Long non-coding RNA FAM83H-AS1 induced by adipose-derived stem cells promotes breast and pancreatic cancer cell proliferation and migration


 Background

Stromal cells recruited to the tumor microenvironment and long non-coding RNAs (lncRNAs) in the tumor cells regulate cancer progression. However, their relationship is largely unknown.
Methods

In the current study, we identified the effects of lncRNA FAM83H-AS1, induced by adipose-derived stem cells (ADSCs) during tumor development, and explored the underlying mechanisms using a coculture cell model. Adipose tissues were obtained from healthy female donors, the expression of stromal markers on cell surface of expanded ADSCs were confirmed using immunofluorescence analysis. The breast and pancreatic cancer cells were cultured with or without ADSCs using 24-well transwell chamber systems with 8.0 µm pore size.
Results

Our results showed that FAM83H-AS1 was upregulated in breast and pancreatic cancers and associated with poor prognosis. ADSCs further induced FAM83H-AS1 and increased tumor cell proliferation via promoting G1/S transition through cyclin D1, CDK4 and CDK6. Wound healing, modified Boyden chamber and immunoblotting assays demonstrated that ADSCs induced epithelial-mesenchymal transition and migration of breast and pancreatic cancer cells in a FAM83H-AS1-dependent manner. And ADSC-induced FAM83H-AS1 increased unfolded protein response through AKT/XBP1 pathway.
Conclusion

In conclusion, our results indicated that ADSCs promoted breast and pancreatic cancer development via inducing cell proliferation and migration, as well as unfolded protein response through FAM83H-AS1.


Introduction
Long noncoding RNAs (lncRNA) are involved in cancer progression [1]. Their expression levels could be used as diagnostic or prognostic markers. Accumulating studies also con rmed the "seed and soil" relations between tumors and the surrounding microenvironment [2]. The tumor microenvironment consists of extracellular matrix and various mesenchymal cell types supported by a vascular network [3].
Tumor stromal is believed to exhibit many similarities in the in ammatory milieu produced by healing wounds, including angiogenesis, in ltration of broblasts and immune cells, and extensive remodeling of the extracellular matrix [4,5].
Endoplasmic reticulum (ER) functions as an essential synthesis and folding manufactory of secretory proteins. Hypoxia will cause the accumulation of misfolded protein in the ER, inducing ER stress. Unfolded protein response (UPR) is activated to clear misfolded proteins and restore normal physiological function [18,19]. Human X-Box binding protein 1 (XBP1) was reported to e ciently induce UPR [20,21].
Previous studies indicated that marrow fat cells induced ER stress in both tumor cells and the neighboring cells [22], And that XBP1 was induced in tumor cells cocultured with adipocyte, suggesting that ER stress in tumor cells was regulated by their microenvironment [23].
In the current study, we examined the role of ADSCs in cancer progression and explored the underlying mechanisms using a coculture cell system. Our data demonstrated that ADSCs enhanced lncRNA FAM83H-AS1 expression in cancer cells, and promoted cancer cells growth and migration via inducing UPR in a FAM83H-AS1-dependent manner.

Materials And Methods
Isolation and culture of human ADSCs Adipose tissues were obtained from healthy female donors, aged 18 to 30 years, undergoing an abdominal liposuction bariatric procedure with written informed consent. This study was approved by the Ethics Committee for Human and Animal Research of Wuhan University. All experiments were performed in accordance with Declaration of Helsinki and other recognized standards. After washing with phosphate buffered saline (PBS), these lipoaspirates were nely minced and enzymatically digested with 2 mg/mL type I collagenase at 37 °C for 40 min. The tissues were then centrifuged to remove buoyant adipocytes. The top layers were retrieved to obtain the stromal vascular fractions (SVFs) and cultured in Dulbecco's modi ed eagle medium (DMEM, Hyclone, USA) F12 supplemented with 10% fetal bovine serum (FBS) and 1% antibiotic/antimycotic. ADSCs were obtained after 3-4 passages of SVFs cultured in the corresponding medium as reports [24].

Identi cation of ADSCs
The expression of stromal markers on cell surface of expanded ADSCs were con rmed using immuno uorescence analysis. ADSCs cultured on 14 mm slides in 24-well plates were xed in 4% paraformaldehyde at room temperature for 30 minutes. After washing with PBS for 3 times, the cells were blocked with 10% goat serum and incubated with primary antibodies against CD31 (rat, Abcam, USA), CD45 (mouse, Santa Cruz, USA), CD90 (mouse, R&D, USA), and CD105 (rat, Santa Cruz, USA) at 4 °C overnight. After incubating with Alexa Fluor 488-and Cy3-conjugated secondary antibodies, 40,6diamidino-2-phenylindole (DAPI, Sigma-Aldrich, USA) was used to stain the nuclei. The images were examined with a confocal microscope system (Nikon C2+ Confocal Microscope, Japan), and quantitatively evaluated using Image-Pro Plus 6.0 software to measure mean densities (IOD/area). For adipogenic differentiation, cells were seeded in 6-well plates. When full con uent, cells were cultured in DMEM/Ham's F12 media supplemented with 10μg/ml transferrin, 0.85 μM insulin, 0.2nM triiodothyronine, 1μM dexamethasone, 500μM isobutylmethylxanthine for 2 weeks. For osteogenic differentiation, cells were seeded in 6-well plates and maintained in DMEM/Ham's F12 media supplemented with 0.1mM dexamethasone, 50mM ascorbate-2-phosphate and 10mM bglycerophosphate for 4 weeks. [25]. Differentiated ADSCs were stained with oil red for 30 minutes or alizarin for 30 minutes. The breast and pancreatic cancer cells were cultured with or without ADSCs for 3 days and then trypsinized. And they were seeded (2×10 3 cells/well) in 96-well plates and cultured in complete medium. Cell viability was measured using Cell Counting Kit-8 (CCK8) every 12 hours. Cancer cells were cultured with or without ADSCs for 48 hours. After washing with PBS for 3 times, the cells were stained with propidium iodide (Multisciences, China) and analyzed by ow cytometer (Beckman, USA). For the colony formation assay, tumor cells were seeded (1-1.5×10 3 cells/well) in the bottom chamber of 6-well plates and maintained in complete medium for 2 weeks, ADSCs were cultured in the upper chamber. The cells were xed with 4% paraformaldehyde for 2 hours, and stained with 1% crystal violet.
Xenograft tumor model BALB/c nude mice aged 3 weeks were obtained from Beijing HFK Bioscience Co., Ltd. in Beijing, China. 1X106 MDA-MB-468/ASPC1 cells with or without 1X106 ADSCs were injected to each mouse. The mice were maintained in a temperature and humidity-controlled and speci c pathogen-free environment in the laboratory animal facility of Zhongnan Hospital of Wuhan University. Tumor sizes were measured every 5 days until the end of the experiment. The experiments were performed under the protocols approved by ethnic committee of Zhongnan Hospital of Wuhan University.

Wound healing assay
The breast and pancreatic cancer cells were seeded in the bottom chamber of 6-well transwell plates with or without ADSCs cultured in the upper chamber. When full con uent, the cell layer was scratched with a 200 μl sterile pipette tip and washed with PBS. Images were acquired at different time points (0, 24, 48 h), and 3 independent experiments were conducted.

Modi ed Boyden chamber assay
The modi ed Boyden chamber assay was performed in 24-well transwell chamber systems (Corning, USA) with 8.0 µm pore size. The breast and pancreatic cancer cells were cultured with or without ADSCs for 3 days and then trypsinized. Cancer cells (5×10 4 cells/well) were seeded in the upper chamber insert and cultured in serum-free media. The lower chamber was lled with complete medium (600 μL, 10% FBS). After incubated at 37°C for 24 hours, the cells on the lower surface of the membrane were xed with 4% paraformaldehyde and stained with crystal violet. The membranes were placed under an inverted phase contrast microscope and imaged to count the migrated cells. Three independent experiments were conducted.

Statistical analysis
Student's t test and one-way ANOVA were used to compare 2 and more groups respectively. Multiple comparison with Bonferroni correction was performed when appropriate. A P value < 0.05 was considered as statistically signi cant and all tests were two-tailed. All statistical tests were performed with Prism 7.0 (GraphPad, USA).

Results
FAM83H-AS1 is up-regulated in breast cancer and pancreatic cancer samples and is associated with poorer prognosis Using the web-based tool Gene Expression Pro ling Interactive Analysis (GEPIA)[26], we found that lncRNA FAM83H-AS1 was upregulated in multiple tumors, including breast carcinoma, colon adenocarcinoma, lung cancer, pancreatic adenocarcinoma, prostate adenocarcinoma, rectum adenocarcinoma, stomach adenocarcinoma. The subsequent correlation assays indicated that the higher FAM83H-AS1 levels were only signi cantly correlated with worse prognosis in breast and pancreatic cancers (Figure 1).
ADSCs-induced FAM83H-AS1 promotes MDA-MA-468 and ASPC1 cell proliferation Previous studies indicated that ADSCs promoted breast and pancreatic cancer development [27,28]. To investigate whether FAM83H-AS1 was involved in the modulation of ADSCs on cancer growth, we downregulated FAM83H-AS1 in the breast and pancreatic cancer cells and examined cancer cell behaviors in vitro. All the 3 shRNAs signi cantly decreased FAM83H-AS1 expression ( Figure 3C, Figure  S1B), and shRNA-1 and shRNA-2 were selected for further analysis. The results of CCK-8 proliferation assays indicated that coculture with ADSCs increased cancer cell growth, and that de ciency of FAM83H-AS1 blocked ADSCs-induced proliferation ( Figure 3D, Figure S1C). Colony formation assay suggested similar results. Compared with control group, the breast and pancreatic cancer cells cultured with ADSCs had signi cant higher colony formation number. FAM83H-AS1-knockdown signi cantly decreased colony formation of cancer cells ( Figure 3E, Figure S1D). We investigated the role of ADSCs and FAM83H-AS1 in tumor growth by xenograft mice models. Our data showed that ADSCs-coculture signi cantly increased tumor growth and FAM83H-AS1 depletion by lentvirus based shRNA decelerated breast tumor growth ( Figure 3F). These results revealed that FAM83H-AS1 was involved in ADSCs-induced breast and pancreatic tumor development via promoting cancer cell growth and proliferation.
ADSCs regulate cancer cell cycle G1-to-S transition via FAM83H-AS1 Flow cytometry analysis was used to investigate the effects of ADSCs and FAM83H-AS1 on cell cycle of breast and pancreatic cancer cells. Coculture with ADSCs signi cantly decreased the percentage of G1 cells but increased that of cells at the S phase, indicating that ADSCs promoted G1-to-S transition of cancer cells. However, de ciency of FAM83H-AS1 abrogated ADSC-induced G1-to-S phase transition. In addition, coculture with ADSCs increased the protein levels of cell cycle promoter cyclinD1, CDK4 and CDK6 in cancer cells, while knockdown of FAM83H-AS1 reversed ADSC-induced cell cycle protein upregulation ( Figure 4, Figure S2).

ADSCs modulate cancer cell migration via FAM83H-AS1
To investigated the effects of ADSCs and FAM83H-AS1 on breast cancer and pancreatic cancer cell migration, MDA-MA-468 and ASPC1 cells were cocultured with ADCSs with or without FAM83H-AS1 knockdown, and applied for wound healing and modi ed Boyden chamber assays. The results showed that ADSC induced cancer cell migration, while de ciency of FAM83H-AS1 impeded this induction ( Figure  5A&B, Figure S3A&B). ADSCs were reported to promote epithelial to mesenchymal transition (EMT) in cancer cells [29,30]. The effects of FAM83H-AS1 on EMT of breast cancer and pancreatic cancer cells were further examined. Coculture with ADSCs resulted in a concurrent increase of N-cadherin, vimentin expression and a signi cant reduction of E-cadherin expression. FAM83H-AS1 knockdown reversed the effects of ADSCs on EMT markers ( Figure 5C, Figure S3C). Taken together, these results indicated that FAM83H-AS1 modulated EMT and was involved in ADSCs-induced cell migration.
In the present study, we isolated ADSCs from normal human adipose and investigated their roles in malignant biological behavior of breast and pancreatic cancer. We observed that lncRNA FAM83H-AS1 was upregulated in both breast and pancreatic cancer cells cocultured with ADSCs. Using the web-based tool GEPIA, we found that FAM83H-AS1 was upregulated in breast and pancreatic tumor tissues, and that elevated expression of FAM83H-AS1 was associated with poorer overall survival of breast and pancreatic cancer patients. Dysregulation of FAM83H-AS1 expression was also observed in colorectal carcinoma (CRC), and high FAM83HAS1 levels in CRC patients was signi cantly associated with advanced stage and poorer overall survival. Knockdown of FAM83H-AS1 dramatically inhibited the proliferation capability of CRC cells, and Notch 1 regulators reversed this effect mediated by FAM83H-AS1 [41]. Analysis of 461 lung adenocarcinomas (LUAD) and 156 normal lung tissues revealed that FAM83H-AS1 was overexpressed in lung cancer and signi cantly associated with worse survival. FAM83H-AS1 was reported to enhance lung cancer ce3ll proliferation, migration and invasion via MET/EGFR signaling pathway [34]. In our analysis, coculture with ADSCs increased cancer cell proliferation and migration, while FAM83H-AS1 knockdown blocked ADSCs-induced proliferation and migration of breast and pancreatic cancer cells. In addition, de ciency of FAM83H-AS1 signi cantly caused G1 phase arrest. Expression levels of cell cycle promoter cyclin D1, CDK4 and CDK6 upregulated by ADSCs were abrogated by FAM83H-AS1 knockdown. Previous studies showed that de ciency of FAM83H-AS1 signi cantly caused the arrest of U251 and U87 cells at G1 phase and reduced protein levels of CDK2, CDK4 and CDK6 through recruiting EZH2 to the promoter of CDKN1A [42]. ADSCs were also reported to promote EMT of tumor cells [29,30], we investigated whether ADSCs enhanced breast and pancreatic cancer cell migration through EMT.
Cocultured with ADSCs resulted in a concurrent increase in N-cadherin, vimentin expression and a signi cant reduction in E-cadherin expression, while knockdown of FAM83H-AS1 reversed these ADSCsinduced effects. Taken together, these data indicated that FAM83H-AS1 regulated EMT and was involved in ADSCs-induced cell migration.
Gene set enrichment analysis of differentially expressed genes with FAM83H-AS1 knockdown in pancreatic cancer suggested that UNFOLDED PROTEIN RESPONSE, HYPOXIA, P53 PATHWAY, TNFA SIGNALING VIA NFKB and KRAS SIGNALING UP were the signi cantly enriched pathways [31]. In our analysis, ADSCs treatment signi cantly upregulated IRE1 phosphorylation and XBP1 splicing in cancer cells in a FAM83H-AS1-dependent manner. Previous studies revealed that AKT pathway was regulated by FAM83H-AS1 and IRE1/XBP1, ADSCs from primary breast cancer tissue promoted cancer cell proliferation via AKT [14,[32][33][34][35]. Our results demonstrated that coculture with ADSCs phosphorylated AKT, and that de ciency of FAM83H-AS1 decreased ADSCs-induced AKT phosphorylation. In addition, IRE1 inhibitor suppressed XBP1s, as well as AKT phosphorylation, induced by ADSCs. Taken together, our results indicated that tumor progression regulated by ADSCs was modulated by FAM83H-AS1 via regulating AKT through XBP1s.
Cancer cells and their surrounding microenvironment constantly communicate with each other during tumor development. In the present study, we found that ADSCs upregulated lncRNA FAM83H-AS1 expression in breast and pancreatic cancer cells, promoted cell proliferation and migration, and induced the unfolded protein response via activating IRE1-XBP1 pathway. FAM83H-AS1 was reported to epidemically silence CDKN1A via recruiting EZH2 to its promoter and stabilize HuR protein via directly binding. The impact of ADSCs on tumor microenvironment and the direct mechanism of FAM83H-AS1 on IRE1-XBP1 pathway still need to be veri ed in further studies.

Conclusion
In conclusion, our present study demonstrated that ADSCs promoted breast and pancreatic cancer cell proliferation and migration, and induced the unfolded protein response in a FAM83H-AS1-dependent manner.

Declarations
Ethics approval and consent to participate YG and GW conceived and designed the study. JT, QC, DZ, and XL performed the experiments. JT and YG wrote the paper. GW reviewed and edited the manuscript. All authors read and approved the manuscript.
Supplementary Legends Figure