Huwe1 inactivation promotes ovarian cancer metastasis by extracellular matrix deregulation

Objective: Huwe1 is critical for promoting the progression of a range of malignancies including ovarian cancer. Still, its role in ovarian cancermetastasis has not been reported. The goal of our studywas to evaluate a correlation betweenHuwe1 and ovarian cancer metastasis. Methods: The mouse ovary surface epithelium (MOSE) cells isolated from Huwe1 mice were identified by flow cytometry assay. Cell migration and invasion were analyzed by transwell system. The molecular mechanism of Huwe1 in ovarian cancer migration and invasion was explored by analyzing the difference of RNAseq data between MOSE-Huwe1-CreER and MOSE-Huwe1-Cre cells. Results: Huwe1 deletion significantly promoted tumor migration and invasion in ovarian cancer cells. Moreover, tumor cells were more frequently detected in the blood when the mice bearing allografts were treated with tamoxifen. Transcriptome analysis indicated that this phenotype may be due to the alteration in cell adhesion caused by the Huwe1 knockout. Conclusion: Huwe1 inactivation promoted ovarian cancer metastasis by the extracellular matrix (ECM) deregulation.


Introduction
Roles of the Huwe1, an E3 ligase, in DNA damage accumulation and tumor initiation, progression, and metastasis are critical. It degrades numerous tumor suppression and promotion substrates, such as Myc, BRCA1, TIAM1, H2AX, p53 and CDC6 [1]. Due to the complexity of its substrate, its roles in cancer remain controversial. Huwe1 served oncogenic roles in breast and prostate cancer, while it served anti-oncogenic roles in certain lung cancers and colorectal cancer [2]. Additionally, only a few studies have demonstrated the effect of Huwe1 on tumor metastasis, and the results were contradictory. In human lung cancers, Huwe1 stimulates cell metastasis by modulating the stability of TIAM1 [3]. In contrast, loss of Huwe1 promoted thyroid cancer cells migration and invasion in allografts [4]. The conflicting data indicated that the role of Huwe1 in tumor metastasis was likely to be cell type-dependent in response to different genetic change.
Ovarian cancer is common malignant tumors of the female genital organs. Because the ovary is located deep inside the pelvis, ovarian cancer is generally asymptomatic in prophase, and typically diagnosed in the terminal stage. Seventy percent of ovarian cancers would have spread to the greater omentum of the uterine adnexa, and various organs of the pelvis at the time of diagnosis [5]. Therefore, identification of key genes involved in ovarian metastasis may be beneficial for ovarian cancer treatment [6].
It was reported that high Huwe1 expression in patients with ovarian cancer was associated with worse prognosis [7,8]. Our previous study confirmed Huwe1 sustained tumor growth via the histone H1.3-H19 cascade [9]. However, the function of Huwe1 in ovarian cancer metastasis was still unclear. Here, MOSE cells were isolated and infected according to previous methods [9]. Subsequently, the potential roles of Huwe1 in ovarian cancer tumor metastasis were investigated, and the results showed that cell migration and invasion were promoted by Huwe1 inactivation in vitro. It was also shown that tumor cells were more frequently detected in the blood of the mice bearing allografts when Huwe1 was deleted by tamoxifen. Further studies suggested that the phenotype may be caused by deregulation of the extracellular matrix (ECM) induced by Huwe1 inactivation.

Quantitative Real-time PCR (qRT-PCR)
Total RNA was isolated using TRIzol reagent (Sigma-Aldrich, USA). DNA contaminants were removed by the TURBO DNA-free TM Kit (Invitrogen, USA). Subsequently, 2 µg of processed RNA was reverse transcribed by the Revert Aid First Strand cDNA Synthesis kit (Thermo Scientific, USA). Quantitative PCR was carried out in triplicate using the SYBR Green method (Life Technologies, USA). All operations were performed according to the manufacturer's protocols. All the primers used in this study are listed in Supplementary Table 1.

RNAseq analysis
The top-scored reads prioritized by FastQC v0.11.2 (available online at https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) achieved higher alignment accuracies. After sequencing, raw data were filtered to remove undesired reads. Clean reads were mapped to the mouse genome (version GRCm38.p4, which contained 21,936 protein-coding and 3495 lincRNA genes) using Tophat-2.1.0. Genes were defined distinctly up-or down-regulated when a false discovery rate (FDR)-adjusted p value ≤ 0.05 and an absolute value of fold change ≥3 using Cuffdiff [12]. The raw RNAseq data has been deposited into the SRA database on the NCBI website (Accession No. SRP081212).

Wound-healing assay
Cells were cultured to approximately 80% confluence in 6-well culture dishes. After mitomycin C (sigma, USA, 5 µg/mL) treatment was carried out, monolayer wounds were made by scratching with 200 µL pipette tips. Next, the cells were permitted to migrate into the wounded area for 8 h, and the scoring of cells in each well was recorded using a microscope. Cells migration was determined through measuring the width of the scratched area using ImageJ T M software (version 1.8.0, NIH, Bethesda, MD, USA).

Transwell invasion assay
To inhibit cell proliferation, the cells were processed using mitomycin C (5 µg/mL). Then, the cells were suspended in a serum-free medium, followed by inoculation in the upper migration chamber (Corning, NY, USA), precoated with Matrigel (BD Biosciences, San Jose, NJ, USA). The medium containing 10% FBS was added to the bottom migration chamber. After 8 h of directional migration, the upper insert was carefully removed. The cells on the bottom transwell chamber fixed in methanol, then stained with 0.1% crystal violet. After 30 min, the crystal violet was eluted via 10% acetic acid, and the absorbance of crystal violet solution was measured at 570 nm on a microplate reader.

Allograft mouse models of ovarian cancer
A total of 2 × 10 6 MOSE cells expressing KRAS G12D -2A-MYC-IRES-CreER-shp53 elements were injected subcutaneously into six weeks old BALB/c nude mice. After the subcutaneous tumors attained about 5-8 mm in diameter, tamoxifen or vehicle was administered by intraperitoneal injection every other day for 18 days until the end of the experiment. DNA was extracted from major tissues of mice and submitted to quantitative qRT-PCR to detect metastatic cells. The primer was located in the exogenous KRAS-2A-MYC element, and the sequence was listed in Supplementary Table 1.

Statistical analysis
All statistical analyses were carried out using Graph-Pad Prism 7 (GraphPad Prism, USA). The data are indicated as mean ± SD. Statistical difference was analyzed using an unpaired Student's t-test. Statistical significance was defined as *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

Huwe1 deletion promoted MOSE cells migration and invasion
To evaluate the roles of Huwe1 in ovarian cancer metastasis, the MOSE cells were isolated as previously described [9]. The MOSE cell infected by the vectors named MOSE-Huwe1 L/L -Cre or MOSE-Huwe1 L/L -CreER, respectively (Fig. 1A). Hereafter, MOSE cells will be adopted to refer to infected MOSE cells unless otherwise stated. To identify the type of isolated cells, a flow cytometry assay was applied, and the results indicated that almost all the tested cells expressed ovarian epithelial cells marker CK8 (Fig. 1B). To delete the Huwe1 gene, a Cre or CreER element was subcloned into the lentiviral vectors. Huwe1 expression was not detectable in cells expressing Cre. When CreER translocated into the nucleus mediated by 4-hydroxytamoxifen (4-OHT), the exons in most MOSE cells were also removed ( Fig. 1C-D). To explore the effects of Huwe1 on cell migration, a wound healing assay was carried out in MOSE cells (Supplementary Fig. 1). The results suggested that Huwe1 deletion significantly promoted the migration of MOSE cells into the central wound area. To further verify the result, a transwell migration assay was also done to test the migration of MOSE cells ( Fig. 2A). Consistent with the wound healing assay results, Huwe1-null cells migrated into the lower compartment of the migration chamber more frequently than the Huwe1wild cells. These results proved that Huwe1 significantly repressed the migration ability of MOSE cells in vitro. The Boyden chamber was modified with a thin layer of Matrigel to mimic the three-dimensional (3D) cell culture environment, and then the MOSE cells were cultured in the upper chambers to investigate the roles of Huwe1 in cell invasion potential. 8 hours later, the number of Huwe1-null cells reaching the other side of the filter was much higher than that of Huwe1-wild cells (Fig. 2B). All the above results proved that Huwe1 deletion promoted MOSE cells migration and invasion in vitro. We subcutaneously injected Huwe1 L/L -CreER cells into mice and used tamoxifen to induce Huwe1 deletion (Supplementary Fig. 2). First, we examined Huwe1 expression in mice when Huwe1 L/L -CreER allografts were induced with tamoxifen (Fig. 2C). The results showed that tamoxifen treatment significantly induced Huwe1 deletion. To observe the distribution of tumor cells in mice, we then investigated the number of tumor cells in each tissue, and only a small difference was observed in the liver and kidney tissues, whereas the number of tumor cells in the blood increased significantly (Fig. 2D).

Silencing Huwe1 promoted human ovarian cancer cell migration and invasion
To further explore the phenotype of Huwe1 loss in human ovarian cancer cells, SKOV-3 and OVCAR-3 cells were infected via lentivirus expressing tetracyclineinducible Huwe1 shRNA (teton-shHuwe1). After the SKOV-3 and OVCAR-3 cells were infected with a teton-shHuwe1 lentivirus, they were treated with tetracycline for 5 days. Huwe1 silencing was demonstrated in DOX-treated cells (Fig. 3A). We performed transwell assays to explore the roles of Huwe1 in SKOV-3 and OVCAR3 cell migration as well as invasion. Huwe1 knockdown evidently increased the number of invading cells compared with the untreated group, indicating that Huwe1 effectively inhibited SKOV-3 cells migration and invasion (Fig. 3B,C). Similarly, Huwe1 knockdown also significantly increased OVCAR3 migration. However, OVCAR3 cells failed to penetrate the basement membrane for invasion (Fig. 3D). This may be caused by the lower invasive capacity of OVCAR3 cells. The results confirmed that Huwe1 also had a similar function in human ovarian cancer cell.

Huwe1 inactivation promoted cancer metastasis by dysregulation biological adhesion
To further find out the molecular mechanism of Huwe1 in ovarian cancer migration and invasion, we carried out RNAseq to analyze the differences between MOSE-Huwe1 L/L -CreER and MOSE-Huwe1 L/L -Cre cells. Genome-wide analysis identified 25,431 genes, 141 of which had increased expression, and 55 had decreased expression (greater than 2-fold change) (Fig. 4A). Cluster analysis of genes with significant expression differences using the DAVID database showed that the differential genes expressed in the two cells were mainly extracellular matrix-associated genes (Fig. 4B). After reviewing these initial data sets, the gene functions of this set of 648 genes were studied. This included a significant reduction in 22 genes associated with biological adhesion, 10 of which were associated with cell-cell adhesion (Fig. 4C), and upregulation of 12 genes related to biological adhesion (Fig. 4D). Our analysis suggested that some adhesion molecules were up-regulated and some were downregulated. We suggested that Huwe1 knockout resulted in increased MOSE cells migration and invasion, owing to the interaction of multiple genes in the extracellular matrix leading to dysregulation of biological connections. The results of the study further confirmed this view.

Discussion
Huwe1 catalyzes a range of protein substrates ubiquitination, such as oncogene Myc, the tumor suppressors p53 and BRCA1, DNA damage repair proteins H2AX and the cell cycle protein Cdc6. It suggested that Huwe1 may be participated in many biological processes regulation, containing tumorigenesis, cell proliferation, migration and DNA repair [1]. Previous studies reported that Huwe1 overexpression was associated with cancers of the lung, colon, ovary, and pancreas [1,13]. Huwe1 promoted the ubiquitination and degradation of the tumor suppressor BRCA1, thereby impeding BRCA1-dependent DNA damage repair in breast cancer cells [14]. Additionally, Huwe1 promoted gastric cancer cells migration and invasion [15]. However, Huwe1, via regulating the stabilization of p53, repressed thyroid cancer progression. Huwe1 also inhibited the de-velopment of mouse skin cancer [16]. Huwe1 inactivation inhibited the colon cancer cell line proliferation and its tumorigenicity in mice [17]. These results suggested that role of Huwe1 is diverse in different cancer types.
The existence of metastases and invasion are responsible for most tumor death, but the relationship between Huwe1 expression and ovarian tumor metastasis has not been reported. In this study, migration was obviously increased in Huwe1 deletion cells compared with Huwe1 ECM, a highly dynamic structure, is composed of collagens, laminins and several other glycoproteins. Not only does the ECMs provided physical scaffolds for cells embedding but also regulated lots of cellular processes, such as growth, differentiation and homeostasis. Some studies even reported that ECM homeostasis deregulation triggered pathological ECM remodeling and even promoted tumor growth and migration [18]. Collagen composition altered obviously in breast cancer where fibrillar collagens I, III and V increased, and type IV collagen was decreased [19]. Heparan sulphate proteoglycans (HSPGs) could be degraded by heparanase (HEP). Under normal physiologi-cal conditions, HEP formed heterodimers and did not cleave HSPGs. Nevertheless, under pathological conditions including cancer, HEP drove cleavage of HSPGs promoting cell migration, invasion, and ultimately, metastatic dissemination [20]. Overexpression of the β3 subunit of laminin-332 (LAMB3) promoted colorectal tumor growth and metastasis [21]. Additionally, the ECM-related genes were obviously different in the two cell lines in the study. These results suggested that ECM deregulation influenced many aspects of cancer cell behaviors, such as migration and invasion.
Huwe1, as we know, sustained tumor development via the histone H1.3-H19 cascade. In Huwe1 deficient cells, increased expression of histone H1.3 could inhibit the expression of non-coding RNA H19 [9]. H19 knockdown led to ten-eleven translocation (TET) family proteins3 (TET3) downregulation through H19/let-7 axis. TET3 downregulation repressed TGF-β signaling and its downstream signaling molecules, such as TGFBR2, TGFBR1 and Smad pro-teins, ultimately leading to ECM change [22]. To explore the relationship between Huwe1, ECM, and tumor migration, and to validate whether Huwe1 inhibits cell migration by regulating some mechanism of the ECM, we performed RNAseq. Cluster analysis of genes with significant expression differences was performed using the DAVID database. It showed that the genes with differential expression in the two cell lines were mainly ECM-related genes. Loss of HAPLN4, SVEP1, and DPT were positively correlated with migration and invasion of tumor cells in previous studies [23][24][25]. Our data further confirmed that these genes were significantly downregulated in the MOSE-Huwe1 L/L -Cre group. Upregulation of HES1, CNTNAP2, and PCDH19 positively correlated with migration and invasion of tumor cells also have been demonstrated [26][27][28][29], and these genes were significantly upregulated in the MOSE-Huwe1 L/L -Cre group. The results demonstrated that the increase of MOSE cells migration and invasion induced by Huwe1 knockout might be due to the ECM homeostasis deregulation caused by multiple gene changes.
Apart from changing the ECM to promote metastasis via the histone H1.3-H19 cascade, the regulation of various substrates by Huwe1 may involve cell invasion and migration [3,4]. For example, the c-Myc/Miz1 complex accumulation, the tumor suppressor p53 loss and other related proteins can promote cell spreading and migration [16,30]. Besides, in a comprehensive bioinformatics analysis of ENO1related genes, the researchers identified three hub genes including Huwe1, all of which are involved in tumor metastasis [31]. However, the specific mechanism of the ECM affecting ovarian cancer cell migration and invasion, remain to the further research.

Conclusions
In summary, this study demonstrated that inactivation of Huwe1 promoted ovarian cancer cells migration and invasion. Moreover, the metastatic inhibitory effect of Huwe1 in ovarian cancer due to its involvement in the homeostasis regulation of ECM.

Author contributions
DY designed the research study. FZ, JG and SY performed the research. FZ and YM analyzed the data. FZ, JG and XZ wrote the manuscript. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript.