Silencing protein kinase C ζ by microRNA-25-5p activates AMPK signaling and inhibits colorectal cancer cell proliferation

Developing novel strategies against human colorectal cancer (CRC) cells is needed. Activation of AMP-activated protein kinase (AMPK) could possibly inhibit CRC cells. Protein kinase C ζ (PKCζ) is an AMPK negative regulator. Here we found that PKCζ expression was significantly elevated in human colon cancer tissues and CRC cells. PKCζ upregulation was correlated with AMPK in-activation and mTOR complex 1 (mTORC1) over-activation. Reversely, PKCζ shRNA knockdown activated AMPK signaling and inhibited HT-29 cell proliferation. Significantly, downregulation of microRNA-25-5p (miR-25-5p), a PKCζ-targeting miRNA, could be the cause of PKCζ upregulation. Exogenous expression of miR-25-5p silenced PKCζ to activate AMPK signaling, which inhibited HT-29 cell proliferation. In vivo studies showed that HT-29 xenograft growth in mice was inhibited after expressing PKCζ shRNA or miR-25-5p. Collectively, PKCζ could be a novel oncogenic protein of human CRC. PKCζ silence, by targeted-shRNA or miR-25-5p expression, activates AMPK and inhibits HT-29 cell proliferation.


INTRODUCTION
Colorectal cancer (CRC) is a major threat to human health [1,2]. It causes large cancer motilities each year [1,2]. With the recent progress achieved in the diagnosis and clinical treatments for CRC, the prognosis of this devastating disease has been improved in the past decades [3,4]. Yet, for the patients with advanced, recurrent and metastatic CRC, the five-year overall survival is still poor [1,2]. Meanwhile, the incidence of this disease is rising in China and other regions of the world [1,2]. Therefore, it is important to indentify novel and key oncogenic proteins of human CRC.

PKCζ upregulation in human colon cancer tissues and CRC cells
First, we tested expression of PKCζ in human colon cancer tissues. Quantitative real-time PCR (qRT-PCR) assay was employed, and results showed that PKCζ mRNA expression level was significantly elevated in fresh colon cancer tissues ("Tum", Figure 1A). It level was over five times higher than that in the surrounding normal colon epithelial tissues ("Nor", Figure 1A). Quantified Western blotting assay results (integrating 10 sets of abovementioned samples) in Figure 1B confirmed PKCζ protein upregulation in cancer tissues. As discussed, PKCζ is a negative regulator of AMPK [23,24]. AMPK signaling was then tested in above tissues. As demonstrated, AMPK activation (p-AMPKα1 at Thr-172) was much lower in colon cancer tissues (see quantified results in Figure  1B), as compared to that in the normal tissues. AMPK activation could inhibit mTORC1 directly or indirectly [13,25,26]. We here showed that p-eIF4E-binding protein 1 (4E-BP1, Ser-65), indicating mTORC1 activation [27], was indeed significantly higher in the cancer tissues (See quantified results in Figure 1B). These results confirmed PKCζ upregulation in human colon cancer tissues, which was correlated with AMPK inhibition and mTORC1 activation.

PKCζ shRNA knockdown activates AMPK and inhibits HT-29 cell proliferation
In order to study the possible function of PKCζ in CRC cells, shRNA strategy was employed. As discussed, a panel of three distant PKCζ-targeting shRNAs ("1#/2#/3#"), with non-overlapping sequences, were applied. The lentiviral shRNA was added to cultured HT-29 cells. After puromycin selection, stable cells were established. qRT-PCR assay results in Figure 2A demonstrated that PKCζ mRNA expression was indeed dramatically decreased in PKCζ shRNA-expressing stable HT-29 cells. Consequently, PKCζ protein expression was also silenced ( Figure 2B). On the other hand, AMPK activation, tested again by p-AMPKα1 and p-LKB1, was increased in the PKCζ-silenced cells ( Figure 2B). mTORC1 activation (p-4E-BP1) was yet inhibited ( Figure  2B).
As discussed, AMPK activation could lead to mTORC1 inhibition and proliferation inhibition in CRC cells [8][9][10]22]. We therefore tested proliferation of above CRC cells. Simple viable cell (trypan blue negative) counting assay results in Figure 2C demonstrated that PKCζ knockdown clearly inhibited HT-29 cell proliferation. The number of viable cells (at 72 hours) was dramatically lower in PKCζ shRNA-expressing HT-29 cells (vs. control parental cells, Figure 2C). Consequently, population doubling time of HT-29 cells was significantly increased after PKCζ knockdown ( Figure 2D). Meanwhile, BrdU ELISA optic density (OD, normalized to the viable cell number) was also significantly decreased following PKCζ knockdown ( Figure 2E). On the other hand, cell apoptosis level, tested by Histone DNA ELISA assay, was increased with PKCζ silence ( Figure 2F). Notably, nonsense shRNA control ("sh-C") failed to have above actions (Figure 2A-2F).

Exogenous expression of miR-25-5p silences PKCζ and inhibits HT-29 cell proliferation
miR-25-5p is a recently-indentified PKCζ-targeting miRNA [24], its level was negatively correlated with PKCζ level in human colon cancer tissues and CRC cells ( Figure 1). miR-25-5p downregulation could therefore be the cause of PKCζ upregulation. To support this hypothesis, the pre-miR-25-expressing vector ("miR-25-Vec", a gift from Dr. Cui [24]) was introduced to the HT-29 cells. Via selection, two stable HT-29 cell lines expressing miR-25-Vec were established, named as "miR-25-Vec-L1" and "miR-25-Vec-L2". In line with previous findings [24], expression of miR-25-5p was significantly elevated in the two lines ( Figure 3A). miR-25-3p level was not changed in these cells (Data not shown), suggesting that miR-25-5p could be the primary product of the vector (reported in [24]). Exogenous expression of miR-25-5p led to a dramatic reduction of PKCζ mRNA UTR luciferase activity ( Figure 3B). PKCζ mRNA ( Figure 3C) and protein expression ( Figure 3D) was also depleted in the two HT-29 cell lines, where AMPK activation (p-AMPKα1) and mTORC1 (p-4E-BP1) inhibition were subsequently observed ( Figure 3D). As compared to the parental control cells, proliferation of the two stable cell lines was significantly inhibited, which was again tested by cell counting assay ( Figure 3E) and BrdU ELISA assay ( Figure 3F). Notably, the vector control ("Vec") failed to change above gene expression and HT-29 cell proliferation ( Figure 3A-3F). These results indicate that PKCζ could be the direct and primary target of miR-25-5p in HT-29 cells.

HT-29 xenograft growth in SCID mice is inhibited after expressing PKCζ shRNA or miR-25-5p
The potential effect of PKCζ shRNA or miR-25-5p on CRC cell growth in vivo was tested next. Same amount of HT-29 cells, bearing PKCζ shRNA ("1#", see Figure  2), miR-25-5p ("L1", see Figure 3) or the parental control HT-29 cells ("Par", or control tumors) were inoculated to the SCID mice via s.c. injection. Tumor recordings were initiated when the tumor volume was about 100 mm 3 of each group. Weekly tumor growth curve result in Figure 4A demonstrated that the in vivo growth of HT-29    xenografts was significantly inhibited after expressing the PKCζ shRNA or miR-25-5p. The tumor sizes of the two groups were much smaller than those of the control tumors ( Figure 4A). Mice body weight was not significantly different between the three groups ( Figure 4B). To test signaling changes in above tumor tissues, at Week-2 and Week-4, one HT-29 tumor per group was isolated. Western blotting assay was again applied to test above signaling proteins in fresh tumor lysates. As compared to the control tumors, depleted PKCζ, enhanced AMPK activation and decreased p-4E-BP1 were noticed in tumors bearing PKCζ shRNA or miR-25-5p ( Figure 4C). Meanwhile, downregulation of PCNA (a proliferation marker) and induction of cleaved-PARP (an apoptosis marker) were observed in tumor tissues with PKCζ shRNA or miR-25-5p ( Figure 4C) Therefore, the in vivo signaling changes were in line with the in vitro findings.

DISCUSSION
Recent studies have implied that forced AMPK activation could be a novel and efficient strategy to inhibit CRC cells. For example, Chen et al., showed that AMPK activation mediated plumbagin-induced growth inhibition of CRC cells [9]. Kang et al., demonstrated that Widdrolinduced CRC cell death requires AMPK activation [29]. Aqueous Oldenlandia diffusa extracts and capsaicin also activated AMPK-dependent death pathway in CRC cells [8,30]. Further, activation of AMPK-dependent autophagic pathway contributed to C6 ceramide-induced inhibition of HT-29 cells [20]. Reversely, AMPK inhibition, for example via expressing microRNA-451, could promote CRC cell proliferation and progression [10]. In the current study, we proposed that upregulation of PKCζ, the negative regulator of AMPK [23,24], could likely lead to AMPK inhibition in human colon cancer tissues and CRC cells. On the other hand, PKCζ shRNA knockdown activated AMPK signaling and inhibited HT-29 cell proliferation in vitro and in vivo. These results imply that PKCζ could be a novel oncogenic protein for human CRC, possibly via shutting down AMPK signaling.
miRNA-mediated gene alteration is important in controlling gene expression at the post-transcriptional level [44,45]. miRNAs could be essential in regulating many key biologic processes of human cells, possibly via regulating expression of signaling molecules including growth factors, cytokines, transcription factors and others [46,47]. It is known that miRNAs play vital functions in promoting CRC tumorigenesis and progression [46,47]. miRNA dysregulation is now known as a hallmark of CRC and many other malignancies [46,47]. In the current study, we showed that miR-25-5p, an anti-PKCζ miRNA [24], was downregulated in human colon cancer tissues and CRC cells, which could be the cause of PKCζ upregulation. Remarkably, forced exogenous expression of miR-25-5p silenced PKCζ, activated AMPK and inhibited HT-29 cell proliferation. More importantly, in vivo growth of HT-29 xenografts was largely suppressed after expressing miR-25-5p. These results imply that miR-25-5p could be an anti-cancer miRNA in CRC cells.
We here showed that expression of miR-25-5p inhibited CRC cell proliferation possibly via downregulating PKCζ. It is certainly possible that other targets of miR-25-5p could also be involved in above actions. For instance, several potential miR-25's targeted genes have been identified thus far in cancer cells, including the apoptosis protein Bim [48] and mitochondrial calcium uniporter [49]. Further studies will be needed to identify possible other targets of miR-25-5p in CRC cells.

Chemicals and reagents
Puromycin and neomycin were purchased from Sigma Aldrich (Shanghai, China). The reagents for cell culture were all obtained from Gibco (Shanghai, China). The PKCζ antibody was a gift from Dr. Cui [24]. All other antibodies were purchased from Santa Cruz Biotech (Santa Cruz, CA).

Cell culture
As described, HT-29, DLD-1, HCT-116, Lovo, SW403 and SW48 human CRC cell lines were provided by Shanghai Institute of Biological Science (Shanghai, China). FHC colon epithelial cells were purchased from the iBS Fudan Cell Bank (Shanghai, China). Cells were maintained in DMEM plus 10% fetal bovine serum (FBS). The number of viable cells (trypan blue exclusive) was recorded via the TC20 automatic counter (Bio-Rad, Shanghai, China). All cell lines utilized in this study were subjected to mycoplasma and microbial contamination examination every two months. Population doubling time, colony forming efficiency, and morphology were also examined routinely for cell authentication.

Human tissues
A total of 10 primary colon cancer patients (All polyp adenocarcinoma; 6 male and 4 females; Broders stages II-III) were enrolled in this study. The surgeryisolated colon cancer tissues and matched surrounding normal colon tissues were thoroughly washed. Tissues were then mechanically dissociated and lysed by the tissue lysis buffer (Sigma). Expressions of listed genes and miR-25-5p in fresh tissue lysates were examined. The protocols utilizing human tissue samples were in accordance with the principles expressed in the Declaration of Helsinki, and were approved by Fujian Medical University. Writteninformed consent was obtained from each participant.

BrdU assay
To test cell proliferation, cells were incubated with 10 μM of BrdU (Cell Signaling Tech, Shanghai, China). BrdU incorporation was determined in the ELISA format. BrdU OD value was always normalized to cell number.

Histone-DNA ELISA assay
Histone-DNA ELISA PLUS kit (Roche Applied Science, Shanghai, China) was applied to quantify cell apoptosis according to the manufacturer's protocol. ELISA OD at 450 nm was recorded.

Western blotting assay
As described [54,55], aliquots of 30 μg (per sample) tissue or cell lysates (in RIPA lysis buffer) were electrotransferred to 10-12% SDS-PAGE gel, following by transfer to PVDF membranes. The blots were blocked by 10% milk and were then incubated with designated primary and secondary antibodies. The antigen-antibody binding was detected via enhanced chemiluminescence (ECL) reagents (GE Healthcare, Shanghai, China). The blot was quantified via ImageJ software.

PKCζ shRNA
Three different lentiviral shRNAs (packed into the GV248-puromycin vector) against human PKCζ were reported early [24] and were provided by Genepharm Co. (Shanghai, China). HT-29 cells were seeded onto 6-well plate at 50% confluence. The lentiviral shRNA (10 μL/mL medium per well) was added to cells for 24 hours. Cells were then cultured in puromycin (1 μg/mL)-containing medium for another 5 days. PKCζ expression in the stable cells was tested by Western blotting assay and/or qRT-PCR assay. The scramble lentiviral shRNA (Genepharm Co.) was added to the control HT-29 cells.

AMPKα1 shRNA
The AMPKα1 shRNA lentiviral particles (Santa Cruz, sc-29673-V) were added to HT-29 for 24 h. Afterwards, cell culture medium was replaced with puromycin for a total of 5 days, until resistant stable colonies were formed. AMPKα1 expression in resulting stable cells was tested by Western blotting assay.

PKCζ mRNA luciferase assay
The UTR reporter vector that contains the 3′-UTR of PKCζ carrying the miR-25-5p site was provided again by Dr. Cui [24]. The complementary oligonucleotides for the selected region were hybridized to form doublestranded DNA and inserted into pmIR-Reporter firefly luciferase vector (Genepharm). The construct was then co-transfected with miR-25 vector into HT-29 cells. The cells were then lysed via a luciferase assay kit (Promega, Shanghai, China), which were then tested on a luminescence microplate reader.

Xenograft assay
As described [54], the female severe combined immunodeficient (SCID) mice (age 4-5 week, weight 18-19 g) were obtained from the Animal Center of Fujian Medical University (Fuzhou, China), and were housed under standard procedures. Animals were randomized into three groups. HT-29 cells (5 × 10 6 cells in 100 μL of saline/Matrigel, 1:1 v/v) of different genetic manipulation were inoculated subcutaneously into the flanks of the SCID mice. When the tumor reached approximately 100 mm 3 , recordings were started. The size of the tumors was measured by caliper every week, and tumor volumes were calculated using the following formula: V (volume) = 0.5328 × Long × Width × High (mm 3 ) [8,56,57]. For recording mouse body weight, the estimated tumor weight (tumor volumes × 1g/cm 3 ) was subtracted from total weight of each mouse. Mice were maintained under the following conditions: 12hour dark/12-hour light cycle, 24 ± 2°C temperatures, and 50±10% humidity. The protocol was approved by the Nanjing Medical University's Institutional Animal Care and Use Committee (IACUC) and Ethics Review Board (ERB).

Statistics
The data presented were mean ± standard deviation (SD). Statistical differences were analyzed by one-way ANOVA followed by multiple comparisons with post hoc Bonferroni test (SPSS version 18.0). Values of p < 0.05 were considered statistically significant.

CONCLUSION
Collectively, these results suggest that PKCζ could be a novel oncogenic protein for human CRC, possibly via shutting down AMPK signaling. On the other hand, PKCζ silence, by targeted-shRNA or miR-25-5p expression, activates AMPK and inhibits HT-29 cell proliferation.

Author contributions
All authors carried out the experiments, participated in the design of the study and performed the statistical analysis, conceived of the study, and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.