, WI, USA). Firefly luciferase activity was measured in a luminometer and normalized on the basis of Renilla luciferase activity.
Production of adenovirus expressing RIP1. RIP1 wild type or DKD mutant with a deletion of the kinase domain (deletion of amino acids 1–303) was cloned into an adenoviral vector. This resulted in a Tet operon‐minimal CMV promoter‐driven cassette in place of the AdE1 region; Ad‐tTA (tetracycline‐controlled transactivator) was also prepared. A multiplicity of infection of 50 was used in the experiments. A p27Kip1 adenovirus was obtained from Vector Biolabs (Philadelphia, PA, USA). Cells were exposed to RIP1 adenovirus in the presence or absence of tetracycline in this Tet‐off system.
Cell‐cycle analysis. Cell‐cycle analysis was performed according to standard protocols. For nocodazole experiments, cells were serum starved for 48 h followed by the addition of RIP1 for 24 h. At 24 h after RIP1 addition, nocodazole (100 ng/ml) was added for a further 16 h followed by cell‐cycle analysis.
BrdU incorporation assay. RIP−/− cells were plated in a 96‐well plate. Cells were serum starved for 48 h followed by the addition of RIP1 for 24 h and later subjected to a BrdU incorporation assay using a kit obtained from EMD Biosciences (no. QIA58, Gibbstown, NJ, USA) according to the manufacturer's instructions. BrdU was added for 2 h before starting the assay.
Statistical analysis. Error bars represent the means±s.d. of three independent experiments. To determine the statistical significance between different samples, we used Student's t‐test. A P‐value of <0.05 was considered to be statistically significant.
Supplementary information is available at EMBO reports online (http://www.emboreports.org).
Conflict of Interest
The authors declare that they have no conflict of interest.
Supplementary Figs 1–6
We thank M. Kelliher for RIP‐knockout MEFs, M. Karin for JNK‐knockout MEF cell lines, D. Castrillon for FoxO plasmids and T. Sakai for p27Kip1 luciferase construct. This work was supported in part by the National Institutes of Health grant CA 78741 to A.A.H. and by funding from the Annette Strauss Center for Neuro‐oncology at the University of Texas Southwestern Medical Center.
- Copyright © 2008 European Molecular Biology Organization