Macrophage-Mediated Downregulation of lncRNA Carmn in Mouse Abdominal Aortic Aneurysm

Abstract

The long noncoding RNA (lncRNA) CARMN (cardiac mesoderm enhancer associated noncoding RNA) is a highly conserved lncRNA that expresses primarily by smooth muscle cells (SMCs). Recent literature demonstrates that CARMN plays a critical role in the differentiation and maintaining of the contractile state of vascular SMCs. Because aortic SMCs show diminished contractile proteins in abdominal aortic aneurysms (AAAs), we hypothesize that the expression of CARMN is downregulated in the aortic wall affected by aneurysm. In this study, we analyzed publicly available single-cell or bulk RNA sequencing data comparing healthy and aneurysmal mouse aortic tissues. In both healthy and diseased aortas, Carmn expression was enriched in SMCs characterized by the high expression of SMC-specific contractile proteins including Myh11 and Acta2. Carmn expression levels varied among the sub-clusters of SMCs and consequently along the aortic tree. Comparing to the corresponding sham aorta, aortas from 3 distinct AAA models contained less Carmn. To validate the Carmn downregulation, we induced AAA using the Angiotensin II and CaCl₂ models. In situ hybridization showed that Carmn mRNA located in the nuclei of SMCs and became downregulated within a few days following the aneurysm induction. Mechanistically, we tested whether Carmn expression is regulated by infiltrating macrophages --- the predominant inflammatory cells found in aneurysmal tissues --- by treating healthy mouse aortic SMCs with media conditioned by macrophages primed with pro-inflammatory or anti-inflammatory cytokines. PCR analysis showed that inflammatory macrophages reduced the expression of Carmn and contractile genes including Myh11 and Acta2. Taken together, our results from bioinformatic and experimental analyses demonstrate that Carmn is downregulated in different AAA models, likely by inflammatory macrophages. The negative regulation of Carmn in AAA tissues may explain at least in part the loss of SMC contractile state during the pathogenesis of this progressive degenerative disease.

Graphical Abstract

Introduction

Long noncoding RNAs (lncRNAs) are RNAs that are longer than 200 nucleotides and do not encode proteins ^1,2. LncRNAs often exhibit cell- and tissue-specific expression patterns and play important roles in normal development and disease pathogenesis ^1–3. Among the growing list of lncRNAs, cardiac mesoderm enhancer associated noncoding RNA (CARMN in human, Carmn in mouse) has recently been identified as a highly abundant and conserved SMC-specific lncRNA ⁴. Previous studies have demonstrated that CARMN involves in regulating vascular smooth muscle cell (SMC) plasticity and atherosclerosis by interacting with the SMC transcription factor serum response factor (SRF) and the transcriptional cofactor myocardin, which constitute molecular switch of vascular smooth muscle contractile phenotype ^4–6.

Abdominal aortic aneurysm (AAA) is a potentially fatal condition characterized by the progressive dilation and weakening of the abdominal aorta ^7,8. One of the major pathological features of AAA is the depletion of SMCs ⁷. Accumulating experimental evidence suggests that SMCs undergo phenotypic modulation or dedifferentiation during aneurysm development and growth. This process causes SMCs to loss their contractile functionality and acquire a phenotype that is pro-inflammatory ^7,8. Decrease in CARMN expression has been observed in thoracic aortic aneurysm (TAA) ⁹. However, the expression pattern and distribution of CARMN during AAA development remain unknown.

In this study, we analyzed publicly available single-cell RNA sequencing (scRNA-seq) data from our own study and other relevant studies, our results suggest a potential role of CARMN in AAA. Additionally, we performed in situ hybridization to validate our analysis. Our findings revealed distinct expression levels of CARMN in various segments of the aorta tree. Furthermore, AAA induction reduced CARMN expression in experimental AAA models, likely by pro-inflammatory signaling from macrophages.

Materials and Methods

The supporting data for this study can be obtained from the corresponding author upon a reasonable request.

Human AAA Samples

We acquired de-identified human AAA tissues during routine surgical repair procedures. The utilization of these human tissues, which would otherwise be considered surgical waste, was approved by the institutional ethics committee of the University of Wisconsin—Madison (IRB#: 2021-0215).

Animal Studies

The Institute Animal Care and Use Committee at the University of Wisconsin – Madison approved all animal studies conducted in this research (Protocol #M005792). Male wildtype mice, aged between 8 to 12 weeks and matched for weight, were used in all experiments. This choice of mice was made because AAAs primarily affect males ¹⁰. Throughout the study, the mice had unrestricted access to a standard laboratory diet (2018 Teklad global 18% protein rodent diets, ENVIGO) and water.

CaCl₂ induced murine AAA model

The modified CaCl₂ AAA model, as previously described ¹¹, was employed in this study. Mice were anesthetized with isoflurane, and 0.6 mg/kg sustained-release buprenorphine was subcutaneously administered before surgery. A midline laparotomy was performed to expose the infrarenal abdominal aorta. Perivascularly, a small gauze soaked in 0.5 mol/L CaCl₂ was applied for a duration of 10 minutes. This gauze was then replaced with another piece soaked in PBS, which was left for 5 minutes. Mice in the sham group underwent a similar procedure but with treatment using 0.5 mol/L NaCl for 10 minutes, followed by PBS-soaked gauze for 5 minutes. The abdomen was closed using a layered technique. After indicated days, mice were euthanized, and their aortae were harvested, embedded in optimal cutting temperature compound (Sakura Tissue Tek), and cut into six-micrometer cross sections.

PCSK9 AAV Angiotensin II model of AAA

The PCSK9 AAV Angiotensin II model, previously described ¹², was employed in this study. Mice were intraperitoneally injected with 20×10¹⁰ genomic copies of adeno-associated virus (AAV) expressing the gain-of-function mutation mouse PCSK9D377Y. After two weeks, mice with serum cholesterol levels exceeding 250 mg/dl were subcutaneously implanted with Alzet osmotic minipumps (Model 2004, DURECT Corporation). These minipumps were filled with either a PBS vehicle or Angiotensin II solutions (A9525, Sigma-Aldrich), which delivered a dose of 1000 ng/kg per minute of Angiotensin II. After 3, 7, or 28 days, the mice were euthanized, perfused, and their aortae were harvested. The aortae were then embedded with optimal cutting temperature compound (Sakura Tissue Tek) and cut into six-micrometer cross sections.

RNA In Situ Hybridization

Fresh frozen sections of murine aorta were analyzed using the Advanced Cell Diagnostics RNAscope Fluorescent Multiplex Reagent Kit (#320850). We utilized the RNAscope Probe targeting mouse Carmn (493541) or mouse Myh11 (316101-C2). For paraffin-embedded human AAA tissue, the RNAscope 2.5 HD Duplex Reagent Kit (322430) was employed with a probe specifically designed for human CARMN (493421). Tissue pretreatment, probe hybridization, and signal amplification steps were conducted following the manufacturer’s instructions. Images were captured using a Nikon A1RS confocal microscope system, and subsequent analysis was performed using Fiji/ImageJ software.

Bioinformatic analysis

Raw count matrices of scRNA-seq data were acquired from Gene Expression Omnibus (GEO, GSE191226, GSE221789, GSE164678) and subjected to Scanpy (Ver. 1.9.1) ¹³ to remove low quality cells. The cutoffs for removal were set at: (1) number of genes expressed per cell > 3,000 and <10,000, (2) total counts per cell > 10,000 and < 100,000, (3) percent of mitochondrial gene counts < 15%. The preprocessed AnnData objects was then concatenated into one AnnData object in scanpy and passed to scvi-tools (single-cell variational inference tools) ¹⁴ for making and training of the SCVI deep learning model with the correction of the batch effects among different sample. X_scVI, the latent representation of each cell, was extracted from the trained model and used to compute, embed and cluster the neighborhood graph in two dimensional UMAP. Bulk RNA-seq data was downloaded from GEO (GSE153097, GSE153425) and analyzed with DESeq2 ¹⁵ to obtain the differentially expressed genes. Microarray data was downloaded from GEO (GSE7084) and analyzed with GEO2R to obtain the differentially expressed genes.

Cell culture

Primary aortic SMCs were isolated from the abdominal aorta of C57BL/6J mice (The Jackson Laboratory, RRID: IMSR_JAX:000664), as described previously ¹⁶. Cells at passages 3 to 7 were used. Bone marrow derived macrophages were isolated as described before ¹¹. In brief, bone marrow cells were collected from femurs and tibias of C57BL/6J mice and differentiated into bone marrow–derived macrophages (BMDMs) using a macrophage medium comprising DMEM with 10% FBS, 1% penicillin/streptomycin antibiotic, and 20 ng/ml mouse macrophage colony-stimulating factor (M-CSF). Medium was changed after 3 days of incubation at 37 °C, 5% CO₂. After an additional 4 days of culture, cells were prepared for experiments.

Conditioned medium study

BMDM were treated with 20 ng/ml IL4 or BSA for 24 hours, or primed with 100 ng/ml LPS for 2 hours followed by 20 ng/ml IFNγ for 24 hours. Cell culture medium was discarded, cells were thoroughly washed three times with PBS. Fresh DMEM containing 0.3% FBS and 20 ng/ml M-CSF was then added to cells. After 24 hours, conditioned medium was collected and centrifuged at 2000 rpm for 5 min to removed dead cells.

Primary SMCs were cultured in DMEM with 0.3% FBS for 48 hours, then cell culture medium was replaced with BMDM conditioned medium and further incubated for 48 hours. Cells were collected and analyzed by real-time PCR. Primer sequences were listed in Supplementary Table 1.

Statistics

The bioinformatic analysis was described above. Other results were presented as mean ± SD. Data normality was assessed using the Shapiro-Wilk normality test. Two-tailed Student’s t-test was used for normally distributed data, and Mann-Whitney U nonparametric test was used for skewed data when comparing two conditions. For comparing three or more means, one-way ANOVA with Tukey’s post hoc test was used for normally distributed data, and the Kruskal-Wallis nonparametric test was used for skewed data. Two-way ANOVA followed by Sidak’s multiple comparisons test was performed to analyze the impact of two factors on a response. Statistical analyses were conducted using GraphPad Prism 9 software (GraphPad Software, Inc). Experiments were repeated as indicated. Differences with P < 0.05 were considered statistically significant.

Results

lncRNA Carmn shows region-dependent expression in mouse aorta

To determine the expression and distribution of Carmn in the aorta tree, we analyzed the published single-cell RNA sequencing (scRNA-seq) dataset generated from 5 male C57BL/6 mice (GSE191226) ¹⁷. In this dataset, the aortas were divided into five segments: the ascending aorta and aortic arch (AOAR), the upper descending thoracic aorta (TA1), the lower descending thoracic aorta (TA2), the suprarenal abdominal aorta (AA1), and the infrarenal abdominal aorta (AA2). In all aortic segments, Carmn was mostly expressed in SMCs, characterized by high expression of the SMC marker gene Myh11 (Fig. 1A). Analysis of average expression of Carmn along with SMC markers (Myh11, Acta2, Cnn1, Tpm1, Tpm2, Tagln) revealed that infrarenal abdominal aorta exhibited the highest level of Carmn expression, followed by ascending aorta and aortic arch. Descending thoracic aorta expressed much lower level of Carmn (Fig. 1B).

Figure 1:

Expression and distribution of LncRNA Carmn in different segments of the naïve mouse aorta. (A) The distribution of SMC subtypes as well as expression of Myh11 and Carmn in different segments were manifested by the UMAP plot generated from the reanalyzed publicly available dataset (GSE191226). Only SMCs were labeled on cluster distribution plot. (B) Dot plot showing expression of LncRNA Carmn along with other contraction associated genes in different segments of aorta and different subtypes of SMCs. (C) Stacked bar graph of proportion of SMC subclusters among all SMCs in different segments of aorta. SMC, smooth muscle cells; AOAR, ascending aorta and aortic arch; TA, descending thoracic aorta; AA, abdominal aorta.

The transcriptomics of SMCs showed six distinct subclusters (Fig. 1B). While all clusters express markedly high levels of SMC marker genes compared to non-SMC cells, variations in Myh11, Acta2, Cnn1 and other makers were found among the clusters. For example, SMC-1 and SMC-2 were characterized by relatively low levels of Myh11 but high levels of Acta2 and Cnn1 (Fig. 1B). While both represented the major subclusters, the presence of SMC-1 was the highest in the abdominal aorta (AA1 and AA2, accounting for approximately 39.7% and 41.1% of cells, respectively), whereas SMC-2 was the largest population in the descending thoracic aorta (TA1 and TA2, accounting for approximately 59.2% and 54% of cells, respectively) (Fig. 1C). The region-dependent distribution was also observed in the smaller subsclusters. For instance, SMC-5 was more prominent in AOAR than other segments, whereas SMC-4 was more prominent in AA2 (Fig. 1C).

The level of Carmn expression was uneven among the six SMC subclusters. Paradoxically, Carmn was most abundant in SMC-5. However, this particular subcluster represented only a small fraction of the overall SMC population, and showed the lowest expression of SMC markers such as Myh11, Acta2, Cnn1, and Tagln (Fig. 1A and B).

Carmn expression is decreased in experimental AAA mouse models

To investigate potential alterations in Carmn expression in AAA, we first conducted a comprehensive analysis of published datasets on experimental AAA models. We analyzed scRNA-seq datasets from the Angiotensin II (Ang II) infusion model in Apoe^−/− mice (GSE221789) ¹⁸ and the CaCl₂ model in C57BL/6 mice (GSE164678) ^19,20. Additionally, we analyzed two bulk RNA-seq datasets, one from porcine pancreas elastase (PPE) infusion model in C57BL/6 mice (GSE153097) ²¹, and another from DOCA/high salt induction model (GSE153425) ²².

In agreement with Yu’s dataset from naïve C57BL/6 mice, datasets from the mouse AAA models showed that Carmn expression was mostly enriched in SMCs, identified by high expression of the SMC marker genes including Myh11 (Fig. 2A and C). Compared to sham controls, Carmn expression in the AAA groups ---in both the Ang II (Day 28) and CaCl₂ (Day 4) models --- was markedly reduced. The decrease in Carmn expression was consistent with the AAA-associated reduction in contractile genes, including Myh11, Acta2, Cnn1, Tpm1, Tpm2, and Tagln (Fig 2B and D).

Figure 2:

Expression of Carmn in experimental AAA models. (A&B) scRNA-seq analysis of Carmn expression in Day 28 Angiotensin II (Ang II) induced AAA model in Apoe^−/− mice. (C&D) scRNA-seq analysis of Carmn expression in Day 4 CaCl2 induced AAA model in C57BL/6 mice. (E) Bulk RNA-seq analysis of Carmn, Myh11, and Cnn1 expression at one or two weeks after AAA induction in the porcine pancreas elastase (PPE) induced AAA model using bulk RNA-seq. One-way ANOVA was used for statistical analysis. *p<0.05. **p<0.01. ns: not significant.

The bulk RNA-seq dataset generated by Gabel et al contain two timepoints following AAA induction by PPE (GSE153097) ²¹, week 1 and week 2, which allowed us to determine how Carmn expression is altered during the development of AAA. We found that the expression levels of Carmn, along with SMC markers such as Myh11 and Cnn1, were significantly diminished at week 1. However, by week 2, these expression levels were restored to the normal level (Fig. 2E). A trend of reduction of Carmn, Myh11, and Cnn1 was observed in the DOCA/high salt-induced experimental AAA mouse model. However, due to substantial variations within the control group, the downregulation did not reach a statistical significance (Supplementary Fig. S1A).

The available scRNA-seq data of human AAA samples did not contain non-coding RNAs, likely due to the low abundance and sequencing depth. To determine whether CARMN downregulation occurs in human AAA, we analyzed a microarray dataset derived from human AAA samples (GSE7084) ²³. We observed lower levels of CARMN in AAA tissues than the controls (Supplementary Fig. S1B). However, we were unable to perform statistical analysis due to the limited availability of data. Only one AAA sample and one control sample, both of which were performed on the GPL570 array platform, included the human CARMN gene.

Carmn expression determined by RNA fluorescent in situ hybridization (FISH)

Next, we sought to validate the bioinformatical findings in mouse AAA models. We first utilized the AAV PCSK9/Ang II model, as described by Lu et al. ¹² , which induced aneurysmal dilatations in the suprarenal abdominal aorta, as well as other regions (Fig. 3A–C). Since the bulk RNAseq analysis in mouse PPE model suggests the reduction of Carmn expression may occur during the early phase of aneurysm development, we prepared aortic cross sections from the suprarenal abdominal aorta after 3 and 7 days of Ang II infusion. RNA FISH was utilized to label Carmn and Myh11 in the abdominal aortic wall. As shown in Fig. 3D&G, both Carmn and Myh11 were exclusive localized in the tunic media, presumably in SMCs. Interestingly, Carmn appeared within the nuclei of SMCs in the tunica media layer, whereas Myh11 mRNA was predominantly present in the perinuclear area of SMCs.

Figure 3:

Fluorescent in situ hybridization (FISH) analysis of Carmn expression and distribution in the AAV PCSK9 Angiotensin II (Ang II) mouse AAA model. (A) Diagram of the AAA model. (B&C) Representative photos (B) and percentage increase of maximal external aortic diameter of AAA (C) after 28 days of Ang II infusion. (D-I) Representative images and quantifications of Carmn or Myh11 expression in Day 3 or Day 7 PCSK9/Ang II AAA tissue. Student t-test was used for statistical analysis. *p<0.05. **p<0.01. ns: not significant.

When comparing the AAA group to the sham group, we observed a significant decrease in the number of Carmn puncta within the media layer of Ang II-treated tissues. Specifically, 3 days of Ang II infusion caused a significant reduction (61.1%) in Carmn expression as well as in Myh11 (47.4%) (Fig. 3E&F). By day 7, the expression of Myh11 appeared to be recovered, while the reduction Carmn persisted but tempered to 56.2% (Fig. 3H&I).

Because different mouse AAA models capture different aspects of human AAAs ²⁴, we repeated the validation study in the CaCl₂ model. As we previously reported, perivascular application of CaCl₂ caused a 70% expansion in infrarenal aorta measured 28 days post-surgery model ¹¹ (Fig. 4A–C). RNA FISH analysis showed an approximate 75.9% reduction in Carmn on day-4 post CaCl₂ induction (Fig. 4D&E). Myh11 was reduced at the same time, however the reduction did not reach the statistical significance (Fig. 4F).

Figure 4:

FISH analysis of Carmn expression and distribution in the CaCl₂ induced mouse AAA model. (A) Diagram of the AAA model. (B&C) Representative photos (B) and percentage increase of maximal external aortic diameter of AAA (C) 28 days after AAA induction. (D-F) Representative images and quantifications of Carmn or Myh11 expression in Day 4 AAA tissues. Student t-test was used for statistical analysis. *p<0.05. ns: not significant.

CARMN distribution in human AAA wall

To assess CARMN expression in human AAA tissue, we collected samples from both ruptured and non-ruptured AAA patients. Cross-sections of these samples were subjected to RNA in situ hybridization. Consistent with what was observed in mouse aortic samples, CARMN expression was localized to the nuclei within the tunica media layer in both ruptured and non-ruptured AAA tissues (Fig. 5-2). Strong CARMN signals were also detected in the vasa vasorum of the adventitia, while no detectable CARMN signal was found in the intima or in non-vasa vasorum areas of the adventitia (Fig. 5-1 and 3). Due to the lack of age-matched non-aneurysm infrarenal aortic tissues, we were unable to evaluate how CARMN may be altered by AAA. However, we did not observe meaningful differences in CARMN expression between intact AAA and ruptured AAA tissues.

Figure 5:

Distribution of CARMN in human non-ruptured or ruptured AAA tissue analyzed by in situ hybridization. Magnified images are indicated by red boxes. CARMN positive signals were highlighted by red arrows.

Inflammatory macrophages increase Carmn expression in SMCs

We have previously reported that macrophage-to-SMC signaling is the strongest cell-cell crosstalk during AAA ²⁰. To examine whether signals released by macrophages modulate Carmn expression in SMCs, we derived macrophages from the bone marrow of C57BL/6J mice. Macrophages were treated with BSA (control), LPS+IFNγ, or IL4 for 24 hours. The inflammatory status of treated macrophages was evaluated based on the expression of Il1b and Arg1. Consistent with the literature, LPS+IFNγ-treated macrophages showed significantly higher expression of Il1b compared to the control (Fig. 6A), indicating a pro-inflammatory state. In comparison, IL4-treated macrophages showed reduced levels of Il1b but elevated Arg1 expression compared to the control (Fig. 6A), indicating an anti-inflammatory state. Henceforward, we refer LPS+IFNγ - treated macrophages as M1, IL4-treated as M2, and BSA-treated as M0, albeit macrophages exist in aneurysm tissues as a continuum spectrum of phenotypes.

Figure 6:

Inflammatory macrophages decrease Carmn expression in SMCs. (A) Bone marrow–derived macrophages (BMDMs) were treated with 20 ng/ml IL4 or BSA for 24 hours; or primed with 100 ng/ml LPS for 2 hours, then treated with 20 ng/ml IFNγ for 24 hours. Cells were analyzed by real-time PCR. (B&C) BMDM were “polarized” as described in (A). At the end of treatment, BMDM were thoroughly washed three times with PBS and incubated with fresh DMEM containing 0.3% FBS and 20 ng/ml M-CSF for 24 hours. Conditioned medium was collected, and potential cell debris was removed by centrifugation. Primary mouse aortic SMCs, following a 48 hours of serum starvation with 0.3% FBS, were incubated with BMDM conditioned medium for 48 hours. Cells were lysed for RNA isolation and analysis by real-time PCR. One-way ANOVA was used for statistical analysis. *p < 0.05. **p<0.01. ****p<0.0001 ns: not significant.

Next, we investigated how different macrophage states may affect Carmn expression using primary aortic SMCs isolated from C57BL/6J mice. As shown in Fig. 6B, treating SMCs for 48 hours with the M1 macrophage conditioned medium reduced levels of Carmn relative to SMCs cultured in regular cell culture medium. In contrast, treatment with M0- or M2-conditioned medium did not alter Carmn expression (Fig. 6B). Similarly, the contractile genes Myh11, Cnn1, Tagln, and Acta2 responded to M1-conditioned medium with reduced expression, with Cnn1 mostly affected (Fig. 6B). However, Carmn expression was not affected by M2-conditioned medium, while a statistically significant upregulation was noted in all the contractile genes examined (Fig. 6B). The expression levels of Carmn and contractile genes, except for Acta2, were not altered by M0-conditioned medium (Fig. 6B). To examine the inflammatory state of SMCs, we evaluated the expression of Ccl2 and Il6 by SMCs in the regular medium or macrophage-conditioned medium. Compared to the regular medium, M1-conditioned medium increased Ccl2 and Il6 expression by ~1800 and ~200 folds, respectively (Fig. 6C). In contrast, M0- or M2-conditioned medium had no effects on the expression of these pro-inflammatory genes (Fig. 6C).

Discussion

In this study, we investigated the expression and distribution of CARMN in AAA through in silico analysis of published datasets, and validation in experimental AAA models and human AAA tissues. We found Carmn enriched in infrarenal abdominal aorta in mouse, and its expression was reduced in mouse AAA models. In human AAA tissue, CARMN was mainly expressed in the nuclei of the tunica media and vasa vasorum. Our findings reveal CARMN expression and regulation in AAA contexts.

CARMN is a conserved lncRNA that is specifically expressed in adult SMCs with some expression in the fetal heart^4,25. By analyzing publicly available datasets, we identified a consistent reduction in Carmn/CARMN expression across four distinct mouse AAA models and human AAA tissues. Data from the PPE-induced AAA model suggests that the expression of Carmn, along with SMC markers such as Myh11 and Cnn1, was affected during the early phase of aneurysm development. By week 2, the expression levels of Carmn and contractile genes were restored to normal. However, in human AAA tissues, which were collected at an advanced stage, we inferred reduced CARMN expression. Again, this discrepancy emphasizes the limitation of mouse models particularly the acute AAA models such as the PPE model, which aneurysmal dilatation occurs rapidly and plateaus within weeks. It would be interesting to examine the time course of Carmn expression in an animal model in which aneurysm grows continuously and eventually ruptures.

Dong et al has demonstrated that Carmn is enriched in the aorta ⁴. Our current findings on Carmn distribution in different segments of the aorta tree extend the knowledge. Given the function of CARMN in SMC differentiation, we speculate that various regulatory mechanisms, such as origins of the SMCs and the flow patterns in different aortic segments, may govern the expression of this lncRNA. Of note, human AAAs predominantly occur in the infrarenal abdominal aorta. Although the full explanation for this regional vulnerability remains elusive, hemodynamic factors are believed to play a significant role. Whether the prominent presence of Carmn in mouse infrarenal aorta can be extended to human, and more importantly, how CARMN may contribute to the vulnerability of the abdominal aortic region to aneurysm formation, are intriguing questions. Unfortunately, the regional differences in Carmn expression impose challenges to investigate this lncRNA in human AAAs as age-matched healthy infrarenal aortic tissues are rarely available. However, we were able to compare aortic samples from ruptured and non-ruptured AAAs but found no difference in CARMN expression.

Our results highlight the consistent reduction of Carmn expression in AAA models, indicating its potential involvement in AAA pathogenesis. The observed correlation between Carmn downregulation and decreased expression of SMC markers further suggests a role for Carmn in modulating SMC phenotype during AAA development. Mechanistically, it is probable that CARMN directly interacts with the transcriptional cofactor myocardin, forming the CARMN/MYOCD/SRF complex ^4,26. This complex binds to CArG elements within the promoter regions of genes responsible for SMC contraction. As a result, the presence of this complex helps maintain the contractile phenotype of SMCs ^4,26.

Our in situ analyses of aortic tissues revealed distinct distribution patterns of Carmn and Myh11 mRNAs. The localization of Carmn in the nuclei of SMCs is consistent with observations reported by Ni et al⁵ and Dong et al⁴, suggesting that the function of Carmn primarily takes place in the nucleus . The observed subcellular localization of Myh11 mRNA is interesting. A similar perinuclear localization was reported for alpha myosin heavy chain mRNA (a-MyHC) in cardiac myocytes by Goldspink and colleagues²⁷. The authors showed that a-MyHC mRNA in spontaneously contracting myocytes is dispersed throughout the cytoplasm but becomes restricted to areas around the nucleus when contraction is arrested by a calcium channel blocker²⁷. Whether the similarly restricted Myh11 mRNA distribution is associated with a less SMC contractile state needs to be investigated.

Multiple pathological factors present in aneurysm tissues such as inflammation, oxidative stress, and degradation of extracellular matrix proteins could all contribute to the downregulation of Carmn. Our previous scRNA-seq analysis revealed that signals sent from inflammatory macrophage to contractile SMCs was most elevated in early stage of AAA compared to sham control ²⁰. The notion that macrophages modulate SMCs in aneurysm is supported by experimental data. For example, Hadi et al reported that macrophage-derived netrin-1 could activate MMP3 in SMCs ²⁸, and Lu showed that galectin-3 expressing macrophages promoted SMC apoptosis ²⁹. Furthermore, in the context of atherosclerosis, ox-LDL-treated macrophages secrete IL-1β, which consequently downregulates α-SMA in SMCs by inhibiting STAT3 ³⁰. In this study, we showed that M1-like macrophages reduced the expression of Carmn and contractile genes, and increased the expression of pro-inflammatory genes by SMCs. Since these modulations were produced by macrophage-conditioned medium, it is plausible to speculate direct cell-cell contacts might not be necessary for macrophages to modulate the behavior of SMCs during AAA development. In relation to the downregulation of Carmn by M1-like macrophages, Ni et al reported that TGF-β1 can induce Carmn in a Smad2/3 dependent manner⁵. It is possible that M1-like macrophages secrete less Carmn-promoting TGF-β1. Another possible explanation is that M1-macrophages produce factors that actively suppress Carmn expression.

The functional role of Carmn has been investigated in other vascular diseases through depletion or overexpression experiments conducted in vivo. Vacante et al. generated Carmn global knockout mice and observed increased plaque volume and size in an AAV PCSK9 high cholesterol diet atherosclerosis model ⁶. Similarly, Dong et al. generated SMC-specific Carmn deficient mice and observed significantly exacerbated neointima formation in a carotid artery ligation model. Conversely, they also demonstrated that overexpression of Carmn using adenovirus attenuated neointima formation in rats after balloon injury ⁴. However, in contrast to these findings, Ni et al. conducted experiments in Ldlr^−/− mice where they utilized antisense oligonucleotides to knock down CARMN. Their results showed a significant reduction in the formation of atherosclerotic lesions and a suppression of SMC proliferation, while the apoptosis was not affected ⁵. The presence of inconsistent results highlights the intricate nature of Carmn’s function. Therefore, additional studies are needed to investigate and manipulate Carmn expression specifically within the context of AAA. These further investigations will help to elucidate the functional role of CARMN in AAA and explore its potential as a promising therapeutic target.