Identification of endoplasmic reticulum stress-associated lncRNAs influencing inflammation and VSMC function in abdominal aortic aneurysm

Abdominal aortic aneurysm (AAA) is an age-related degenerative disorder with high morbidity and mortality and is commonly asymptomatic [1-3]. It is defined as a severe pathological condition in which the abdominal aorta enlarges reaching aortic diameter values >3.0 cm. In subjects aged over 65 years, the prevalence of AAAs is estimated between 2% and 6%, depending on the population under study, and AAA rupture is the most severe consequence of this disease, being a worldwide leading cause of death [3,4]. Although the risk of rupture increases exponentially as the aortic diameter grows, aneurysm size is not always a good predictor since small aneurysms occasionally rupture and a proportion of large AAA remains stable over time [4-6]. In this context, the management of these patients consists of watchful observation, being aneurysm diameter as the principal surrogate marker for disease progression and rupture, and the main determinant for surgical repair, the only therapeutic option for this disorder [7]. Therefore, effective therapeutic approaches able to halt the expansion of AAA are urgently needed. Potential drug candidates previously assessed in clinical trials did not show a clear therapeutic effect, and even in some cases, they could lead to a paradoxical worse progression with faster AAA growth [8-10]. Consequently, the development of therapeutic strategies for slowing AAA expansion and providing better clinical outcomes requires a deep understanding of AAA pathogenesis.

AAA is a complex disease involving various risk factors, such as smoking, aging, male sex, history of AAA in first-degree relatives, atherosclerotic disease, and hypertension [3-5]. The pathogenic mechanisms underlying the destructive vascular remodeling characteristic of this disease entail protease activation, extracellular matrix (ECM) degradation, vascular smooth muscle cells (VSMCs) phenotypic switch and death, and extensive medial and adventitial degeneration combined with an exacerbated inflammatory response [11-13]. Interestingly, we and others have reported the critical role of endoplasmic reticulum (ER) stress in the pathogenesis of several cardiovascular diseases including hypertension, atherosclerosis, heart failure, and AAA [14-16]. Environmental factors related to aneurysmal disease, such as hypoxia, reactive oxygen species, aging, and genetic factors, disrupt ER function leading to an accumulation of misfolded and unfolded proteins in the ER lumen, which promotes ER stress. When ER stress occurs, the unfolding protein response (UPR) is activated, initiating adaptive responses for maintaining cellular homeostasis. However, UPR is a double-edged sword as temporary activation can restore ER homeostasis, but, if the ER stress is prolonged, the homeostatic capacity of the adaptive UPR becomes saturated and the UPR response is dominated by the maladaptive process leading to apoptosis [17-19].

Noncoding RNAs have been gaining much attention as epigenetic regulators in cardiovascular diseases. Among them, long noncoding RNAs (lncRNAs) stand out as essential factors in health and disease with a significant, but not yet clearly defined, role in vascular physiology. Accumulating evidence supports the influence of lncRNAs in the regulation of vascular function and inflammation [20-23]. However, studies focused on lncRNAs and AAA are still scarce, and their specific contribution to aneurysm disease development and progression remains largely unexplored [23,24], particularly concerning ER stress. In the present study, we have performed a whole-genome transcriptomic analysis characterizing the expression profile of lncRNAs in AAA, identifying a signature of ER-stress-associated lncRNAs, most of which are also associated with mitochondrial function, that may be useful for the development of new pharmacological interventions in AAA.

Human abdominal aneurysmal aortas were obtained from patients diagnosed with infrarenal AAA and undergoing open surgical repair at the Hospital de la Santa Creu i Sant Pau (HSCSP; Barcelona, Spain). The diagnosis of AAA was confirmed by computed tomography scan. Patients with AAA and with negative histories of rheumatological, immunological diseases, aortitis, or genetic syndromes such as Marfan disease were included in the study. Other exclusion criteria were juxtarenal aneurysms and mycotic aneurysms. Healthy abdominal aortas came from multiorgan donors. Approval to use the discarded human tissue was given by the Ethics Committee of the HSCSP, and the participation in the study of patients and controls was based on the informed consent of patients or legal representatives. Research was conducted in accordance with the Declaration of Helsinki of 1975. Abdominal aorta segments were obtained according to standard operating procedures and ethical guidelines. Samples of control subjects had no postmortem evidence of AAA. For RNA studies, parts of the tissue samples were collected, snap-frozen, and stored at −80°C for subsequent RNA extraction.

Human VSMCs were isolated from the aorta of multiorgan donors by an explant procedure as previously described [16]. Endothelium-denuded medial tissue was cut into 2–4-mm cubes that were transferred to a 23-cm² culture flask containing 5 ml of prewarmed culture medium M199 (Gibco, Carslbad, Cam U.S.A.) supplemented with 10% fetal calf serum (FCS; Biological Industries, Kibbutz Beit-Haemek, Israel) and antibiotics (100 U/ml penicillin and 0.1 mg/ml streptomycin). VSMCs migrate out from the explants within two to three weeks. Then, after removing the explants from the flask surface, cells were trypsinized, used as P1 stage cells, and routinely subcultured. Human VSMCs were cultured in M199 supplemented with 20% FCS, 2% human serum, 2 mmol/l L-glutamine (Invitrogen), and antibiotics incubated at 37°C in a 5% CO₂-humidified incubator.

For experimental procedures, cells between passages 3 and 6 were seeded in six-multiwell plates. Subconfluent cells were serum-starved in a medium supplemented with 1% FCS for 24 h prior to the addition of 1 µg/ml of tunicamycin (Sigma-Aldrich) or vehicle (dimethyl sulfoxide, DMSO, Sigma-Aldrich) for 8 h. In some experiments, cells were pretreated with the ER stress inhibitor tauroursodeoxycholic acid (TUDCA; 500 µg/ml), 1 h prior to exposure to tunicamycin [25].

The human T-acute lymphoblastic leukemia Jurkat cell line was purchased from American Type Culture Collection (ATCC, Manassas, VA), maintained in suspension, propagated in RPMI 1640 medium, GlutaMAX (Gibco, Carslbad, Cam U.S.A.) supplemented with 20% FCS and 1% antibiotics (100 U/ml penicillin and 0.1 mg/ml streptomycin), and incubated at 37°C in a 5% CO₂-humidified incubator. For experimental procedures, Jurkat cells were seeded at a density of 5 × 10⁵ per well, pretreated or not for 8 h with 100 µg/ml of TUDCA [26], and then incubated with 1 µg/ml of tunicamycin for 48 h.

The RNeasy Fibrous Mini Kit (Qiagen, Venlo, Netherlands) was used to isolate total RNA from human aortic samples following the manufacturer’s recommendation. Total RNA isolation from human VSMCs and Jurkat cell was performed using the TRI Reagent (Zymo, Research, U.S.A.) following the manufacturer’s instructions. Total RNA was determined and quantified by a NanoDrop 1000 Spectrophotometer (Thermo Scientific). For RNA sequencing (RNA-seq), RNA integrity and quality was assessed on Agilent 2100 Bioanalyzer (Aligent Technologies, Santa Clara, CA, U.S.A.).

To study up-regulated and down-regulated genes in different patient groups, a whole-genome analysis was performed using RNA-seq for whole-genome transcriptome using RNA-seq at the Genomics/Transcriptomics Core of the CNAG (Barcelona). Total RNA was used as starting material to synthesize cDNA libraries using the Illumina Stranded mRNA Prep kit based on the poly-A selection method, according to the manufacturer’s protocol (Illumina, San Diego, CA, U.S.A.). Libraries were run simultaneously on an Illumina NextSeq 500 sequencing system by the NextSeq 500/550 High-Output reagent kit v2.5 of 150 cycles (2 × 74 bp paired end) (Illumina).

Paired sequence files in FASTQ format were analyzed on the BaseSpace Sequence Hub platform with the Dragen RNA Pipeline v3.7.5 (Illumina). The reference genome used was UCSC hg19. Differential expression analyses were performed with the Dragen Differential Expression application (Illumina). The compared subset of samples were AAA and healthy donor samples. Significantly differentially expressed genes (DEGs) were defined as those with adjusted P-values of <0.05 and visualized by EnhancedVolcano tool from R Studio (EnhancedVolcano tool from Bioconductor package.https://bioconductor.org/packages/devel/bioc/vignettes/EnhancedVolcano/inst/doc/EnhancedVolcano.html), and DEGs were ranked by log2 fold change.

Functional enrichment studies were carried out using String Database (version 12.0) (https://string-db.org/) and Reactome Pathways and visualized using SRplot [27]. The complete list of enriched terms can be found in Table S1. LncRNA enrichment analysis was performed by introducing GO terms related to ER stress and mitochondria into the lncRNA Ontology (http://bio-bigdata.hrbmu.edu.cn/lncrnaontology/). The complete list of lncRNAs associated with ER stress and mitochondria can be found in Tables S3 and S4.

For DE mRNAs-lncRNA co-expression matrix, we evaluate whether the normalized expression value of the selected lncRNA and the DE mRNA is linearly correlated by calculating the Person correlation coefficient. We considered correlating the pairs with a Pearson correlation coefficient greater than |0.9|. The complete list of values can be found in Table S2. All DE-mRNA correlated to LINC00494, PCED1B-AS1, TCL6, and LINC00892 were used to perform a functional enrichment study as described above. Reactome pathways selected were used to construct protein–protein interaction networks according to the String database and visualized using Cytoscape software (http://www.cytoscape.org).

DNase I-treated total RNA (1 µg) was reverse-transcribed into cDNA using the High-Capacity cDNA Archive Kit (Applied Biosystems, Foster City, CA, U.S.A.). The mRNA levels of DE-lncRNAs and ER stress markers were quantified by using specific probes provided by the assay-on-demand system (Applied Biosystems): ATF-4 (Hs00909569_g1), ATF-6 (Hs00232586_m1), CHOP or DDIT3 (Hs99999172_m1), cysteine-rich with EGF-like domains 2 (CRELD2, Hs00360923_g1), heat shock protein 5 (HSPA5), or GRP78 (Hs99999174_m1), IRE1 or ERN1 (Hs00176385_m1), Suppressor/Enhancer of Lin-12-like (SEL1L, Hs01071406_m1), and XBP1 (Hs00231936_m1); and the lncRNAs: CCDC26 (Hs00540885_m1), ITGB2-AS1 (Hs04274319_m1), IL21-AS1 (Hs04976181_s1), ITPKB-IT1 (Hs05032319_s1), LY86-AS1 (Hs00543584_m1), MIR137HG (Hs01939875_s1), PCED1B-AS1 (Hs05027356_m1), TCL6 (Hs00220956_m1), WASIR2 (Hs00415846_m1), LINC00158 (Hs00365202_m1), LINC00381 (Hs01381499_g1), LINC00426 (Hs03680869_m1), LINC00494 (Hs01052916_s1), LINC00544 (Hs00417034_m1), LINC00582 (Hs04274021_m1), LINC00626 (Hs00365491_m1), LINC00861 (HS04274413_m1), LINC00892 (Hs00866930_s1), and LINC00926 (Hs03680805_m1). As housekeeping genes, glyceraldehyde 3-phosphate dehydrogenase (Hs02758991_g1) and β-actin (Hs99999903_m1) or r18SRNA (Hs.PT.39a.22214856.q) were used. Each sample was amplified in duplicate by quantitative RT-PCR in an ABI PRISM 7900HT Sequence Detection System (Applied Biosystems). The same results were obtained after the normalization to both housekeeping genes. Relative mRNA levels were determined using the 2−∆∆Ct method.

Results are expressed as mean ± SEM of the number (n) of samples indicated in the figure legends. Demographic and clinical characteristics of the human individuals were expressed as mean ± SD. The normal distribution of the variables was determined by the Shapiro–Wilk test. When data followed a normal distribution, differences between the two groups were assessed using the Student’s t-test (two-tailed), and when normality failed, the Mann–Whitney rank sum test was applied. To compare more than two groups, the one-way ANOVA or the Kruskal–Wallis ANOVA test was used and corrected for multiple comparisons by controlling the false discovery rate with the two-stage step-up method of Benjamini, Krieger, and Yekutieli. Pearson product–moment correlation coefficient was used to study the correlation between variables following a normal distribution, whereas Spearman’s rank-order correlation was applied to determine the association between variables when data failed the normality test. Receiver-operating characteristic (ROC) curve analysis was used to evaluate the ability of each lncRNA to discriminate patients from healthy controls. Data analysis was carried out using GraphPad Prism version 9.2.20 software (La Jolla, CA, U.S.A.) and the R Software version 4.4.0 (R Foundation for Statistical Computing, Vienna, Austria. https://www.r-project.org/foundation/). Values of P≤0.05 were considered significant.

A transcriptomic analysis of abdominal aorta samples from patients and controls was performed by RNA-seq (n = 12 AAA patients and n = 7 donors). Demographics and clinical characteristics of patients and donors included in the RNA-seq closely matched for age, sex, and smoking habit as depicted in Table 1. All the patients included in the study had an abdominal aorta diameter greater than 55 mm, which is the current threshold size indicated for AAA surgical repair [1,2]. Most AAA patients were male, active smokers or ex-smokers, and suffered from hypertension and dyslipidemia (Table 1). The unsupervised hierarchical clustering for gene expression data revealed a distinctive profile associated with aneurysmal disease (Supplementary Figure 1A). A total of 6572 genes were differentially expressed in human AAA in comparison with donors, with 4035 genes up-regulated and 2538 down-regulated (Figure 1A). Functional enrichment analysis of these DEGs indicates an affectation of processes such as smooth muscle contraction, elastic fiber formation, immune system signaling, and mitochondrial biology (Figure 1B). Notably, besides mRNA transcripts, our analysis also identified 1283 DE-lncRNAs in AAA (Supplementary Figure S1B; Figure 1C).

Our previous research highlighted the importance of ER stress in the pathophysiology of human and experimental AAA [16,28]. Of note, as shown in the Venn diagram (Figure 1D), functional enrichment analysis in the LncRNA Ontology database revealed that a set of 368 DE-lncRNAs (256 up-regulated and 112 down-regulated) were members of the gene ontology category GO0034976, corresponding to the response to ER stress. Since ER and mitochondrial stress are closely related processes, we also aimed to explore whether these DE-lncRNAs could be related to mitochondrial dysfunction. As shown in Figure 1D, 436 DE-lncRNAs (273 up-regulated and 163 down-regulated) were associated with the GO term mitochondrion (GO0005739). It should be noted that 164 of these (108 up-regulated and 56 down-regulated) are DE-lncRNAs common to both processes (Figure 1D). We specifically focused on the most significantly increased lncRNAs of those up-regulated 256 lncRNAs that include, ITGB2-AS1, IL21-AS1, ITPKB-IT1, LY86-AS1, MIR137HG, PCED1B-AS1, TCL6, WASIR2, LINC00158, LINC00426, LINC00494, LINC00582, LINC00626, LINC00861, LINC00892, and LINC00926 (Figure 1C; highlighted in yellow those DE-lncRNAs common with mitochondrion). The up-regulation in human AAA of these lncRNAs was ratified by real-time PCR in the same set of aneurysmal and donor samples used for the transcriptomic profiling, as shown in Supplementary Figure S2. The expression analysis was also carried out in a large cohort of AAA patients (n = 100) and healthy donors (n = 39) (Table 2), showing similar characteristics than those described above, that is, about a quarter of AAA patients were diabetics, had chronic kidney disease, and/or suffered from ischemic heart disease. Roughly 60% of the patients were being treated with cholesterol-lowering drugs, antiplatelets, and anticoagulants and 33% with ACE inhibitors (Table 2). Interestingly, the induction of those nine lncRNAs exhibiting the greatest expression in abdominal aorta (IL21-AS1, ITPKB-IT1, PCED1B-AS1, TCL6, LINC00494, LINC00582, LINC00626, LINC00861, and LINC00892) was further confirmed in this large cohort of aneurysmal samples (Figure 2).

Because of the participation of the selected lncRNAs in the response to ER stress cluster, we sought to investigate whether there was any correlation between their expression and that of ER stress and ER-associated degradation (ERAD) markers in this large cohort. As depicted in Table 3, significant correlations were found between HSPA5, ATF-6, CHOP, and IRE-1/ERN1 and some of the selected lncRNAs, being XBP1 the most outstanding, since it significantly correlated with ITPKB-IT1, TCL6, and LINC00582. No correlations were detected for ATF-4. More remarkable were the co-expression detected with the ERAD markers SEL1L and CRELD2, the latter correlating with IL21-AS1, ITPKB-IT1, LINC00626, LINC00861, and LINC00892 (Table 3).

Interestingly, the ROC curve analyses of the nine lncRNAs showed that all lncRNAs discriminated AAA patients from donor subjects. LINC0861 and PCED1B-AS1 displayed the highest performance. The area under the curve (AUC) was 0.919 (95% CI 0.87–0.96) for LINC0861, and 0.914 (95% CI 0.86–0.96) for PCED1B-AS1, indicating that these two lncRNAs show the best sensitivity and specificity and supporting the reliability of each lncRNA in distinguishing aneurysmal from donor samples (Figure 3A). Additionally, we performed a logistic regression analysis to figure out the association of each lncRNA to AAA diagnosis (Figure 3B). In the univariate analysis, the lncRNAs IL21-AS1 and LINC0582 were excluded due to the loss of the association with AAA. After adjusting for age, gender, and smoking, the rest of the lncRNAs increased their association with AAA. PCED1B-AS1 and LINC0861 exhibited the best-adjusted OR values and were selected for further analysis in ROC curves. Combining these lncRNAs in a single model did not result in a significantly superior discriminatory capability compared with each lncRNA alone with an AUC of 0.930 (95% IC 0.888–0.971). Furthermore, we extended this analysis by performing a backward stepwise regression analysis resulting in a combination of four lncRNAs (ITPKB-IT + LINC0626 + LINC0861 + LINC0892), but the discriminatory performance of this combination analyzed by ROC curves resulted in a non-significant AUC improvement of 0.933 (95% IC 0.893–0.974), compared with each lncRNA alone (data not shown).

Next, we aimed to characterize the ER stress-dependent regulation of the identified lncRNAs in human VSMC and Jurkat cells, a human immortalized T-lymphocyte cell line. The treatment with tunicamycin, which is extensively used as an experimental tool to induce ER stress, significantly increased the expression of the ER stress markers ATF-4, ATF-6, CHOP, and XBP-1 in both cell types (Figures 4 and 5). Concurrently, tunicamycin up-regulated the levels of five of the selected lncRNAs (IL21-AS1, PCED1B-AS1, LINC00494, LINC00626, and LINC00892) in VSMCs (Figure 4), while in Jurkat, it enhanced those of IL21-AS1, ITPKB-IT1, LINC00494, LINC00892, PCED1B-AS1, and TCL6 (Figure 5). The treatment with TUDCA, a classic ER stress inhibitor that significantly abolished the tunicamycin-mediated increase in ER stress markers in both cell types (Figures 4 and 5), normalized IL21-AS1, PCED1B-AS1, LINC00494, and LINC00626 in VSMC (Figure 4), while it attenuated the up-regulation of all lncRNAs induced by tunicamycin in Jurkat cells, except that of LINC00494 (Figures 4 and 5).

To establish the possible functional relationship between the selected lncRNAs and the DE-mRNAs identified in human AAA, we generated a co-expression matrix. Both LINC00494 and PCED1B-AS1 correlate with more than 1000 DE-mRNAs (Figure 6A), most of which are common (Figure 6B). In turn, we observed that TCL6 and LINC00892 correlate with 185 and 310 DE-mRNAs, respectively (Figure 6A), and found that these genes are largely shared with those that correlate with LINC00494 and PCED1B-AS1 (Figure 6B). The ontological analysis of the mRNA transcripts differentially regulated in AAA that correlate with LINC00494, PCED1B-AS1, LINC00892, and TCL6 evidenced the participation of these DE-mRNAs in muscle contraction and in different aspects of the immune response (immunoregulatory interactions, TCR signaling, cytokine signaling in immune system, adaptive immune system, innate immune system, and neutrophil degranulation in 336 different genes among different categories) (Figure 6C). Protein interactions among the DE-mRNAs included in the clusters ‘muscle contraction’ and ‘adaptive immune system’ were analyzed by STRING as depicted in Figure 6D, where it can be observed how most of the DE-mRNAs involved in muscle contraction were down-regulated in AAA tissue (blue bubbles), while those related to the adaptive immune system were up-regulated (red bubbles).

Although remarkable progress has been made in AAA research over the last few decades, the development of new pharmacological strategies to fight against this disease has been a challenge which is still unsolved [10]. Notably, accumulating evidence supports that epigenetics is critical in aneurysmal disease and that its control may hold potential for therapeutic intervention. Certainly, epigenetic mechanisms involving histone posttranscriptional modifications, DNA methylation, and non-coding RNAs seem to play an essential role in the control of fundamental processes in aneurysmal disease, such as VSMC function, inflammation, and ECM remodeling, and emerge as potential targets for pharmacological interventions and biomarkers [10,29]. However, among them, the regulation of gene expression through lncRNAs is less understood and has been poorly characterized. Here, our transcriptome profiling in human abdominal aorta samples has revealed a more complete picture of lncRNAs in this disease and our bioinformatic analysis has provided a comprehensive analysis of relevant lncRNAs related to ER stress in the pathogenesis of AAA.

Similar to previous studies using RNA-seq in human AAA and healthy aortic tissue samples [30-32], our high-throughput sequencing and functional enrichment analysis revealed the disturbance of genes involved in the control of vascular function and inflammation in this disease. Patients and donors included in our transcriptional profiling study were matched for age, sex, and smoking habit, allowing to minimize biased observations. Interestingly, our analysis identified a bulk of transcripts corresponding to lncRNAs whose expression was differentially regulated in aneurysmal samples. Previous high-throughput analysis in human AAA samples focused on the study of lncRNAs was limited, and few lncRNAs were confirmed as differentially regulated in human disease [29-31,33].

Of note, a huge number of the differentially regulated lncRNAs found in our study are involved in the response to ER stress, which plays a critical role in AAA development. In fact, we have previously reported the induction of ER stress in human AAA, its relationship with oxidative stress, and that ER stress inhibition abrogates aneurysm development in a preclinical model [16,28]. Specifically, our validation in a large cohort of patients with AAA and donors identified nine ER stress-related lncRNAs significantly up-regulated in aneurysmal tissue. Moreover, a significant correlation with the expression of ER stress and ERAD markers was detected for most of these lncRNAs. Further, PCED1B-AS1, LINC0861, IPTKB-1, LINC0626, LINC0494, and TCL6 exhibited the highest AUC values in ROC curves, indicating a high sensitivity and specificity regarding their ability to distinguish those individuals with AAA. The adjusted analysis by confounding factors showed that most of the selected lncRNAs kept their association with AAA excepting IL21-AS1 and LINC0582.

Recent works have focused on the mechanisms linking specific lncRNAs to ER stress and their contribution to high-impact diseases such as heart disease, cancer, and diabetes. These studies have unraveled how the downstream components of the UPR regulate transcription and translation reprogramming to ensure selective gene and lncRNAs expression in response to pathological stimulus [34-36]. It should be noted that some of the lncRNAs investigated in the present study, such as TCL6, LINC0892, and IL21-AS1, were reported to be involved in metastasis and tumor progression [37,38] and in other pathological conditions such as preeclampsia [39,40]; however, this is the first study involving these selected lncRNAs in AAA.

The contribution of lncRNAs to human AAA has been scarcely addressed, although recent studies have revealed that specific lncRNAs affect some of the mechanisms underlying AAA growth [29-33]. A recent single-cell RNA sequencing analysis revealed the down-regulation of the lncRNA CARMN in human aneurysmal VSMC associated with the altered contractile function of these cells [41]. Accordingly, in our study, CARMN expression was found down-regulated in AAA samples, and it is also one of the lncRNA related to ER stress (data not shown). The up-regulation of the lncRNA H19 was confirmed in a small cohort of tissue samples from donors and AAA patients, and studies in preclinical models support its role in the progression of AAA through the control of VSMC survival and inflammation [29,42,43]. Similarly, GAS5 and PVT1 have been involved in human AAA development by inducing VSMC apoptosis among other mechanisms [44,45]. In accordance, our RNA-seq analysis detected the up-regulation of H19 and PVT1, but not that of GAS5, in AAA samples (data not shown); however, neither H19 nor PVT1 was among the ER stress-related lncRNAs identified in our study.

Both medial VSMC paucity and vascular inflammation are critical contributors to vascular degeneration and weakening in AAA [12,46,47] and are closely associated with ER stress and mitochondrial dysfunction [13,16,28]. Although our data in human samples could not ascertain whether the response of these lncRNAs i a cause or a consequence of the response related to ER stress, studies on VSMC and T cells in culture indicate that, in both cell types, the activation of ER stress significantly up-regulated the expression of most of the DE-lncRNAs found in human aneurysmal lesions, an effect abolished by the ER stress inhibitor TUDCA. Thus, these data support the direct relationship between these transcripts and ER stress in vascular and inflammatory cells. Further, we have established the lncRNA-mRNA network in human AAA. Based on the functional enrichment analysis of the DEGs and from the ontological analysis of the correlation between the DE-mRNA and DE-lncRNAs, the identified lncRNAs signature has been related to both vascular inflammation and VSMC contractile dysfunction, and similar results were found by analyzing the protein interaction profile. Although some of the lncRNAs investigated in our work have been described as biomarkers in other disorders and lncRNA H19 shows evidence to be a diagnostic biomarker for cardiovascular diseases, no specific data have been reported in the setting of AAA [48]. The differential profile of other lncRNAs rather than those described here has been recently reported in peripheral blood mononuclear cells and serum from AAA patients [49,50], and their interest as biomarkers deserves further investigation.

In summary, based on our data, lncRNAs might be considered as key regulatory molecules involved in the response to ER stress in AAA, affecting both vascular inflammation and VSMC function. Although the identified lncRNAs have no murine homologous counterparts, hampering in vivo studies, further research is warranted to determine whether the lncRNAs signature described in this work is involved in the progression of the disease and might become the basis for the development of new pharmacologic interventions in AAA.

The data that support the findings of this study are included as supplementary files (Tables S1, S2, and S3) in Excel format and may be asked of the corresponding authors upon reasonable request.

The authors declare no conflict of interest.

This work was funded in part by the Instituto de Salud Carlos III (ISCIII) under grant [PI20/01004] to M.G., [PI21/01048] to C.R., in part by “ERDF A way of making Europe,” and in part by Ministerio de Ciencia, Innovación y Universidades (MCIN) under grant [PID2020 116498RB-I00] to A.G-R. R.A-P is supported by a PFIS fellowship [FI21/00125, ISCIII]. M.G. was hired under the Miguel Servet Program [ISCIII. Ref. CP20/00013]. The study has been carried out within the framework of the Doctorate Program in Pharmacology of the Autonomous University of Barcelona.

This article has been published open access under our Subscribe to Open programme, made possible through the support of our subscribing institutions, learn more here: https://portlandpress.com/pages/open_access_options_and_prices#conditional

M.G.: Conceptualization, Supervision, Investigation and Methodology. A.G.R. Formal analysis and Methodology. R.A.P., M.C., M.G., and A.G.R: Investigation, Formal analysis, and Data curation. C.R., E.P.M., D.S.I., and S.C.: Conceptualization, Investigation, and Data curation. M.G., A.G.R., and C.R: Writing - original draft and Writing - review & editing. J.L.S.Q., G.M.G., and J.M.G: Formal analysis, Investigation, and Writing - review & editing. All the authors revised the manuscript and gave their final approval of the submitted version.

We thank the technical support provided by Silvia Aguiló.

AAA

abdominal aortic aneurysm

DEGs

differentially expressed genes

ER

endoplasmic reticulum

RNA-seq

RNA sequencing

VSMCs

vascular smooth muscle cells

lncRNA

long non-coding RNA