Systems Biology Strategy Reveals PKCδ is Key for Sensitizing TRAIL-Resistant Human Fibrosarcoma

Cancer cells are highly variable and largely resistant to therapeutic intervention. Recently, the use of the tumor necrosis factor related apoptosis-inducing ligand (TRAIL) induced treatment is gaining momentum due to TRAIL’s ability to specifically target cancers with limited effect on normal cells. Nevertheless, several malignant cancer types still remain non-sensitive to TRAIL. Previously, we developed a dynamic computational model, based on perturbation-response differential equations approach, and predicted protein kinase C (PKC) as the most effective target, with over 95% capacity to kill human fibrosarcoma (HT1080) in TRAIL stimulation (1). Here, to validate the model prediction, which has significant implications for cancer treatment, we conducted experiments on two TRAIL-resistant cancer cell lines (HT1080 and HT29). Using PKC inhibitor bisindolylmaleimide I, we demonstrated that cell viability is significantly impaired with over 95% death of both cancer types, in consistency with our previous model. Next, we measured caspase-3, Poly (ADP-ribose) polymerase (PARP), p38, and JNK activations in HT1080, and confirmed cell death occurs through apoptosis with significant increment in caspase-3 and PARP activations. Finally, to identify a crucial PKC isoform, from 10 known members, we analyzed each isoform mRNA expressions in HT1080 cells and shortlisted the highest 4 for further siRNA knock-down (KD) experiments. From these KDs, PKCδ produced the most cancer cell death in conjunction with TRAIL. Overall, our approach combining model predictions with experimental validation holds promise for systems biology based cancer therapy.


INTRODUCTION
Numerous recent studies have revealed the close link between inflammation and cancer. First, various types of immune cells, which support tumor growth progression, are found within the tumor microenvironment (2,3). Second, the vicinity of cancer cells displays increased proinflammatory activity, through the detection of elevated levels of major cytokines such as the tumor necrosis factor (TNF) (4,5). One notable cytokine found within the tumor microenvironment is the TNF related apoptosis-inducing ligand or TRAIL, which has been shown to induce apoptosis in certain types of malignant cancers with no significant effect on normal cells (6,7). The findings have led to a major stride in the ongoing research aimed at optimizing TRAIL-induced cancer therapy (8,9). Despite some success, TRAIL-based therapies still show dismal results for several types of cancers such as the breast cancer, neuroblastoma, adenocarcinoma, and glioma (10)(11)(12)(13).
Here, we tested the model prediction by experimentally verifying whether targeting PKC will enhance apoptosis in TRAILresistant cancer cell lines. Experiments were performed on TRAILinduced human fibrosarcoma (HT1080) and human colon adenocarcinoma (HT29) cells, and the cell viability was compared with control normal fibroblasts (TIG-1 and MRC-5). Moreover, to investigate the intracellular mechanisms for resultant cell viability, we measured time-course activation levels of caspase-3, PARP, p38, and JNK. Subsequently, we analyzed the expressions of each PKC isoform member in HT1080 cells. To identify a crucial target member for enhanced cancer apoptosis, we prepared relevant siRNA KD experiments. In summary, our study investigates (i) whether the model prediction of PKC suppression will enhance cancer cell death is true, and (ii) whether computational modeling using perturbation-response approach is valuable for biological research focusing on cancer treatment.

CELL VIABILITY ASSAY
The cell viability was measured by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay and trypan blue exclusion. MTT assay: cells (10 × 10 4 ) were inoculated in each well and incubated for 24 h. Thereafter, 50 µL of MTT (2 mg/mL in PBS) was added to each well and the plates were incubated for a further 2 h. The resultant formazan was dissolved with 100 µL of dimethyl sulfoxide (DMSO) after aspiration of culture medium. Plates were placed on a plate shaker for 1 min and then read immediately at 570 nm using TECAN microplate reader with Magellan software (Männedorf, Switzerland). Trypan blue exclusion: cells were detached with 1 mL of trypsin and suspended in DMEM. After staining with trypan blue, viable cells were counted using microscopy (n = 3). The percentage of trypan blue exclusive viable cells was determined as a percentage of the total number of cells.

WESTERN BLOT ANALYSIS
Anti-PARP, anti-phospho-p38, and anti-β-actin antibodies were purchased from Cell Signaling Technology. Proteins were extracted from the cell lines using radioimmunoprecipitation assay (RIPA) buffer according to the manufacturer's instructions. Next, their concentrations were measured by Bradford protein assay. Equal amounts of protein were loaded in each well and separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), which was subsequently transferred onto a polyvinylidene difluoride (PVDF) membrane. The membrane was blocked for 1 h with 5% BSA in TBST on the shaker at room temperature. The membrane was placed on PARP and p-p38 antibody diluted at a 1:1000 proportion in diluent buffer [5% (w/v) BSA and 0.1% Tween 20 in TBS] and incubated overnight at 4°C on the shaker. The membrane was washed three times in TBS as above and incubated with secondary antibody diluted at a 1:10000 proportion for 1 h on the shaker at room temperature. The membrane was again washed three times for 5 min each time as above and finally the results were generated by using an enhanced chemiluminescence (ECL) Western blotting kit.

QUANTITATIVE REAL-TIME PCR ANALYSIS
Total cellular RNA was extracted from cells using the TRIzol reagent according to the manufacturer's instructions (Invitrogen). One microgram of RNA was reverse-transcribed using a firststrand cDNA synthesis kit (ReverTra Aceα; Toyobo). Quantitative real-time PCR (qRT-PCR) was performed using SYBR premix Ex Taq (Takara) on the Applied Biosystems StepOnePlus™ according to the technical brochure of the company. qRT-PCR primers used in this study are listed in Table 1. Quantitative measurements were determined using the ∆∆Ct method and expressions of GAPDH gene for PKC genes and RPL27 gene for rela, mtor, bcl2, bax, cytoc, and jun were used as the internal control. Melt curve analyses of all qRT-PCR products were performed and shown to produce the sole DNA duplex.

EFFECT OF PKC INHIBITOR IN TRAIL-RESISTANT HT1080 CELLS
Based on our previous computational TRAIL model, the removal of PKC family members would enhance HT1080 cell death by 95% (1). Here, we investigated the actual experimental effect of PKC inhibition to HT1080 cells in TRAIL stimulation. HT1080 cells were stimulated with 1000 ng/mL of TRAIL in the presence or absence of 10 µM of PKC inhibitor (31)(32)(33), BIM-I, pre-treatment and compared with unstimulated control with and without BIM-I pre-treatment ( Figure 1A). We observed, phenotypically, that HT1080 cell death was significantly increased in combinatorial treatment of TRAIL and BIM-I ( Figure 1A, forth column), while control pre-treated with BIM-I did not induce any noticeable cell death ( Figure 1A, second column).
Next, we investigated cell survival ratio using MTT assays for HT1080 cells pre-treated with BIM-I with increasing dosage (0, 3, and 10 µM) for 30 min prior to increasing TRAIL stimulation (0, 100, 200, 400, and 1000 ng/mL) for 24 h ( Figure 1B). Notably, from these experiments, it is clear that HT1080 cell death is almost unaffected with any dosage of BIM-I without TRAIL stimulation. However, when BIM-I was treated in the presence of TRAIL, the effect synergistically produced significant cell death, compared with TRAIL alone (Figure 1B). Remarkably, as predicted by our previous computational TRAIL model (1), the inhibition of PKC (with 10 µM of BIM-I) resulted in about 99% cell death for TRAIL stimulation (with 100 ng/mL or more) in HT1080 cells. We further investigated the cell viability of HT1080 with respect to stimulation time, and noticed that significant cell death occurs at 3 h and onward ( Figure 1C).
Next, in addition to HT1080, we also investigated another TRAIL-resistant cancer cell type (HT29) and compared with normal fibroblasts (TIG-1 and MRC-5). Experiment-matched MTT assays revealed that both HT1080 and HT29 cell cultures treated with BIM-I were sensitized to TRAIL-induced cell death (approximately 99 and 95% cell death, respectively), while normal TIG-1 Frontiers in Immunology | Tumor Immunity and MRC-5 largely survived (Figures 1D,E). These results indicate that PKC inhibitor, BIM-I, has specific ability to enhance cell death in TRAIL-resistant cancer cells while having little effect on normal cells.

TREATMENT OF PKC INHIBITOR WITH TRAIL ENHANCES CELL DEATH THROUGH APOPTOSIS
The experimental results, so far, are consistent with our previous model simulations. To further scrutinize the result, that is, to explore the origins of cell death, we performed analysis to observe intracellular markers prior to cell death. According to our model, PKC inhibition causes enhancement of apoptotic pathways through signaling flux redistribution (SFR) (1,22). To check whether apoptosis is increased in TRAIL stimulated and BIM-I treated HT1080 cells, we measured PARP cleavage and p38 phosphorylation using western blotting assays, and caspase-3 activation and JNK phosphorylation using enzyme linked immunosorbent assays (ELISAs) (Figure 2). Consistent with the prediction of computational model, we observed substantial induction of PARP and caspase-3 cleavage, indicating increased apoptosis in HT1080 cells treated with BIM-I when compared with untreated cells in TRAIL stimulation (Figure 2A, top panel and Figure 2B, right panel). We further noticed enhanced p38 activation and low activity of JNK in TRAILstimulated cells treated with BIM-I (Figure 2A, middle panel and Figure 2B, left panel), in agreement with our model predictions for SFR at p62 pathway junction (1) (Figure 2C). Note that the housekeeping protein β-actin remained almost unaffected in the western blots. These results clearly demonstrate that BIM-I is a potential therapeutic target for HT1080 treatment.
To examine the expression levels of appropriate genes in TRAIL-stimulated HT1080, with and without BIM-I, we performed qRT-PCR experiments for several survival and apoptotic genes (rela, mtor, bcl2, bax, cytoc, and jun) at 0, 20, 40, 60, 120, and 180 min (Figure 3). Except for jun, the levels of genes were stable for up to 60 min, after which their expressions were significantly reduced, especially for BIM-I treated HT1080 cells, in correlation with the cell death dynamics (Figure 1C). These data indicate that, except for jun, transcription of the genes does not occur, perhaps due to the increased signaling flux through the apoptosis process depriving transcriptional signaling and, or due to the repression of pre-and post-transcriptional mechanisms found during apoptosis (34)(35)(36)(37)(38). Our observations are also consistent with other TRAILinduced apoptosis studies investigating gene expressions in HeLa (35) and MCF7 (36) cells.
Interestingly, jun levels showed an initial decrease during the first 20 min and then increased and stabilized after 120 min. This pattern indicates jun may evade the global transcriptional repression and play a role during apoptosis. Such behavior has been previously observed for other genes, in particular, genes translated through internal ribosome entry site (IRES)-mediated translation, which is known to occur during apoptosis after TRAIL stimulation of MCF7 cells (36,39). Notably, the presence of IRESs in jun transcriptional machinery has also been previously shown (40). Nevertheless, further investigation is required to define the exact role of jun during TRAIL and BIM-I mediated apoptosis.
Overall, the experiments demonstrate that the enhancement of cell death of BIM-I pre-treated TRAIL-stimulated cancer occurs through apoptosis.

IDENTIFICATION OF SPECIFIC PKC ISOFORM TARGET FOR ENHANCED CELL DEATH
Although we have demonstrated that PKC is a key target to enhance apoptosis in TRAIL-resistant cancer cells, it is unknown which PKC family isoform, among the 10 major members (α, β, γ, δ, ε, ι, θ, η, ζ, and µ), is a crucial single target. To investigate this, we first www.frontiersin.org    Table 1) in unstimulated HT1080 cells using qRT-PCR.
We observed the gene expressions of four PKC isoforms (α, δ, ε, and ι) were noticeably elevated, indicating that these isoforms may be crucial targets (Figure 4A). To investigate the effect of suppressing each of the four isoforms in TRAIL-stimulated HT1080 cells, we next performed siRNA-mediated PKC (α, δ, ε, and ι) KDs. The effect of each PKC KD was first confirmed after 24 h (Figure 4B). Consequently, we investigated cell viability by trypan blue for each of the four PKC KD conditions with and without TRAIL stimulation (200 ng/mL). Notably, PKCδ KD produced the most significant cell death of approximately 83% after 3 h (Figure 4C). Note that this result is almost identical to TRAIL-stimulated HT1080 www.frontiersin.org pre-treated with BIM-I at 3 h ( Figure 1C). Thus, our experiments reveal that PKCδ is the optimal single target for enhancing cancer apoptosis in TRAIL-based therapy.

DISCUSSION
TRAIL, a proinflammatory cytokine produced by the mammalian immune system, is known to induce apoptosis in cancer cells while leaving non-diseased cells largely unharmed (41,42). Hence, there has been intense interest in using TRAIL has a therapeutic target to treat cancers (43,44). However, not all cancers respond to TRAIL (45,46).
Previously, we investigated the TRAIL-resistant mechanism in HT1080 cells using a computational model (1). We predicted that the suppression of a novel pro-survival molecule would result in significant enhancement of apoptosis through signaling flux redistribution (22). PPI database search indicated that the Frontiers in Immunology | Tumor Immunity pro-survival molecule is a member of PKC. To experimentally validate this result, in this article, we investigated the effects of two TRAIL-resistant cancer cells to PKC inhibition.
First, using different doses of PKC inhibitor BIM-I together with various levels of TRAIL stimulation, we observed approximately 99 and 95% cell death occurred for HT1080 and HT29 cells, respectively (Figure 1). Notably, the effect on control TIG-1 and MRC-5 cells were less significant, at approximately 40 and 20% cell death, respectively.
Second, to confirm the mechanism for cell death is through apoptosis, we measured the activations of PARP and caspase-3 over 3 h in TRAIL-stimulated HT1080 cells untreated and treated with BIM-I, and compared with activations of p38 and JNK. We found that PARP, caspase-3 cleavages and p38 phosphorylation were significantly enhanced in BIM-I treated cells (Figure 2), while JNK activity was very low. These results are in consistency with the previous prediction of our computational model (1). We also investigated the expressions of major pro-and antiapoptotic genes, and found them to be mostly repressed at their transcription levels, especially after 1 h for BIM-I treated cells (Figure 3).
Third, to identify the crucial PKC family member for single specific target, we investigated the mRNA expressions of all 10 major isoforms in HT1080 cells. We selected the top four significantly expressed isoforms for developing siRNA KDs, and subsequent experiments demonstrated that PKCδ is a key target for enhancing cell death in TRAIL-resistant HT1080 cells (Figure 4).
It is worthy to mention other previous works that have studied PKC in different cancer types (47)(48)(49)(50). Although these works have demonstrated the importance of PKC, the investigations were performed in different cell lines or stimulations. In this work, however, we focused mainly on HT1080 and limitedly on HT29 cells. In addition, we bring to the attention the power of using multidisciplinary research to systemically identify a key target that can be experimentally tested. Therefore, to our knowledge, this is the first time the usefulness of a computational model is shown to identify a consistent and key target for regulating TRAIL-resistance. In summary, our work provides further evidence for the utility of systemic approaches in providing effective treatment strategies to tackle complex diseases.

AUTHOR CONTRIBUTIONS
Kentaro Hayashi and Kumar Selvarajoo conceptualized and designed the study. Kentaro Hayashi and Sho Tabata performed the wet lab experiments. Masaru Tomita and Kumar Selvarajoo provided cells, reagents, and discussions. Kentaro Hayashi, Vincent Piras, and Kumar Selvarajoo wrote the article. All authors read and approved the final manuscript.