Cell death/proliferation roles for nc886, a non-coding RNA, in the protein kinase R pathway in cholangiocarcinoma

Abstract

We have recently identified nc886 (pre-miR-886 or vtRNA2-1) as a novel type of non-coding RNA that inhibits activation of protein kinase R (PKR). PKR's pro-apoptotic role through eukaryotic initiation factor 2 α (eIF2α) phosphorylation is well established in the host defense against viral infection. Paradoxically, some cancer patients have elevated PKR activity; however, its cause and consequence are not understood. Initially, we evaluated the expression of nc886, PKR and eIF2α in non-malignant cholangiocyte and cholangiocarcinoma (CCA) cells. nc886 is repressed in CCA cells and this repression is the cause of PKR's activation therein. nc886 alone is necessary and sufficient for suppression of PKR via direct physical interaction. Consistently, artificial suppression of nc886 in cholangiocyte cells activates the canonical PKR/eIF2α cell death pathway, suggesting a potential significance of the nc886 suppression and the consequent PKR activation in eliminating pre-malignant cells during tumorigenesis. In comparison, active PKR in CCA cells does not induce phospho-eIF2α nor apoptosis, but promotes the pro-survival nuclear factor-κB pathway. Thus, PKR has a dual life or death role during tumorigenesis. Similarly to the CCA cell lines, nc886 tends to be decreased but PKR tends to be activated in our clinical samples from CCA patients. Collectively from our data, we propose a tumor surveillance model for nc886's role in the PKR pathway during tumorigenesis.

Figure 1. Expression of nc886, vtRNA1-1, PKR, and eIF2α in cholangiocyte and CCA cells

Top three panels: Northern hybridization of n886 and vtRNA1-1, with EtBr (ethidium bromide) staining of total RNA shown for equal loading. Bottom five panels: Western blot of PKR and eIF2α, with β-actin as a loading control. The molecular size markers in kilodalton (kD) are indicated on the right.

Figure 2. nc886 suppressed PKR in CCA cells

A. Western blot of p58IPK. The equal loading is shown in Fig 1. The expression profile of nc886, P-PKR, and P-eIF2α from Fig 1 is summarized and indicated on the bottom. All other descriptions are the same as in Fig 1. B. Northern hybridization of nc886 and Western blot of indicated proteins. MMNK1 cells were harvested for sample preparation, at 48 hrs after two rounds of transfection (at 0 and 24 hrs) of each indicated anti-oligo at 50 nM. All other descriptions are the same as in Fig 1. C. PKR pulldown assays to measure PKR/nc886 association. M156 and M214 cells were harvested at 48 hrs after one round of transfection of each indicated anti-oligo at 50 nM. Cell lysates were subjected to PKR immunoprecipitation, as described in Materials and Methods. RNAs from immunoprecipitates (designated as “IP” above the gel) and from cell lysates (designated as “input”) were subjected to RT-PCR measurement of nc886, vtRNA1-1 and β-actin mRNA. For comparison, genomic DNAs from each cell line (designated as “gen. DNA”) were included in PCR reactions. D. Western blot of indicated proteins in M156 and M214 cell lysates. Transfection conditions were the same as in panel C. All other descriptions are the same as in panel B. E. Autoradiogram of in vitro PKR kinase assays using MMNK1 lysates, without (lane 1) or with 50 nM of each indicated anti-oligo (lane 2-3). PKR immunoprecipitates were pre-incubated with anti-oligos for 2 hrs, before adding [γ-³²P]-ATP to initiate the kinase reaction. F. In vitro PKR kinase assays of M213 lysates, with indicated amounts of synthetic biotinylated nc886 or yeast tRNA which were incubated with PKR immunoprecipitates for 10 min before the kinase reaction. G. EMSA gels with indicated amounts of PKR(dsRBM1-2). Identities of the bands are indicated on the right.

Figure 3. nc886 depletion provoked the canonical PKR/eIF2α pathway leading to apoptosis in cholangiocyte MMNK1 cells, but not in CCA cells

A. The same MMNK1 samples used in Fig 2B were subjected to eIF2α Western blot (upper panels) and global protein synthesis assays (lower panels). In lower panels, 10 μg of total protein was loaded onto a 10% SDS-polyacrylamide gel, and visualized by Coomassie blue staining (for total protein) and by autoradiogram (for [³⁵S]-methionine incorporation). Incorporated amounts of [³⁵S]-methionine were quantified by scintillation counting and indicated under each lane. B. Cell proliferation assays (MTT assays) of MMNK1 cells after the same treatment in Fig 2B and panel A. Mock transfection without an anti-oligo (designated as “no anti-oligo”) was included as another negative control in addition to an anti-oligo against vtRNA to rule out any effect of vtRNA depletion on cell proliferation. All experiments were performed in triplicate from which an average and a standard deviation were calculated. C. Western blot of apoptotic proteins in the indicated cell lines. Transfection conditions for each cell line are as described in earlier figures (Fig 2B for MMNK1, Fig 2C for M156 and M214) and all other descriptions are same as in Fig 2B. D. eIF2α Western blot (upper panels) and global protein synthesis assays of M156 and M214 cells. All other descriptions are same as in panel A and C.

Figure 4. P-PKR induced by dsRNA activated the NF-κB branch but not the eIF2α branch in CCA cells

A. Western blot of indicated proteins (upper panels) and global protein synthesis assays (lower panels). Cells were harvested at 24 hrs after transfecting 0.2 μg/ml of Poly(I:C) or yeast tRNA. All other descriptions are same as in Fig 3A. B. The NF-κB activity was measured by luciferase assays. Assays were performed at 24 hrs after transfection of luciferase plasmids combined with 0.2 μg/ml of Poly(I:C) or yeast tRNA. At the time of transfection, 2-AP (2-aminopurine in ethanol) was added to indicated samples at a final concentration of 2.5 mM and the corresponding volume of ethanol (vehicle) was added to the other samples. The relative luciferase values (y-axis) were calculated through several normalizations. First, firefly luciferase (Pp) values from pNF-κB-Luc were normalized to Renilla luciferase (Rr) values from co-transfected pRL-SV40. The Pp/Rr values of the NF-κB reporter plasmid were subjected to the second normalization to Pp/Rr values of the CMV promoter from pcDNA3.1-Zeo(+)-Pp, yielding relative Pp/Rr values (y-axis). The values of yeast tRNA were set as 1. For each sample, an average and a standard deviation were calculated from triplicate reactions.

Figure 5. NF-κB was elevated by P-PKR in CCA cells and played a pro-proliferative role therein

A. In vitro PKR kinase assays (top panel) of indicated cell lines. Aliquots of immunoprecipitates were subjected to PKR Western blot (using antibody against total PKR) to confirm comparable pull-down efficiencies among cell lines (bottom panel). B. NF-κB activity measured by luciferase assays. The relative Pp/Rr value of MMNK1 was set as 1. All other descriptions are the same as Fig 4B. C. qRT-PCR measurement of NF-κB target genes (IL-8, A20, and IκBα) and 18S rRNA as a control for equal RNA quantities. Ct (cycle threshold) values were converted to “relative cDNA amounts” by a regression line of titrating amounts of cDNA (a pool of cDNA from MMNK1, M214, and M213) and their Ct values. An average and a standard deviation were calculated from triplicate samples. For each mRNA, the value of MMNK1 was set as 1. D. qRT-PCR measurement of NF-κB target genes (A20, IL-6, and IL-8). M156 and M214 cells were harvested for RNA preparation, at 48 hrs after transfection of each indicated anti-oligo and/or siRNA at 50 nM. Data processing was done as described in panel C, except that values of each mRNA (“relative cDNA amounts”) were normalized to values of 18S rRNA (x-axis) and normalized values of “no transfection” was set as 1 for each mRNA. E. Cell proliferation (MTT) assays at 24 hrs after treating CAPE, a chemical inhibitor for NF-κB. An average and a standard deviation from triplicate samples are shown. F. BrdU incorporation assays were done after treating 2-AP, a chemical inhibitor for PKR. All other descriptions are the same as panel E.

Figure 6. The intrinsic expression of P-eIF2α, which was independent of P-PKR, was bypassed by eIF2B overexpression in CCA cells

A. Western blot of indicated proteins after PKR knockdown. Cells were harvested at 48 hrs after transfection of indicated siRNAs at 50 nM. siGFP (targeting 5'-GGCTACGTCCAGGAGCGCA-3') was included as a negative control. B. Western blot of eIF2B-ε, β-actin, and MetAP2(p67). All descriptions are the same as Fig 2A.

Figure 7. nc886/P-PKR expression in clinical specimens from CCA patients

A. qRT-PCR measurement of nc886 was performed in 16 pairs of cancerous and non-cancerous portion of CCA liver tissues (samples KK#1-#16 from Khon Khan University). Ct values were converted to “equivalent ng of genomic DNA” by a regression line of titrating amounts of genomic DNA and their Ct values. The converted values of nc886 were normalized to those of snU6 (y-axis). An average and a standard deviation were calculated from triplicate reactions. Samples were sorted into three groups (increased, not changed, and decreased), with a cutoff p value < 0.05. B. IHC of P-PKR. A representative pair (MDA patient #3) is shown among eight samples from MD Anderson Cancer Center (MDA samples #1-8 in panel C). C. The IHC signal of P-PKR was scored in a semi-quantitative fashion based on the staining intensity; +3 (strong), +2 (moderate) or +1 (weak). In case of five pairs (patient # 1, 2, 3, 5, 6), bile duct areas of matched normal tissues were used as “matched normal”. In the other three cases (patient # 4, 7, 8), matched normal bile duct tissues were unavailable and thus normal liver tissues from the same patient were used as “matched normal”.

Figure 8. Summary of data depicting the tumor surveillance model

Cell lines are in brackets. Black arrows: active pathways; grey arrows: inactivated pathways.