In this study, among many genes associated with MAPK, we found that knockdown of SON remarkably suppressed the proliferation, survival, and tumor formation of pancreatic cancer cells. The suppressive effect was less pronounced in normally phenotypic ductal cells. In primary pancreatic cancer tissues, SON was overexpressed in ductal adenocarcinomas compared with normal duct cells and PanINs. Knockdown of SON induced G2/M arrest and apoptosis. SON shuttled between the nucleus and cytoplasm depending on the phase of cell cycle. These results indicate that SON plays a crucial role in the proliferation, survival, and tumorigenicity of pancreatic cancer cells, thus suggesting that this molecule could be a prime therapeutic molecular target for pancreatic cancer.
Our investigation showed that knockdown of MAPK-associated molecules suppressed the proliferation of pancreatic cancer cells in vitro to variable degrees. We found that knockdown of AURKB, CENPA, EBNA1BP2, GOLT1A, KIF11, NEDD4L, SON, TPX2, or WDR5 strongly suppressed the proliferation. AURKB encodes aurora kinase B (AURKB), which is involved in chromosome segregation and cytokinesis during mitosis . CENPA encodes centromere protein A (CENPA), which, by functioning as a replacement for histone H3 in centromeric nucleosomes, plays an essential role in kinetochore formation and functions in cellular mitosis . EBNA1BP2 encodes a ribonucleoprotein, Epstein-Barr virus nuclear antigen 1-binding protein 2 (EBNA1BP2), which serves as a scaffold for ribosome biogenesis . GOLT1A encodes Golgi transport 1A (GOLT1A), which functions as a transporter on the Golgi membrane . KIF11 encodes a microtubule-dependent motor protein, kinesin family member 11 (KIF11), which plays a critical role in chromosome positioning during mitosis . NEDD4L encodes neural precursor cell expressed, developmentally down-regulated 4-like, an E3 ubiquitin protein ligase (NEDD4L) that plays a role in polyubiquitination and proteasomal destruction of SMAD2/3 . TPX2 encodes a homologue of Tpx2 of Xenopus (TPX2), a binding partner of aurora kinase A (AURKA) that plays a role in microtubule spindle formation . WDR5 encodes WD repeat domain 5 (WDR5), which binds methylated histone H3 lysine 4 (H3K4) and is required for recruiting H3K4 methyltransferase . Among these, AURKB, CENPA, KIF11, and TPX2 are involved in functions of the microtubule spindles and kinetochores, which are considered essential for cell mitosis. Because we screened by assaying the effects of knockdown of the MAPK-associated genes on in vitro proliferation of pancreatic cancer cells, molecules associated with the microtubules and kinetochores might be selectively represented in our screening. Interestingly, these microtubule kinetochore-associated molecules have already been studied as molecular targets in various cancers [22–25]. Nevertheless, of these MAPK-associated molecules, we found that knockdown of SON most remarkably suppressed proliferation, which led us to investigate SON in detail as a candidate molecular target.
SON encodes SON, a large protein harboring a serine or arginine-rich domain. It was first cloned as a gene encoding a protein with DNA-binding activity. However, subsequently, it turned out to be a nuclear speckle protein involved in RNA processing and required for proper and efficient splicing of pre-mRNAs [26–30]. In our study, knockdown of SON attenuated the proliferation, survival, and tumorigenicity of pancreatic cancer cells. These suppressive effects were attributable to cell cycle arrest at the G2/M phase and apoptosis induced by depletion of SON. The association between the depletion of SON and G2/M arrest has been reported to be associated with impairment of spindle pole separation, microtubule dynamics, and genome integrity due to inadequate RNA splicing of a specific set of cell cycle-related genes with weak splice sites, i.e., splice sites without the conserved sequence .
Pancreatic cancer cells were more susceptible to depletion of SON than normally phenotypic cells. This may be due to rapid progression through the cell cycle in cancer cells, which results in exaggerated dependence on SON to maintain efficient RNA processing of the cell cycle-related genes. This interpretation could be endorsed by the overexpression of SON we found in most ductal adenocarcinomas, compared with normal ductal cells or precursor lesions, which suggests that adenocarcinoma cells depend on SON more strongly than normal ductal cells and precursor lesions to maintain their phenotypes. These results suggest that depletion of SON may specifically lead to an anticancer phenotype. SON overexpression is purportedly due to the constitutive activation of MAPK in ductal adenocarcinoma; however, other possible causes, such as gene amplification or aberrations in protein turnover, cannot be ruled out and will be a subject of further study.
The dynamics of SON distribution during the cell cycle is not well known. We performed live-cell imaging of cells expressing EGFP-SON and observed that SON dispersed in the cytoplasm during early mitotic phase formed small foci in the cytoplasm in the late mitotic phase, and gradually redistributed as speckles in the nucleus as foci in the cytoplasm faded. The cytoplasmic small foci are supposed to be mitotic interchromatin granules that correspond to accumulations of nuclear speckle proteins in the cytoplasm in the late mitotic phase [31, 32]. These dynamics seem similar to the dynamics of another speckle protein, SF2, and are consistent with the idea that SON plays a role in the appropriate organization of RNA splicing factors [29, 33, 34].
The knockdown of SON by RNA interference showed sufficient anti-cancer phenotypes experimentally. For the RNA interference, vector-mediated stable transduction appeared to be more effective than oligonucleotide-based transient transduction as shown in Figure 2. Although the stable knockdown of SON by RNA interference could be an efficient molecular therapy for pancreatic cancer, the lack of a conventional method for tissue-specific, stable delivery of short, double-stranded RNA could limit the use of this approach in clinical therapeutics. Indeed, the use of RNA interference in clinical practice is generally not warranted. Recently, however, systemic delivery of siRNA combined with a special nanoparticle successfully knocked down a target gene in melanoma in a clinical trial . The use of such a technique to attempt specific knockdown of SON in pancreatic cancer cells in a clinical model is worth trying and is an issue to be resolved in a future study. The results of this study also suggest that development of a molecule-oriented chemical substance against SON as therapy for pancreatic cancer is warranted.