COL5A2 drives regorafenib resistance-induced metastatic phenotype via reducing LIFR expression in hepatocellular carcinoma.

Systemic therapies, the ultimate strategies for patients with advanced hepatocellular carcinoma (HCC), are suffering from serious clinical challenges, such as the occurrence and development of drug resistance. Treatment resistance aggravates tumor progression partly by inducing tumor metastasis. Regorafenib-resistant HCC cells exhibit a highly striking metastatic phenotype, but the detailed mechanisms underlying these aggressive behaviors remain elusive. Here, we conduct transcriptome sequencing analysis to identify COL5A2 as a crucial driver of the metastatic characteristics of regorafenib-resistant HCC cells. COL5A2 is aberrantly highly expressed in resistant cells, and its genetic depletion significantly suppresses proliferation, migration, invasion, vasculogenic mimicry (VM) formation and lung metastasis in vitro and in vivo, concomitant with the downregulation of VE-cadherin, EphA2, Twist1, p-p38 and p-STAT3 expressions. LIFR is confirmed to be an essential downstream molecule of COL5A2, and its expression is observably elevated by COL5A2 depletion. Ectopic overexpression of LIFR drastically attenuates the proliferation, migration, invasion and VM of regorafenib-resistant cells and represses the expressions of VM-related molecules and the activation of p38/STAT3 signaling pathway. Interestingly, rescue experiments show that the inhibition of the above aggressive features of resistant cells by COL5A2 loss is clearly alleviated by silencing of LIFR. Collectively, our results reveal that COL5A2 promotes the ability of regorafenib-resistant HCC cells to acquire a metastatic phenotype by attenuating LIFR expression and suggest that therapeutic regimens targeting the COL5A2/LIFR axis may be beneficial for HCC patients with therapeutic resistance.

Genetic analysis of ANXA5 haplotype and its effect on recurrent pregnancy loss.

Recurrent pregnancy loss (RPL) is often associated with dysregulated Annexin A5 (ANXA5) expression. Moreover, the variants of Anxa5, a protein that is enriched in the placenta to prevent coagulation, have been reported to affect RPL risks. The haplotypes M1 [including single nucleotide polymorphisms (SNPs) 1A/C and 27T/C] and M2 (including SNPs 19G/A, 1A/C, 27T/C and 76G/A) of ANXA5 were also reported to affect RPL risks. The present study aimed to investigate the association between the haplotype located in the promoter region of ANXA5 and the risk of RPL. Patients with RPL (n=235) or intrauterine fetus death (IUFD; n=154), as well as healthy control subjects (n=375) were enrolled in the current research. Their haplotypes of ANXA5 were determined using genotyping, and the association between ANXA5 haplotypes and the risk of RPL was accordingly analyzed. A luciferase assay was conducted to investigate the haplotype responsible for ANXA5 activity. Reverse transcription­quantitative PCR, western blot analysis, immunohistochemistry and ELISA were performed to assess the expression level and activity of ANXA5 in patients with RPL. Consequently, the majority (n=214) of patients with RPL had a history of early RPL, whereas 31 patients with RPL had a history of both early and late RPL episodes. A significant difference was found between cases and controls in terms of gravidity and parity, whereas no significant differences were found in terms of age. The percentage of patients with RPL carrying the M2 haplotype of ANXA5 was significantly higher compared with that in control subjects, indicating that the M2 haplotype of ANXA5 was an independent risk of RPL as it influenced the transcription efficiency of ANXA5 promoter. In patients with RPL, ANXA5 activity was suppressed and the mRNA and protein expression levels of Anxa5 were decreased. Thus, the ANXA5 M2 haplotype may be an independent risk factor of RPL by affecting Anxa5 activity.