lncRNA HOTAIRM1 Activated by HOXA4 Drives HUVEC Proliferation Through Direct Interaction with Protein Partner HSPA5

This study was approved by the Ethics Committee of Guizhou Provincial People’s Hospital. The animal experimental protocol was approved by the Animal Ethics Committee of Guizhou Provincial People’s Hospital (License number: 2022–038).

Male ApoE mice in the AS group were fed a high-fat diet (21% fat and 0.15% cholesterol) at 4 weeks of age. Male ApoE mice in the control group were fed a normal diet (4% fat and 0% cholesterol). After 12 weeks, the mice were sacrificed, and the aortas were removed for further experiments.

HUVECs were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA, AC337632). The cells were cultured in an F-12 K medium (Invitrogen, Carlsbad, CA, USA, 21127022) containing 10% FBS (Gibco) and 1% penicillin–streptomycin (Thermo Fisher Scientific, 15140122). All cells were grown at 37 ℃ in a humidified atmosphere containing 5% CO2. For ox-LDL induction experiments, HUVECs were induced with ox-LDL (100 μg/mL, UnionBiol, Beijing, China, jk-002) for 24 h when the cells reached 70–80% confluence. HOXA4 and HSPA5 in HUVECs were knocked down using small interfering RNAs (siRNAs), and HOTAIRM1 was silenced using antisense oligonucleotides (ASOs). The ASOs and siRNAs were synthesized by GenePharma (Shanghai, China). The pcDNA3.1 vectors from Invitrogen were used to construct pcDNA3.1/HOXA4, pcDNA3.1/HOTAIRM1, and pcDNA3.1/HSPA5, with the empty vector as the negative control. Lipofectamine 3000 (Invitrogen, L3000015) was used for the transfection experiments.

Total RNA was extracted from the arterial tissue and HUVECs using TRIzol reagent according to the manufacturer’s protocol (Thermo Fisher Scientific, 12183555), and cDNA was acquired following reverse transcription. Quantitative real-time polymerase chain reaction (RT-qPCR) was performed using the SYBR Green Master Mix (Takara, Dalian, China, RR036A). The results were normalized to the expression of β-actin, and data were analyzed based on the 2 − ΔΔCt method. Supplementary File 1 lists the forward and reverse primers.

Total cellular protein was extracted using RIPA lysis buffer (Beyotime, Shanghai, China, P0013B) and subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Separated proteins were transferred to PVDF membranes, treated with 5% skim milk at room temperature for 2 h, and then incubated with corresponding primary antibodies against HOXA4 (Abcam, 131049), HSPA5 (Abcam, ab108615), PCNA (CST, 13110), cyclin D1 (Abcam, ab16663), CKD4 (Abcam, ab108357), caspase-3 (Abcam, ab32351), caspase-9 (Abcam, ab32539), Bax (Abcam, ab32503), Bcl-2 (Abcam, ab185002), and GAPDH (Abcam, ab181602) overnight at 4 °C, followed by conjugation with horseradish peroxidase at 37 °C for 2 h. Band density was visualized using an ECL kit (Invitrogen, Carlsbad, CA, USA, WP20005).

Tissues were stained with filtered Oil Red O solution (Sigma). Oil red staining was visualized under a microscope. The root and aortic arch bifurcation vessels were immediately fixed in 4% neutral formaldehyde solution, dehydrated by soaking in 75% distilled water, and immersed in formaldehyde solution for H&E staining.

For the Cell Counting Kit-8 (CCK-8) assay, to detect the viability of HUVECs, transfected-HUVECs were added to CCK-8 solution (Dojindo, KMJ, Japan, CK04) at the indicated time and incubated for 2 h at 37 °C; the absorbance was determined at the wavelength of 450 nm. For the EdU assay, HUVECs were incubated in a 24-well plate for transfection for 72 h and subsequently added to 100 µL of EdU (RIBOBIO, Shanghai, China, C10310-1). A fluorescence microscope was used to visualize the cells (Olympus, Tokyo, Japan). For cell cycle assay, after transfection for 72 h, cells were digested and washed with PBS and then fixed in 70% ice-cold ethanol at 4 °C overnight. The cells were treated with 500 µL of RNase A, stained with 50 μg/mL propidium iodide (PI) for 15 min, and examined by flow cytometry.

For apoptosis analysis, cells were double-stained with 7-AAD and PE per the manufacturer’s instructions (559763; PE Annexin V Apoptosis Detection Kit I, BD, USA, 559763). GFP-positive cells were assessed by flow cytometry (FACSCantoII, 338960; BD Biosciences, San Jose, CA, USA, 338960). Terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) staining was carried out to test cell apoptosis. Briefly, the cells were fixed using 4% formaldehyde, followed by staining with the TUNEL kit according to the manufacturer’s instructions (Vazyme, TUNEL Bright-Red Apoptosis Detection Kit, A113). The TUNEL-positive cells were counted using a fluorescence microscope.

Chromatin immunoprecipitation (ChIP) analysis was conducted using a Millipore ChIP Assay Kit (Millipore, MA, USA, 17–295) according to the manufacturer’s instructions. HUVECs were cross-linked with formaldehyde and sonicated at 200–1000 bp. Sonicated lysates were cleared and incubated overnight at 4 °C with magnetic beads coupled with HOXA4 antibody (Santa Cruz, sc-398426). RT-qPCR was used to analyze the precipitated chromatin DNA.

To dissect the transcriptional regulation of HOXA4 on HOTAIRM1, the sequence of the wild-type or indicated site-mutated HOTAIRM1 promoter (WT, Site 1-MUT, Site 2-MUT, Site 3-MUT, Site 4-MUT, and MUT) was synthesized and cloned into the pGL3 vector to construct the corresponding reporter plasmids. Thereafter, the indicated recombinant reporter was co-transfected with an empty vector or a HOXA4 plasmid into HUVECs. After 48 h, the luciferase activity in each group was measured using a Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA).

A fluorescent in situ Hybridization (FISH) Kit was used to analyze HOTAIRM1 subcellular location (GenePharma, Suzhou, China). Cy3-labeled HOTAIRM1 probe was also synthesized. Images were acquired using a confocal microscope (Olympus). GAPDH and U6 served as controls for cytoplasmic and nuclear RNA, respectively.

Biotin-labeled sense and antisense HOTAIRM1 sequences were obtained using the Transcript Aid T7 High-Yield Transcription Kit (Thermo Scientific, K0441). A MEGAclearTM Kit (Thermo Scientific) was used to recycle the sequences per the manufacturer’s instructions. The sequences were incubated with cell lysates for 4 h at 4 °C, and then the biotin-labeled RNAs and their binding protein partner were pulled down by streptavidin magnetic beads (Thermo, USA, 20164) at 4 °C overnight. Proteins were separated by electrophoresis and visualized using a Coomassie Blue Staining Kit (Beyotime, China, P0017A). Different bands between sense and antisense HOTAIRM1 were identified using mass spectrometry (Thermo Scientific, Bremen, Germany, 24600).

An RNA immunoprecipitation (RIP) assay was used to determine whether HOTAIRM1 interacted with HOXA4/HSPA5 using the Magna RIP RNA-binding protein immunoprecipitation kit (Millipore, Billerica, MA, USA, 17–700). Cells were lysed with RIP buffer, and the cell extract was mixed with agarose beads carrying anti-HOXA4/HSPA5 or anti-lgG antibodies. Enrichment of the coprecipitated RNAs was detected by RT-qPCR.

Statistical data from each group were analyzed using GraphPad 6.0. Data are expressed as mean ± standard deviation (SD). Intergroup differences were assessed using Student’s t-test, and multivariate statistical analysis was performed using analysis of variance (ANOVA). Statistical significance was considered at p < 0.05.

As a first step in elucidating the biological function of HOTAIRM1 in AS, ox-LDL was used to simulate a cell model of AS. To screen the appropriate ox-LDL concentration, the viability and apoptosis were measured in HUVECs stimulated with different ox-LDL concentrations (0, 25, 50, 75, 100, and 125 μg/mL). The results indicated that the introduction of ox-LDL impaired HUVEC viability and contributed to apoptosis compared to the untreated group (Fig. 1a, b). Ultimately, 100 μg/mL was selected for subsequent cell experiments. RT-qPCR was performed to examine whether the presence of ox-LDL affected HOTAIRM1 expression. Ox-LDL enhanced HOTAIRM1 expression in HUVECs in a dose- and time-dependent manner (Fig. 1c, d). In addition, Hotairm1 expression was up-regulated in the aorta of HFD-fed ApoE − / − mice (Fig. 1e). Oil red O and H&E staining represented that the area of arterial tissue damage in the ApoE − / − mice fed with high-fat-diet (HFD) was increased compared with that in the ApoE − / − mice fed with normal diet (ND) (Fig. 1f, g). These findings suggest that HOTAIRM1 is possibly a regulator in the biological regulation of HUVECs.

Next, we sought to elucidate the biological role of HOTAIRM1 in HUVECs. Due to ox-LDL can induce HOTAIRM1 expression, ASOs that mediate HOTAIRM1 knockdown were screened; therefore, three individual ASOs were used to knock down their expression, and the silencing efficacy was evaluated by qPCR (Fig. 2a). In order to explore the effects of HOTAIRM1 on the proliferation and apoptosis of HUVEC, firstly, EdU and CCK-8 assays revealed that HOTAIRM1 knockdown remarkably reduced HUVEC proliferation irrespective of ox‐LDL treatment (Fig. 2b, c), and later, flow cytometry assay was performed to determine whether HOTAIRM1 depletion dampened cell proliferation by altering cell cycle progression. The results showed that the cell cycle was dramatically stalled at the G1-G0 phase in HOTAIRM1-silenced HUVECs (Fig. 2d), indicating that HOTAIRM1 deficiency impaired HUVEC proliferation, partly by regulating cell cycle progression. Furthermore, the silencing of HOTAIRM1 strikingly expedited HUVEC apoptosis (Fig. 2e, g). Finally, western blotting results revealed that the expression of cell cycle- and proliferation-related proteins, such as cyclin D1, CKD4, and PCNA, dramatically reduced upon HOTAIRM1 knockdown with a concurrent increase in apoptosis markers, including caspase-3, caspase-9, and Bax (Fig. 2F, H). Thus, HOTAIRM1 may play an important role in modulating HUVEC proliferation and apoptosis.

Mounting evidence suggests that the transcription of several lncRNAs can be regulated by transcription factors (TFs) via direct binding. To tease out potential upstream mediators exquisitely governing HOTAIRM1 in HUVECs, we conducted a motif analysis in the HOTAIRM1 promoter region using JASPAR (http://​jaspar.​binf.​ku.​dk/​). The result implicated that HOXA4 might be a potential TF with probable binding motif “ATTA” (Fig. 3a), indicating that HOTAIRM1 may be modulated by HOXA4 at the transcriptional level. Because HOXA4 appeared to harbor numerous seemingly binding sites predicted by JASPAR in the HOTAIRM1 promoter region (Supplementary File 2), we constructed different HOTAIRM1 promoter truncations to identify the core promoter binding region. The HOTAIRM1 promoter region containing fragments of nucleotides − 2000 to 0 bp, − 1500 to 0 bp, − 1000 to 0 bp, and − 500 to 0 bp were cloned into a PGL3 luciferase reporter construct. When the reporter was co-transfected with a HOXA4 overexpression construct into HUVECs, it displayed remarkable activation by HOXA4 within the promoter region of HOTAIRM1 ranging from − 2000 to 0 bp, − 1500 to 0 bp, and − 1000 to 0 bp. Furthermore, compared to the control group, multiple luciferase activities were enhanced in the HOTAIRM1 promoter regions. Luciferase activity from − 2000 to 0 bp was akin to that from − 1500 to 0 bp, while luciferase activity from − 1500 to 0 bp was higher than that from − 500 to 0 bp. However, there was no influence on the promoter activity of HOTAIRM1 in the region from − 500 to 0 bp (Fig. 3b). A thorough analysis of the aforementioned results showed that HOXA4 promotes HOTAIRM1 transcription. Of note, the core binding region located between − 1500 and − 500 bp was defined, and the JASPAR website again predicted four possible binding sites for HOXA4 in this region, namely, binding site 1 located between − 1130 and − 1123 bp (ACCATTAG), site 2 between − 1109 and − 1102 bp (GTGATTAA), site 3 between − 916 and − 909 bp (ATCATTAT), and site 4 between − 869 and − 862 bp (ATAATCAA) of the HOTAIRM1 promoter region (Fig. 3c). To identify the specific binding site of HOXA4 responsible for its binding to HOTAIRM1, a series of small-scale deletion mutants were generated. As shown in Fig. 3d, HOXA4 overexpression led to a substantial increase in the luciferase activity of HOTAIRM1 promoter with wild-type or sites 1/3/4 mutations in HUVECs, with barely altered luciferase activity of HOTAIRM1 promoter with site 2 or all four sites mutations, thereby revealing that the site 2 with − 1109 to − 1102 bp was important for HOXA4-mediated HOTAIRM1 activation. ChIP-qPCR assay was used to further examine the role of HOXA4 in the HOTAIRM1 promoter region. Interestingly, the occupation of HOXA4 in the HOTAIRM1 promoter region was substantially increased compared to that in the IgG control (Fig. 3e). Collectively, the investigation of transcriptional regulation revealed that HOXA4 directly binds to the HOTAIRM1 promoter and facilitates its transcription.

The regulatory role of HOXA4 on HOTAIRM1 has not been addressed thus far. The siRNAs specifically targeting HOXA4 were generated, and silencing efficiency was evaluated by qPCR and western blotting (Fig. 3f, g). Indeed, knocking down or overexpressing HOXA4 led to decreased or increased HOTAIRM1 expression, respectively (Fig. 3h). Functionally, CCK-8 and EdU assays showed that HOXA4 knockdown dampened the proliferation of HUVECs (Fig. 3i, j). Whether HOTAIRM1 is involved in the contribution of HOXA4 to cell proliferation regulation remains unknown. Rescue experiments showed that HOTAIRM1 overexpression partially counteracted the HOXA4 depletion-mediated proliferation impairment (Fig. 3k, l). Cumulatively, the effect may account for the fact that HOXA4 functions, at least in part, by being contingent upon HOTAIRM1 and indicates that HOXA4 can bind and activate the HOTAIRM1 promoter.

The introduction of ox-LDL led to a dramatic increase in the protein level of HOXA4 in HUVECs (Fig. 4a). To gain insight into how HOTAIRM1 affects cell proliferation, we first examined the localization of HOTAIRM1 through NCBI (https://​www.​ncbi.​nlm.​nih.​gov/​). HOTAIRM1 and HOXA4 are located adjacent to the same chromosome and in opposite directions of transcription (Fig. 4b). Existing data have demonstrated that lncRNAs can function in cis to regulate the expression of adjacent protein-coding genes [31], indicating that HOTAIRM1 may modulate HOXA4 expression, which has yet to be imprinted in HUVECs. HOTAIRM1 knockdown substantially diminished, whereas its overexpression markedly augmented HOXA4 expression at the mRNA and protein levels (Fig. 4c, d). To further elucidate how HOTAIRM1 influences HOXA4 expression, the possible interaction between HOTAIRM1 and HOXA4 was assessed using bioinformatics analyses. Using the RNA–protein interaction prediction (RPISeq) website (http://​pridb.​gdcb.​iastate.​edu/​RPISeq/​), HOXA4 was shown to interact with HOTAIRM1 using the random forest (RF) classifier (= 0.90) and support vector machine (SVM) classifier (= 0.96) (Fig. 4e), suggesting that HOTAIRM1 could interact with HOXA4. Inspired by these observations, follow-up RIP, RNA pull-down, and western blotting experiments were performed to determine whether HOTAIRM1 could interact with HOXA4 in an RBP manner. Unfortunately, the experimental results described here yielded no evidence of the corresponding RBP patterns (Fig. 4f, g). Overall, these discoveries, coupled with our existing knowledge presented above on the transcription promotion of HOTAIRM1 initiated by HOXA4, support the notion that HOTAIRM1 tuned the expression of HOXA4 in a non-RBP manner. Of course, this raises a major intriguing problem in the HOTAIRM1-mediated regulatory axis, which is a candidate for downstream HOTAIRM1-associated protein partners.

To delineate the RNA-binding protein landscape in the HOTAIRM1-mediated signaling network, we aimed to identify HOTAIRM1-associated proteins responsible for modulating cell proliferation. The subcellular localization of lncRNAs determines their modes of action [32]. Hence, the subcellular location of HOTAIRM1 was predicted using lncATLAS (https://​lncatlas.​crg.​eu/​), and the results indicated that it was preferentially localized in the nuclei of HUVECs (Fig. 5a). To unequivocally visualize HOTAIRM1 distribution as described by the RNA-FISH assay, it was also principally located in the nucleus of HUVECs (Fig. 5b), which enabled HOTAIRM1 to play an important role at the transcriptional level. Overwhelming evidence has conferred multiple regulatory functions to lncRNAs via interactions with diverse RNA-binding protein partners in the nucleus. Toward this end, HOTAIRM1-interacted proteins were identified and characterized by RNA pull-down assay, followed by mass spectrometry in HUVECs (Fig. 5c). Consequently, 233 proteins associated with HOTAIRM1 were identified using mass spectrometry (Supplementary File 3), and HSPA5 was identified as a prominent HOTAIRM1-binding protein among the top ten hit candidate proteins (Fig. 5d).

To further confirm the HOTAIRM1-HSPA5 interaction, the RNA–protein interaction prediction (RPISeq) website (http://​pridb.​gdcb.​iastate.​edu/​RPISeq/​) was used to predict whether HOTAIRM1 could bind to the HSPA5 protein. The scores for the random forest (RF) classifier and support vector machine (SVM) classifier are 0.7 and 0.85 (both over 0.5) (Fig. 5e), respectively, raising the likelihood that HOTAIRM1 could interact with HSPA5. A follow-up RIP experiment was carried out for the RNA-HSPA5 complex utilizing the HSPA5 antibody to corroborate this speculation. Notably, upon comparison with the IgG-control group, HSPA5 caused a robust HOTAIRM1 enrichment in HUVECs (Fig. 5f). Overall, these data indicate that HOTAIRM1 physically interacts with HSPA5 in HUVECs in an RBP manner.

Based on these results, we further investigated whether HOTAIRM1 could regulate HSPA5 expression in HUVECs and found that the protein level of HSPA5 was markedly up-regulated in HUVECs (Fig. 6a), which was in accordance with HOTAIRM1 expression in the presence of ox-LDL. The gain-and-loss-of-function experiments were performed to elucidate the regulatory role of HOTAIRM1 in HSPA5 expression. HOTAIRM1 knockdown mitigated this effect; however, its overexpression increased HSPA5 protein expression (Fig. 6b, c). Down-regulation or up-regulation of HOTAIRM1 remarkably decreased or increased HSPA5 expression at the mRNA level, respectively (Fig. 6d), indicating that HOTAIRM1 modulates the expression of HSPA5 in a transcription-dependent manner. To determine whether HSPA5 is responsible for the role of HOTAIRM1 in HUVEC proliferation, siRNA and overexpression plasmids specifically targeting HSPA5 were utilized to evaluate the protein expression level in HUVECs (Fig. 6e). EdU and CCK-8 assays demonstrated that HSPA5 knockdown markedly retarded HUVEC proliferation (Fig. 6f, g). We then transfected HOTAIRM1-silenced HUVECs with an HSPA5-overexpressing plasmid or empty vector and found that HSPA5 overexpression partly countered the effect of HOTAIRM1 depletion on HUVEC proliferation suppression (Fig. 6h, i), thereby showing that HOTAIRM1-mediated proliferation modulation was dependent on HSPA5. Cell fate determination was tightly orchestrated by a series of molecular signals to unveil the regulatory network of the HOXA4-HOTAIRM1-HSPA5 axis, and the effect of HSPA5 on HOXA4 expression was monitored. HSPA5 knockdown blunted, whereas its overexpression potentiated HOXA4 expression at the mRNA and protein levels in HUVECs (Fig. 6j, k).