CircRNA-DAPK1 promotes vascular cell pyroptosis in diabetes by regulating the circ-DAPK1/miR-4454/TXNIP axis.

Circular Circ-DAPK1 was selected from circular RNA sequencing data of HUVECs treated with high glucose medium and normal medium. RT-qPCR was used to determine the expression of circ-DAPK1 in vivo and in vitro. Dual luciferase reporter assay, uorescence in situ hybridization (FISH) and RNA immunoprecipitation (RIP) were performed to prove the interaction of circ-DAPK1, miRNA-4454 and thioredoxin-interactingprotein (TXNIP). Adeno-associated virus (AAV) was injected intravenously to establish mouse models. PI staining, western-blot and transmission electron microscopy (TEM) analyses were performed to identify the role of circ-DAPK1 in promoting pyroptosis.


Background
Pyroptosis is de ned as a kind of programmed cell death that differs from necrosis and apoptosis [1]. Previous studies have demonstrated that canonical and noncanonical signaling pathways are involved in pyroptosis. In the canonical pathway, nucleotide-binding oligomerization domain-like receptor family 3 (NLRP3) senses external stimuli. Then, the adapter apoptosis-associated speck-like protein (ASC) assembles with NLRP3 to activate the in ammatory protease caspase-1. Subsequently, caspase-1 cleaves the proin ammatory cytokines interleukin-18 (IL-18) and pro-interleukin-1β(IL-1β) to initiate the downstream immune response [2]. In the noncanonical pathway, caspase-4/5/11 can sense bacterial lipopolysaccharide (LPS) to form the in ammasome, which cleaves the N-terminus of GSDMD to cause the secretion of in ammatory mediators that induce in ammatory reactions [3]. Therefore, in transmission electron microscopy (TEM) imaging, pyroptotic cells can be identi ed by the presence of multiple holes in the membranes [4]. Circular RNAs (circRNAs) are a subclass of endogenous noncoding RNAs that are characterized by a closed-loop structure without a 3'-5' polyadenylated tail or polarity. CircRNAs are conserved among different species including in human vascular cells [5]. Several studies have reported their vital role in cancers, osteoarthritis, atherosclerosis, vascular dysfunction and many other diseases [3,[6][7][8][9].
Diabetes has been shown to accelerate the progression of atherosclerosis, which promotes the progression of coronary artery disease [10,11] and arteriosclerosis obliterans [12]. However, our understanding of the roles of insulin resistance and hyperglycemia in atherosclerosis is relatively incomplete [13,14]. Recently, some studies have discovered the factors involved in pyroptosis-induced atherosclerosis, such as shear stress, mircoRNAs (miRNAs) and nicotine [15][16][17]. However, the underlying molecular mechanism and cellular pathways involved in the link between circRNAs and vascular cell pyroptosis in diabetes remain to be clari ed.
In this study, we identi ed a new circRNA, circRNA-021774 (also named circ-DAPK1) and discovered its contributions to vascular cell pyroptosis in diabetes in vitro and in vivo. We also aimed to demonstrate the underlying mechanisms of circ-DAPK1 in promoting pyroptosis.

Ethics statement
All animal studies were approved by the Ethical Review Committee of the First A liated Hospital, Zhejiang University School of Medicine.

Cell culture and transfection
Human umbilical vein endothelial cells (HUVECs) were cultured in DMEM medium (Biological Industries, Israel) with 10% fetal bovine serum and 1% penicillin/streptomycin solution (Biological Industries, Israel) at 37°C under 5% CO 2 atmosphere. HUVECs were treated with 33 mM glucose in the high glucose (HG) group and 5 mM glucose in the control normal glucose (NG) group.

Western-blot
The cells were lysed on ice for 60 min in RIPA buffer (Beyotime, China) supplemented with phosphatase and protease inhibitors. The supernatant was collected with centrifugation (12,000 ×g, 4°C). The concentrations were analyzed with a Pierce BCA protein assay kit (Thermo Scienti c, USA).

RNA immunoprecipitation assay (RIP)
The RIP assay was performed using an Imprint RNA Immunoprecipitation Kit (Merck, USA). AGO2 and vector plasmids were transfected into HEK293 cells. Then, 1.5×10 7 cells were harvested and lysed in 100 µL RIP lysis buffer containing an RNase and protease inhibitor cocktail. Then, the cell lysates were incubated with 4 µL AGO2 antibody(Abcam, UK) or IgG-coated beads. After incubation, the cell lysates were rotated at 4°C for 24 h. Immunoprecipitated RNA was extracted with an RNeasy MinElute Cleanup Kit (Qiagen, USA) and reverse transcribed using HiScript Reverse Transcriptase (Vazyme, China). The expression of circRNA-DAPK1 and miRNA-4454 was detected by PCR.

Enzyme-linked immunosorbent assay (ELISA)
Mouse serum concentrations of interleukin-1β and IL-18 were determined by ELISA kits (Neobioscience, China) according to the manufacturer's instructions.

Dual luciferase reporter assay
Luciferase reporters containing wild-type (WT) or mutated circ-DAPK1 plasmid (Mut) were constructed using the psi-CHECK2 vector (Promega, USA). The luciferase vector (2 µg) containing the pcDNA3.1 circ-DAPK1 plasmid was cotransfected with miR-4454 mimics, miR-4454 inhibitor or negative control (scramble sequence of miR-4454) into human embryonic kidney 293 (HEK293) cells using Lipofectamine™ 2000 (Invitrogen, USA) with 10 ng of a Renilla luciferase reporter used as an internal control. After 48 h, the cells were collected and lysed. Then, luciferase activities were analyzed by the Dual-Luciferase Reporter Assay System (Promega, USA) following the manufacturer's instructions.
Establishment of the animal model C57BL/KsJ-db/db mice (Cavens Lab Animal Co.,Ltd, China) were used to construct the animal models. A total of 7×10 7 VG adeno-associated virus (AAV) was injected intravenously in each group. The animals were sacri ced 4 weeks after AAV injection, serum and aorta of mice were harvested for further detection ( Fig. 7A).
The stably infected cells were stained with 5 µL Hoechst 33342 at 37 o C in the dark for 10 min, and then incubated with 3 µL PI solution at room temperature for 15 min. Finally, the images were captured by uorescence microscopy (Olympus BX53, Japan).
Transmission electron microscopy (TEM) TEM imaging procedures were performed according to our previous work [18]. HUVEC cell lines were used to construct the cell model.

Immuno uorescence (IF) staining
For immuno uorescence analysis, 4×10 4 HUVECs were plated in a confocal slide and cultured overnight.
The cells were xed with 4% paraformaldehyde for 15 min and then permeabilized with 0.5% Triton X-100 for 20min. After permeabilization, the cells were blocked with goat serum albumin for 30 min at room temperature. Subsequently, the primary antibodies against caspase-1 (1:100, Abcam) and ASC (1:100, Affbiotech) were added to the slides for 24 h incubation at 4°C. The next day, the cells were incubated with the secondary antibodies for 1 h. DAPI was utilized to stain the nuclei for 5 min at room temperature. Finally, images were captured by uorescence microscopy (Olympus BX53, Japan).

RNase R resistance assay
Total RNA (5 µg) from HUVECs were treated with RNase R(3U/µg, Lucigen) previously and incubated at 37°C for 60 min. The RNA was reverse transcribed with primers and analyzed by PCR.

Statistical analysis
All data of this study are presented as the mean ± SD. Differences between two groups were analyzed by Student's t-test unless indicated otherwise. P values less than 0.05 were considered statistically signi cant.

Results
Circ-DAPK1 expression is relatively high in HUVECs cultured in HG medium and is located in the cytoplasm.
For a circRNA sequencing, we constructed a cell model using HUVECs. Three samples cultured with HG medium (HG group) for 72 h and their controls (NG group) were prepared to conduct RNA-sEq. The results showed that 46 circRNAs were upregulated while 37 circRNAs were downregulated in the HG group (Fig. 1A, B). Among them, circ-DAPK1 was signi cantly overexpressed in the HG group (Fig. 1B). To con rm the expression of circ-DAPK1 under HG condition, we used RT-qPCR to determine its abundance in the HG and NG groups. Consistent with the results of the RNA-seq analysis, the expression of circ-DAPK1 was higher in the HG group (Fig. 1C). As shown in previous reports, circRNAs are formed by transsplicing [19]. Subsequently, we identi ed the loop structure of circ-DAPK1. According to the circBase database, circ-DAPK1 is produced from the DAPK1 gene locus and contains exon 23-24 (Fig. 1E). As head-to-tail splicing could be the result of trans-splicing and genomic rearrangements, convergent primers for DAPK1 mRNA and divergent primers for circ-DAPK1 were utilized to distinguish these two molecules. PCR results indicated that circ-DAPK1 was ampli ed from cDNA but not gDNA (Fig. 1F). In the presence of RNase R, the linear form of DAPK1 decreased sharply (the right two lanes), while circ-DAPK1 remained stable (the left two lanes) in the Northern-blot analysis (Fig. 1G). Furthermore, FISH experiments revealed that circ-DAPK1 was located mainly in the cytoplasm (Fig. 1D).
To explore the role of circ-DAPK1 in regulating pyroptosis, we transfected the HUVECs with circ-DAPK1speci c siRNA (circ-KD) or overexpression plasmid (circ-OE) respectively and determined the e ciencies of circ-DAPK1 knockdown and overexpression by RT-qPCR ( Fig. 2A). PI staining showed that knockdown of circ-DAPK1 signi cantly decreased the pyroptosis of HUVECs cultured in HG medium, and circ-DAPK1 overexpression increased the PI uptake in NG medium (Fig. 2B, C). Western-blot analysis showed that circ-DAPK1 suppressed the expression of NLRP3, ASC, cleaved caspase-1 (c-caspase-1), GSDMD, IL-18, and IL-1β. In contrast, circ-DAPK1 downregulated the expression of the above pyroptosis-related proteins (Fig. 2D). Moreover, TEM imaging revealed that many pores formed in the cell membranes of the HGtreated group and circ-OE group (Fig. 2E). Since ASC and c-caspase-1 are key regulators in the process of pyroptosis, we performed immuno uorescence staining in HUVECs. In the HG groups, circ-KD inhibited the uorescence intensity, and consistent with this nding, circ-OE enhanced the uorescence intensity in the NG groups (Fig. 2F, G, H, I).
Because circ-DAPK1 is stable in the cytoplasm, we further investigated the capability of circ-DAPK1 to bind miRNAs. We performed a miRNA expression pro le analysis and compared the results with two databases (miRanda and TargetScan). Ten miRNAs were predicted as the potential targets of circ-DAPK1 (Fig. 3A). RT-qPCR showed that only miRNA-4454 increased signi cantly after circ-DAPK1 knockdown (Fig. 3B). To prove the correlation between miRNA-4454 and circ-DAPK1, HEK293 cells were transfected with an AGO2 overexpression plasmid or empty vector and subjected to a RIP assay. By PCR analysis, we found that circ-DAPK1 and miRNA-4454 pulled down by the anti-AGO2 antibody were enriched in the AGO2 overexpression group (Fig. 3C). Furthermore, in the HG-treated group, miRNA-4454 expression was decreased (Fig. 3D). And immuno uorescence staining showed that miRNA-4454 was predominately located in the cytoplasm (Fig. 3E).
We next explored the function of miRNA-4454 in HUVECs via a PI staining assay. As shown in Fig. 3F and 3G, the inhibition of miRNA-4454 resulted in higher pyroptosis rates of HUVECs in the HG group. However, miRNA-4454 overexpression reduced the percentage of PI + cells in the NG group (Fig. 3G). TEM revealed that miRNA-4454 could attenuate the pyroptosis of HUVECs (Fig. 3H). Furthermore, miRNA-4454 overexpression decreased the levels of NLRP3, ASC, c-caspase-1, N-GSDMD, IL-18 and IL-1β (Fig. 3I).
TXNIP is a target gene of miRNA-4454.
To determine which target genes were regulated by miRNA-4454 in HUVECs, we screened the transcriptomes of the HG and NG groups (Fig. 4A). Among the identi ed differentially expressed genes, TXNIP was higher in the HG groups. For validate its expression, RT-qPCR was performed to demonstrate that TXNIP expression was higher in the HG groups (Fig. 4B). Based on RT-qPCR and western-blot analysis, TXNIP expression substantially increased under miRNA-4454 inhibitor treatment (Fig. 4C), while miRNA-4454 overexpression reduced the expression of TXNIP (Fig. 4D). According to the competing endogenous RNA (ceRNA) mechanism, the expression of circ-DAPK1 should be positively correlated with the expression of target genes. The RT-qPCR results indicated that circ-DAPK1 silencing decreased the mRNA expression of TXNIP in the HG group, while circ-DAPK1 overexpression increased the abundance of TXNIP in the NG group (Fig. 4E). Luciferase assays con rmed that miRNA-4454 mimics inhibited luciferase activity in cells transfected with the wildtype(WT) 5'-UTR of TXNIP but not the mutant (Mut) 5'-UTR of TXNIP (Fig. 4F). In addition, FISH imaging revealed that both miRNA-4454 and TXNIP mainly were located in the cytoplasm (Fig. 4G). These results suggest that miRNA-4454 targets TXNIP and that TXNIP seems to be a downstream regulator of circ-DAPK1 and miRNA-4454.
As TXNIP is a downstream factor of circ-DAPK1 and miRNA-4454, we next investigated the relationship of these three molecules. In the RT-qPCR assay, circ-DAPK1 inhibited the expression of miRNA-4454 (Fig. 5A). FISH assays revealed that under HG conditions, both circ-DAPK1 and miRNA-4454 were located in the cytoplasm (Fig. 5B). This result suggested that circ-DAPK1 might regulate miRNA-4454 by acting as a ceRNA. As expected, a luciferase assay con rmed that circ-DAPK1 could bind to miRNA-4454 (Fig. 5C).
We next performed a ceRNA network analysis by circRNA-seq, miRNA pro ling and RNA-sEq. The bioinformatics analysis of the ceRNA network revealed that circ-DAPK1 could regulate miRNA-4454/TXNIP (Fig. 5D). In other words, the circ-DAPK1/miRNA-4454/TXNIP axis exists in the HUVECs. To verify the role of this signaling axis in the process of pyroptosis, we designed dual luciferase reporter and rescue assays. In Fig. 5E, comparing the TXNIP 5'-UTR WT + miRNA-4454 OE group with the TXNIP 5'-UTR WT + miRNA-4454 OE + circ-DAPK1 OE group, circ-DAPK1 overexpression enhanced the luciferase activity.
The circ-DAPK1-regulated mechanism exists in a diabetic mouse model.
To further investigate the role of circ-DAPK1 under HG conditions, we constructed a diabetes model with C57BL/KsJ-db/db mice. After the mice were fed a high-fat diet for 4 weeks, the tail vein blood was collected for blood glucose testing. In the db/db group, blood glucose was higher than that in the db/m group (Fig. 6A). By FISH, circ-DAPK1 was enriched in the db/db group. Contrarily, we found miRNA-4454 expressed higher in db/m group (Fig. 6B). In contrast, by RT-qPCR analysis, the expression of circ-DAPK1 and TXNIP was detected higher, while miRNA-4454 expression was lower, in the db/db group (Fig. 6C, D, E). Western-blot analysis revealed that c-caspase-1, NLRP3 and TXNIP were highly expressed in the db/db group (Fig. 6F). These results suggested that the circ-DAPK1/miRNA-4454/TXNIP axis may exist in vivo.
We next continued to investigate whether the circ-DAPK1/miRNA-4454/TXNIP axis aggravates pyroptosis in vivo. Therefore, we injected circ-DAPK1, miRNA-4454 or TXNIP AAV intravenously (Fig. 7A). As shown in Fig. 7B, AAV had no in uence on the weight of mouse model. RT-qPCR was performed to determine the expression of the circ-DAPK1/miRNA-4454/TXNIP axis in db/db mice, and the results con rmed that circ-DAPK1 knockdown and miRNA-4454 overexpression transfection markedly decreased the level of TXNIP (Fig. 7C). This result suggested that the circ-DAPK1/miRNA-4454/TXNIP axis could function in vivo. Blood glucose tests illustrated that circ-DAPK1 knockdown, miRNA-4454 overexpression and TXNIP knockdown reduced the glucose levels of db/db mice (Fig. 7D). Then, immuno uorescence staining was performed to detect the ASC and caspase-1 in the arteries of each group. The results con rmed that in the circ-KD, miRNA-4454 OE and TXNIP-KD groups, the uorescence intensity of ASC and caspase-1 dramatically declined (Fig. 7E, F). Moreover, the serum and aortas of db/db mice were collected for ELISA and western-blot analysis respectively. The experiments demonstrated that circ-DAPK1 knockdown, miRNA-4454 overexpression and TXNIP knockdown reduced the expression of NLRP3, ASC, TXNIP, ccaspase-1, N-GSDMD, IL-18 and IL-1β in vivo (Fig. 7G, H, I, J, K, L, M, N). These results veri ed that the circ-DAPK1/miRNA-4454/TXNIP axis promotes pyroptosis in vivo.

Discussion
At present, pyroptosis has been reported to induce atherosclerosis [20]. Many molecules and mechanisms, such as circRNAs and ceRNAs, are involved in the process of pyroptosis [21,22]. CircRNAs are a type of noncoding RNA that has been reported to mediate different biological and pathological processes, including vascular diseases [23,24]. However, their functions in the progression of vascular cell pyroptosis and roles in atherosclerosis remain to be clearly de ned. Therefore, understanding their mechanisms in the regulation of pyroptosis may provide prospects for atherosclerosis. We identi ed a potential circRNA -circ-DAPK1 by circRNA sequencing, but the role of circ-DAPK1 in vascular cell pyroptosis in diabetes has not been clari ed. In this study, we focused on the biological role and mechanism of circ-DAPK1.
According to frequent reports, circRNAs can modulate miRNAs by acting as sponges in what we called ceRNA networks [23]. Some well-known circRNAs can mediate biological functions by interacting with miRNAs. CircITCH acts as a sponge of miR-330-5p to upregulate SIRT6, Survivin and SERCA2a expression, thus alleviating doxorubicin-induced cardiomyocyte dysfunction [25].In addition, circ-ZNF532 could induce vascular dysfunction through miR-29a-3p [26]. In our study, we identi ed circ-DAPK1 as a sponge of miRNA-4454 by dual luciferase reporter and RIP assays. Moreover, under HG conditions, circ-DAPK1 knockdown enhanced the expression of miRNA-4454. These results were mirrored by in vivo experiments. Therefore, we propose a mechanism by which circ-DAPK1 act as a sponge of miRNA-4454 to promote vascular cell pyroptosis.
TXNIP is an important factor in various diseases, such as cardiovascular diseases, cancer and diabetes [27]. In recent years, TXNIP has been reported to mediate the process of pyroptosis. According to previous studies, TXNIP directly interacts with NLRP3 to promote cell pyroptosis [28,29]. In our study, we found that circ-DAPK1 acts as a sponge of miRNA-4454 to regulate the expression of TXNIP. Moreover, western-blot analysis revealed that circ-DAPK1 could activate the pyroptosis signaling pathway. In other words, the circ-DAPK1/miRNA-4454/TXNIP signaling axis could promote the development of vascular cell pyroptosis.
In this study, we designed in vitro and in vivo experiments to identify the role of circ-DAPK1 in vascular pyroptosis in diabetes. We will continue to explore its status in clinical specimens and its function in the procession of atherosclerosis in our future work.

Conclusions
In summary, we have demonstrated the vital role of the circ-DAPK1/miRNA-4454/TXNIP axis during the development of cell pyroptosis in diabetes. This signaling axis may serve as a therapeutic target for diabetic vascular diseases.

Declarations
Ethics approval and consent to participate All animal studies were approved by the Ethical Review Committee of the First A liated Hospital, Zhejiang University School of Medicine.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.