High-throughput functional screen identifies YWHAZ as a key regulator of pancreatic cancer metastasis

Abstract

Pancreatic cancer is a leading cause of cancer death due to its early metastasis and limited response to the current therapies. Metastasis is a complicated multistep process, which is determined by complex genetic alterations. Despite the identification of many metastasis-related genes, distinguishing the drivers from numerous passengers and establishing the causality in cancer pathophysiology remains challenging. Here, we established a high-throughput and piggyBac transposon-based genetic screening platform, which enables either reduced or increased expression of chromosomal genes near the incorporation site of the gene search vector cassette that contains a doxycycline-regulated promoter. Using this strategy, we identified YWHAZ as a key regulator of pancreatic cancer metastasis. We demonstrated that functional activation of Ywhaz by the gene search vector led to enhanced metastatic capability in mouse pancreatic cancer cells. The metastasis-promoting role of YWHAZ was further validated in human pancreatic cancer cells. Overexpression of YWHAZ resulted in more aggressive metastatic phenotypes in vitro and a shorter survival rate in vivo by modulating epithelial-to-mesenchymal transition. Hence, our study established a high-throughput screening method to investigate the functional relevance of novel genes and validated YWHAZ as a key regulator of pancreatic cancer metastasis.

Fig. 1. Establishing the regulated high-throughput mutagenic libraries.

A piggyBac transposon-based gene search vector (PB-GSV) contains TRE (tetracycline response element) and CMV-Neo-PA (neo-expression cassette). PB-GSV could be efficiently integrated into the genome when co-transfected with transposase mPB. Tetracycline transactivator (tTA) could initiate mutagenesis when binding to the TRE element. This mutagenic effect was turned off in the presence of tetracycline (+Dox) (Tet-off system). The piggyBac-based mutagenic units combined with the Tet-off system resulted in a regulable high-throughput mutagenic system. B Different mutagenic effects were initiated by the sense RNA or antisense RNA depending on the integration orientation of PB-GSV in the genome. C Flow chart illustrates the screening including the establishment of mutagenic libraries, screening metastatic lesions in mouse models, functional validation of the metastatic ability of the subpopulations cultured from the metastatic lesions in the first-round screening in vivo by turning off the mutagenesis through the Tet-off system. The metastatic lesions in the second-round screening in vivo were cultured, and the integration sites were analyzed. D Titration of the transfection efficiency of the gene search vector. E High integration efficacy was achieved with an optimized ratio of PB-GSV and mPB. F The integration efficacy was represented by the number of G418-resistant clones stained by methylene blue.

Fig. 2. Screening for highly metastatic subclones in vivo to identify the candidate genes.

A Enhanced metastatic capacity of the mutagenic libraries in vivo, representative images were shown. Upper: organs from mice transplanted with control cells Pan02-4B3. Lower: organs from mice transplanted with the library. B Anxa3 was identified by Splinkerette PCR from the metastatic lesions. Target bands were extracted and sequenced. C Sequence alignment in UCSC Blat for Anxa3. PB-GSV initiated an antisense RNA which was opposite to the direction of Anxa3 transcription. D Overall survival time was prolonged when intraperitoneally transplanting GSV-Anxa3 subclone into mice with Dox treatment (+Dox) as compared to the control group (−Dox) (n = 10, *p < 0.05, Kaplan–Meier survival analysis). E Macro-metastatic lung nodules decreased when subcutaneously transplanting GSV-Anxa3 subclone into mice with Dox treatment (+Dox) as compared to the control group (−Dox). Mice were sacrificed on day 55 (n = 10, *p < 0.05).

Fig. 3. Functional validation of Ywhaz as a metastasis-promoting gene.

A Tumor growth was not affected in mice subcutaneously transplanted with GSV-Ywhaz subclones with Dox treatment (n = 6; ns, not significant). B The total number of macro-metastatic lung nodules decreased with Dox treatment (+Dox) as compared to the control group (−Dox). Mice were sacrificed on day 60 after subcutaneous transplantation of GSV-Ywhaz (n = 6, *p < 0.05). C Representative images of macro-metastatic lung nodules. D Overall survival time was prolonged when intravenously transplanting GSV-Ywhaz subclone into mice with or without Dox treatment (n = 6, *p < 0.05, Kaplan–Meier survival analysis). E Western blotting analysis showed the upregulation of YWHAZ expression in GSV-Ywhaz subclones and downregulation of YWHAZ expression after DOX treatment.

Fig. 4. YWHAZ is a regulator of mouse and human pancreatic cancer metastasis.

A Western blotting analysis showed YWHAZ was overexpressed in Pan02 cells (Pan02 Yw-OV) compared to the control cells (Pan02 Yw-WT). B Tumor growth was not affected by YWHAZ overexpression in s.c. model (n = 7; ns, not significant). C YWHAZ overexpression was associated with worse overall survival in s.c. model (n = 12, *p < 0.05, Kaplan–Meier survival analysis). D YWHAZ overexpression resulted in more lung macro-metastasis in s.c. model. Representative images of macro-metastatic lung nodules were shown. E Macro-metastatic lung nodules were compared between Pan02 Yw-OV and Pan02 Yw-WT (n = 7, *p < 0.05). F Validation of YWHAZ overexpression in AsPC-1 by Western blotting analysis (AsPC-1 YW-OV). G and H YWHAZ overexpression resulted in more micro-metastatic (G) and macro-metastatic (H) lung lesions in o.r. model (n = 6, *p < 0.05). I Kaplan–Meier survival analysis showed that YWHAZ overexpression significantly shortened the overall survival time in AsPC-1 cells in o.r. model (n = 5, *p < 0.01). J Kaplan–Meier survival analysis showed that high YWHAZ expression in pancreatic tumor samples was associated with a worse prognosis in clinical patients (data from TCGA and HPA database, N = 173, median survival time in high and low expression group was 15.33 and 23.17 months, respectively). The multivariate COX regression analysis, integrated with patients’ age, gender, and TNM stage, revealed YWHAZ was an independent negative prognostic factor for pancreatic cancer (HR = 2.65, 95%; CI:1.68–4.16; p < 0.0001).

Fig. 5. Overexpression of YWHAZ promotes metastatic phenotypes.

A The GSV-Ywhaz subclone showed enhanced invasion capacity as compared to the parental Pan02-4B3 cells with transwell assay (n = 3, ***p < 0.0001). B YWHAZ overexpression promotes tumor invasion in Pan02 cells with transwell assay (n = 3, *p < 0.05). C YWHAZ overexpression in AsPC-1 cells resulted in enhanced invasion capacity with transwell assay (n = 3, *p < 0.05). D The GSV-Ywhaz subclone showed enhanced anchorage-independent growth capacity in soft agar as compared to the parental Pan02-4B3 cells (n = 5, ***p < 0.0001). E YWHAZ overexpression promotes anchorage-independent growth in soft agar in Pan02 cells (n = 6, ***p < 0.0001). F YWHAZ overexpression resulted in enhanced anchorage-independent growth capacity in soft agar in AsPC-1 cells. (n = 6, ***p < 0.05). The representative image of colonies was shown. G The expression of differentiation and EMT-related markers were investigated by qPCR (n = 3, *p < 0.05). H The expression of differentiation and EMT-related markers in the GSV–Ywhaz subclone was compared with or without Dox treatment by qPCR. The expression of mRNA was normalized against GAPDH (n = 3, *p < 0.05). I Western Blotting analysis showed the downregulation of E-cadherin expression, the upregulation of phosphorylated AKT (pAKT), and Snail expression in GSV-Ywhaz subclone.

Fig. 6. YWHAZ promotes pancreatic cancer metastasis via EMT.

A Heat map of the DEGs identified by RNA-seq in Pan02 control cells (Yw-WT) and YWHAZ overexpression cells (Yw-OV) (>2-fold, p < 0.05, n = 3). B Top enriched GO terms in downregulated genes. C Western blot analysis of the EMT biomarkers and transcription factors, Ywhaz was cloned with a fusion HA-tag and anti-HA shows the expression of YWHAZ. D Hematoxylin-eosin staining and immunohistochemical staining of the EMT biomarkers and transcription factors in subcutaneous tumors. E-cadherin and ZO-1 not only showed a difference in intensity but also a change in location after YWHAZ overexpression (n = 3, p < 0.05).