Gasdermin D deficiency attenuates development of ascending aortic dissections in a novel mouse model

Abstract

Background:

Mechanisms driving development of type A aortic dissection (TAD) are currently poorly understood, and animal models of spontaneous TAD are limited. In the present study, we developed a novel mouse TAD model and evaluated the role of gasdermin D (GSDMD) in TAD development.

Methods:

TADs were created by treating the ascending aorta of adult C57BL/6J mice with active elastase and β-aminopropionitrile (Act E+BAPN). The temporal progress of the TAD pathology was rigorously characterized by histological evaluation and scanning electron microscopy, while potential mechanisms explored using bulk RNA sequencing of specimens collected at multiple timepoints. With this novel TAD model, we conducted additional experiments to investigate the impact of GSDMD deficiency (Gsdmd^−/−) on TAD formation.

Results:

Ascending aortas challenged with Act E+BAPN developed pathology featuring early onset of intimomedial tears (complete penetration) and intramural hematomas, followed by progressive medial loss and aortic dilation. Ingenuity Pathway Analysis and functional annotation of differentially expressed genes suggested that a unique inflammatory micro-environment, rather than general inflammation, promotes the onset of TADs by specifically recruiting neutrophils to the aortic wall. At later stages, T-cell mediated immune injury emerged as the primary driver of pathology. Gsdmd^−/− attenuated medial loss, adventitial fibrosis, and dilation of TADs. This protective effect correlated with a reduced cell death and decreased T-cell infiltration in TADs. Notably, cleaved GSDMD was detected in human TADs but was absent in healthy aortas.

Conclusions:

A novel mouse TAD model was developed, specifically targeting the ascending aorta. This model generates a unique microenvironment that activates specific immune cell subsets, driving the onset and subsequent remodeling of TADs. Consistently, Gsdmd^−/− mitigates TAD development, likely by modulating cell death and T-cell responses. This model provides a valuable tool for studying immune injury mechanisms in TAD pathogenesis.

Graphical Abstract

Introduction

Aortic dissection is a silent killer, with an estimated incidence of 30 to 50 cases per 1 million person-years¹. The thoracic aorta, particularly the ascending aorta is most frequently affected². Although relatively uncommon, aortic dissection is a catastrophic life-threatening cardiovascular event. Currently, diagnosis can only be made after dissection occurs due to the lack of biomarkers capable of identifying at-risk population. Despite the availability of emergency medical care, approximately, approximately 50% of the cases result in death before the patient reaches the hospital³. Improving clinical outcomes for aortic dissections remains an urgent, unmet need.

Our current understanding of the etiology of aortic dissections is quite elusive. While roughly 25% of the cases may be attributable to genetic mutations, the cause remains unknown for majority of the cases ⁴. This gap in knowledge poses significant challenges in developing animal models that faithfully recapitulate the pathophysiology of human aortic dissections. Among existing models, the angiotensin II (AngII) model is the most widely adopted, with various modifications introduced to enhance the rate of TAD formation^5,6. Because of the limitations of individual models, we and others developed an alternative model by inducible deletion of SMC (smooth muscle cell) specific TGF-β receptors^7–9, for cross-model validation of the scientific findings. However, this approach is time-consuming and susceptible to off-target genome modifications, which can complicate data interpretation¹⁰. These challenges highlight the need for alternative options.

TAD is a separate aortic disease that differs from the thoracic aortic aneurysm (TAA) in etiology, pathophysiology, and clinical management^4,11. However, the two conditions share certain pathophysiological mechanisms that drive medial degeneration, rendering the aortic wall prone to dissection and/or dilation. For instance, programmed cell death has been implicated as a significant contributor to the development of both TADs and TAAs¹². Gasdermin D (GSDMD), a membrane pore-forming molecule, is a key executioner protein for pyroptosis¹³. Recent studies suggest that GSDMD may be activated through pathways independent of pyroptosis, such as panoptosomes¹⁴. Furthermore, studies have identified GSDMD as a driver for AAA formation^15,16 and rupture¹⁷, yet its role in the development of TADs remains unexplored. In the present study, we developed a novel mouse TAD model to study the role of GSDMD in TAD formation.

Materials and Methods

Data Availability

Data supporting the findings of this study, including quantification for western blots, is available from the corresponding author upon reasonable request. RNA sequencing data have been submitted to the Gene Expression Omnibus (GEO) and can be accessed at https://www.ncbi.nlm.nih.gov/geo/.

Animal Studies

This study conforms to the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The Institutional Animal Care and Use Committee of the University of Florida approved all procedures. Mice used in this study were on a C57BL/6J background. Sex has been identified as a significant factor in multiple mouse model of TADs^18,19. Since sex-dependent TAD formation was not the focus of this study, only male mice were used. Both wild type (WT, 000664) and Gsdmd^−/− (032663) mice were purchased from Jackson laboratory and were allowed to acclimate for at least one week prior to enrollment in the study.

Elastase, active (Act E) or heat inactivated (Inact E) was delivered to the outer surface of the ascending aorta via minimally invasive surgeries under a dissecting microscope. Briefly, mice were anesthetized by continuous inhalation of 1.5–2.0% isoflurane and secured on a supine position. After removing the fur and disinfection of the surgical area, a midline incision (around 1.0–1.5 cm) was made, extending from the lower neck to the upper chest. A straight scissor was used to cut the manubrium longitudinally, followed by opening the manubrium with a custom-made retractor. The thymus was separated bluntly from the anterior wall to expose the ascending aorta and aortic arch, followed by isolating them from peri-aortic tissue with fine forceps. Elastase (30.0 μl) was then loaded to two cotton patches, with each specifically placed to wrap the back and front of the ascending aorta. After a five-minute incubation, the cotton patches were removed. A 4–0 absorbable suture was used to restore the manubrium and to close the skin incision. Ethiqa-XR, a slow-releasing version of buprenorphine, was administered via a s.c. injection to alleviate pain and stress. In our hands, the successful rate is around 90%, with no surgery-related post-op death noted so far. Beta-aminopropionitrile (BAPN, 0.2%) was delivered on the day of surgery via drinking water and refreshed twice weekly.

Ultrasound imaging

Animals were anesthetized via inhalation of 1.5–2.0% isoflurane and placed in a supine position on a heated platform. Naire was used to remove the fur from the imaging area and then wiped off with warm water. A high-resolution Vevo 2100 Imaging System with a MS550D (central frequency: 40 MHz) linear array transducer (VisualSonics) was then utilized to scan aortas under B-mode. Ascending aortas were imaged via the right parasternal long axis view, with the transducer finely adjusted to a position that captures the largest possible width of the ascending aorta. A cine-loop was then recorded and reviewed to select the frame that displays the widest lumen. On the selected image, lumen diameter was measured using the straight-line tool, with the line placed and oriented so that the length of the line equals the largest lumen diameter of the ascending aorta^8,9.

Histology

Evans blue staining was performed as described previously⁵. Briefly, Evans blue solution (5%) was administered through tail vein injection 30 minutes before tissue collection. During tissue collection, saline was injected through the left ventricle to thoroughly remove Evans blue from the circulatory system, followed by perfusion fixation with 10% neutral buffered formalin. The TADs were cut open longitudinally and spread on slides with the luminal side facing up. After mounting with coverslips, the specimens were evaluated microscopically (Zen lite 2012, Carl Zeiss, Oberkochen, Germany) for Evans blue extravasation and presence of intimal/medial tears.

Van Gieson and Prussian blue staining were performed on paraffin sections. While a standard protocol was followed for Va Gieson staining, a kit (ab150674, Abcam, Cambridge, United Kingdom) was used for Prussian blue staining by following the manufacture’s instruction.

Morphometry

Cross-sections (5.0 μm) were collected at locations 0, 100, and 200 μm from the proximal end of the TADs, with the point displaying the first complete circle taken as the position-0 (p0). Specimens were stained using the Van Gieson protocol and imaged with a Zen blue software (Carl Zeiss, Oberkochen, Germany). Both internal and external elastic laminae were traced to obtain area of media and length of the external elastic laminae (EEL), while area of the adventitia was obtained by tracing the outer boarder of the layer with dense matrix. For each TAD sample, the area of the media and adventitia was normalized to the length of EEL. To evaluate medial loss, surface length was measured for areas where the media contains ≤ 4 layers of elastic lamina (the mouse ascending aorta generally has 8 layers of elastic laminae) or displays ≥ 50% reduction in thickness. This criterion was set so that the starting and the ending points are standardized when lines were drawn to measure medial loss. For a given section, length of lines was summed up, followed by normalization to EEL.

TUNEL Assays

TUNEL assays were performed using a One-step TUNEL In Situ Apoptosis Kit (E-CK-A322, Elabscience) according to the manufacturer’s instructions. Aortas of naïve mice were used as a biological negative control for protocol optimization. Staining was evaluated independently by two observers blinded to the treatment information. The number of TUNEL⁺ cells was counted on individual sections, and the average count from the three locations was calculated for statistical analysis. For fluorescence double labeling of dying SMCs, FITC-conjugated dUTPs/dNTPs (11767291910, Roche) were used for TUNEL labeling, followed by the immunostaining of α-actin with Cy3-conjuated anti α-actin antibody (C6198, Sigma).

Immunohistochemistry (IHC) and immunofluorescence (IF) Assays

IHC and IF assays were performed on paraffin sections collected at the location of P100, following protocols established in our laboratory previously^9,20. Briefly, specimens were rehydrated and antigens unmasked by incubating specimens in Citra buffer (pH 6.0, H3300, Vector Laboratories) heated in a pressure cooker. After blocking non-specific bindings, specimens were incubated with primary antibodies at 4°C overnight, followed by probing secondary antibodies conjugated with a Fluorophore or biotin. For IHC assays, antigen-specific signals were amplified with an ABC kit (PK-6100, Vector Laboratories) and developed with a DAB kit (SK-4100, Vector Laboratories) Nuclear counterstain was completed with hematoxylin and DAPI (D9542, Sigma-Aldrich) for IHC and IF labeling, respectively.

IHC assays were evaluated under a bright field microscope by two observers blinded of grouping information. Both observers graded density of positive cells using a 5-scale scoring system. In cases where the difference was greater than 2 points between the two observers, assays were graded by a third reviewer. An average score was calculated to represent positivity of the assay on each specimen. IF assays were graded following the same protocol as used for IHC assays except reviewing under a fluorescent microscope.

Western Blot assays

Total protein was extracted from TADs with RIPA buffer (89900, Thermo Fisher Scientific), supplemented with protease and phosphatase inhibitor cocktails (1861281, Thermo Fisher Scientific). Protein concentrations were determined using the Pierce BCA Protein Assay kit (23225, Thermo Fisher Scientific). For each sample, 30 μg of total protein from each sample was separated on SDS-PAGE gels and transferred onto PVDF membranes. Membranes were incubated with primary antibodies overnight at 4°C, followed incubation with IRDye 680RD Donkey anti-mouse IgG (926–68072, LICOR) and IRDye 800CW Donkey anti-rabbit IgG (926–32213, LICOR) for 1 hour at room temperature. Blots were imaged using an Odyssey^® CLx imager (LICOR) and quantified using the imaging software. The results were normalized to GAPDH expression for comparative analysis.

Flow cytometry

Flow cytometry analysis was performed following the protocol described previously. Briefly, approximately 100 μL of blood was collected via facial vein bleeding and lysed (lysis buffer; 000–4300-54, Thermo Fisher Scientific) to remove red blood cells. The remaining white blood cells were resuspended in PBS containing 2% FBS. After blocking with a rat anti-CD16/CD32 antibody, cells were stained with an antibody (rat IgG) cocktail containing Zombie Aqua-live/dead, PerCP-CD45, FITC-CD3, PE-CD4, APC-Cy7-CD8, Pacific Blue-CD19, APC-Ly6C, and Alexa fluor 700-Ly6G, all purchased from BioLegend. Each antibody was used at a predetermined dilution factor specified in the Major resources Table. Stained cells were analyzed using a 5-laser Aurora flow cytometry system (Cytek), and the data were processed with FlowJo software (BD Biosciences).

Bulk RNA sequencing

Total RNAs were extracted using a TRIzol plus RNA Purification kit (12183555, Invitrogen) following manufacturer’s instructions. An aliquot was collected for every sample and analyzed with an RNA 6000 Pico Kit (5067–1513, Agilent) on an Agilent 2100 Bioanalyzer (Agilent, CA). All RNA samples produced a RIN > 8.0 while those entering next step had a RIN ≥ 8.5.

Illumina Stranded mRNA Prep and Ligation Kit (Ref# 20040534; Illumina, CA) was used to prepare sample libraries following the manufacturer’s recommendations. Briefly, high-quality total RNA (100–300 ng/μL) was captured and purified, followed by reverse transcription to convert the captured mRNA into first and second strands of cDNA. The resulting double stranded cDNA molecules were then enzymatically fragmented in preparation for unique dual indexing using IDT for Illumina RNA UD Set Ligation Indexes (Ref# 20040534, Illumina). To facilitate ligation of the barcoded adapters, adenine (A) nucleotides was added to the 3ʹ ends of the blunt fragments preventing them from ligating to each other during adapter ligation. Additionally, a corresponding thymine (T) nucleotide on the 3ʹ end of the adapter provided a complementary overhang for ligating the adapter to the fragment. The ligated samples were subsequently enriched through amplification (13x PCR cycles) on Bio-Rad Thermal cycler (T100; Bio-Rad, CA) and purified using AMPure XP magnetic beads (Ref# A63881; Beckman Coulter, CA). The average fragment length of libraries was measured to be ~300–400 bp using the DNA 1000 Kit (Agilent, CA) on Agilent 2100 Bioanalyzer. Finally, the optimized libraries were pooled to a final concentration of 750pM. Afterwards, they were loaded onto Illumina’s P2 flow cell/200-cycle reagent cartridge (Ref# 20100986, Illumina) for sequencing on the NextSeq 1000 instrument (Illumina, CA). The run setup for the Illumina NextSeq 1000 was set to a length of 100bp reads (Read1: 100; Index1: 10; Index2: 10; Read2:100). Denature and Dilute On Board was enabled and a 2% PhiX spike-in was added as positive control. To ensure quality, multi-QC Analysis was conducted for each sequence run using Illumina’s Dragen RNA software (Illumina). Once sequencing was completed, the raw sequence base call files (BCL) were converted to FASTQ files using Illumina’s bcl2fastq conversion software (Illumina, CA) as input for downstream bioinformatics analysis.

Bioinformatics

The raw sequencing reads were quality controlled using FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and mapped to the mouse GRCm39 genome assembly as well as Gencode M32 annotations using STAR aligner²¹. RSEM²² were applied to summarize gene counts, which were further normalized using the TMM method²³ and gene-wise compared between experimental groups using edgeR²⁴ based on negative binomial generalized linear models. The Benjamini & Hochberg procedure was applied to adjust P-values for multiple comparison. Differentially Expressed Genes were identified with an FDR of 10 and an adjusted P<0.05. Principal component analysis was performed on the scaled logarithm transformed FPKM (fragments per kilobase per million) value and samples were examined in the space of the first three principal components (PC1, PC2 and PC3).

IPA (Ingenuity Pathway Analysis) “Core Analysis” was performed to identify canonical pathways, and their upstream regulators predicted to be accountable for DEGs^25,26. A cutoff of 2.0 and 10.0 was set for Z score and P value, respectively. Pathways associated with specific diseases, such as arthritis and cancer, were excluded from the graphical summary due to the irrelevance of those themes to the present study. Functional annotation of DEGs was performed with DAVID^25,26. With an EASE score of 0.01 and a minimum gene count of 10, KEEG pathways were enriched and the annotation chart exported as a text file. Following collapsing of redundant themes, data were plotted as a dot plot.

Statistical Analysis

All data are expressed as the mean ± SEM. Statistical analyses were performed using Sigma Plot 14.0 (San Jose, CA, USA). Datasets were evaluated using normality and equivalence variance testing. For those failing this evaluation, logarithmic and exponential transformations were performed to meet these requirements. Student’s t-test, one-way ANOVA, two-way repeated measures ANOVA, and Mann-Whitney Rank Sum test were performed, when appropriate, with Holm–Sidak analysis being used for post hoc tests. P<0.05 was considered statistically significant.

Results

Elastase and BAPN work in concert to promote TAD formation

Elastase is one of the proteases that are commonly used for the creation of animal models of TAAs and AAAs²⁷. These models appear to develop aortic aneurysms, with rare aortic dissection or rupture occurrence. In the abdominal aorta, the combined challenge of elastase and BAPN induces AAAs with a phenotype featuring persistent aortic dilation and luminal thrombosis²⁸. These results predicted that systemic delivery of BAPN in combination with local elastase insult would yield a model of TAA. However, this same treatment resulted in a phenotype of dissection rather than aneurysm dilation in the ascending aorta; and therefore, the term TAD is used to indicate the phenotype throughout the present study.

The ascending aorta was approached through a midline incision (1.0–1.5 cm) crossing the lower neck and the upper chest and isolated without opening the chest. Elastase (30.0 μl) was loaded to two cotton patches (15.0 μl/piece) and delivered to the outer surface of the ascending aorta by wrapping it from both the anterior and the posterior side for five minutes. Since both low and high concentrations of elastase were used to induce AAAs^28,29, we performed a set of experiments to determine the appropriate dosage for TAD induction. C57BL/6J mice (male, 11–14 weeks of age) were randomized to three groups (n=3–5 per group), with the ascending aorta exposed to active elastase (Act E) at a low (30.0 μl at 4.0 U/ml) or a high dose (30.0 μl at 40.0 U/ml), or heat inactivated elastase (Inact E, 30.0 μl heated high dose solution). BAPN (0.2%) was administered to all three groups, beginning on the day of surgery. Ultrasound imaging showed that the low dose elastase failed to promote aortic dilation in a four-week follow-up period. The high dose elastase resulted in aortic dilation. However, the temporal pattern of aortic dilation differed from that documented for elastase-induced AAAs. Specifically, early dilation (< 1 week) was not dramatic. The overall expansion was much less than 100% (Figure S1A) as seen in the AAA model²⁹.

Because of the different aortic phenotype, we went on to perform rigorous model characterization. C57BL/6J mice (male, 11–14 weeks of age) were randomized to receive treatments as follows (n=6–8 per group): Act E+BAPN, Act E, Inact E+BAPN, or Inact E. Ultrasound imaging showed that both the Act E+BAPN and the Act E promoted an early (in a week), moderate aortic dilation (23 ± 4 % and 23 ± 3 % for Act E+BAPN and Act E, respectively). Thereafter, aortas exposed to Act E+BAPN continued to expand and caused a 71 ± 13% increase in lumen diameter in four weeks. In contrast, those exposed only to Act E displayed a halted progression, showing insignificant changes in diameter during the rest of the follow-up period. The groups challenged with Inact E+BAPN or Inact E only did not result in significant aortic dilation in the entire four weeks (Figure 1A). Peri-aortic adhesion and aortic dilation are the typical pathologies noted during gross examination. Absorbed intramural hemorrhage (yellow coloration of the aortic wall) was detected occasionally (Figure 1B and S1B). Fresh intramural hemorrhage (presence of red clot or dots) was not detected at the four-week time point. Aortic rupture was not noted for this cohort during the four-week follow-up.

Figure 1.

A reliable mouse TAD model created with minimally invasive surgery. Specimens were collected four weeks after the TAD induction. A, dilation of TADs challenged with indicated conditions (C57BL/6J, male, 10 to 14 weeks old, n=6–8). Data was analyzed using two-way RM ANOVA. Time: P<0.001; treatment: P<0.001; Time × treatment: P<0.001; *P=0.0050, Act E+BAPN vs. Act E. Inact E: heat-inactivated elastase; Act E: activate elastase without BAPN; Act E+BAPN: active elastase with BAPN. B, representative gross images of aortas subjected to indicated treatments. Ruler scale: 1.0 mm. C, Van Gieson staining of TADs. High power images (lower row) depict typical pathologies developing under indicated conditions. D, morphometric evaluation (n=6–7 per group). Data was analyzed using two-way ANOVA.

To further characterize the TAD pathology, we performed serial sectioning at a 100 μm interval for specimens collected at the four-week time point. An example is provided in Figure S1B and S1C. Elastic fiber breaks, intimomedial tears, medial thinning or complete loss, and adventitial fibrosis were readily detectable in aortas exposed to Act E with or without BAPN, while not present in those treated with Inact E only (Figure 1C). Blood clots were occasionally located between elastic laminae (Figure S1D) in TADs induced by Act E+BAPN (3 in 8) or Act E (2 in 8). This pathology was defused along the circumference of the aorta except the area adjacent to the pulmonary artery where medial thinning or complete loss was less frequently observed. Along the long axis of TADs, the whole ascending aorta presented the pathologies described above with little variation, except the proximal end where medial loss was not detected in some cases (Figure S1C). Aortas treated with Inact E +BAPN were not free of pathology. Elastic fiber breaks and partial intimomedial tears were frequently detected, although medial thinning was absent in these aortas (Figure 1C). Because of the massive medial loss, we decided not to count elastic fiber breaks or intimomedial tears. Rather, we quantified severity of medial loss (defined as area with ≤3 layers of elastic lamina, Figure S2A and S2B). Act E+BAPN or Act E alone resulted in a significant medial loss, with the pathology much more pronounced in TADs in the presence of BAPN, while neither Inact E nor Inact E+BAPN caused discernible medial loss. Interestingly, despite evident medial loss, the overall medial mass of TADs was not significantly different from aortas exposed to Inact E or Inact E+BAPN (Figure 1D), indicating thickening of the residual media in TADs. Adventitial thickening, particularly in the area without medial loss, was evident in Act E as well as Act E+BAPN TADs (Figure 1C). Consequently, these TADs displayed much thicker adventitia than aortas exposed to Inact E or Inact E +BAPN (Figure 1D).

Inflammatory infiltrates in TADs were labeled with immunohistochemistry (IHC) assays. T-cells (CD4+ or CD8+), B-cells (CD19+), and macrophages (CD68+), were readily detectable in aortas challenged with Act E+BAPN and more frequently located in the adventitia (Figure S3A). Compared with aortas exposed to Inact E, Inact E +BAPN, or Act E, TADs caused by Act E+BAPN displayed significantly more B-cells and macrophages (Figure S3B). Neutrophils (Ly6b.2+) were not detected on any of the specimens (data not shown). Additionally, we evaluated angiogenesis with IHC assays for expression of CD31. “Floppy vessels” were frequently found in the adventitial layer of TADs and aortas challenged with Act E, while not detected in those exposed to Inact E or Inact E+BAPN (Figure S4A). Quantitatively, TADs and aortas exposed to Act E had a greater degree of angiogenesis than aortas under other conditions (Figure S4B). Collectively, these results suggest that TADs suffer chronic inflammation at the advanced stage.

Act E+BAPN causes early onset of intimomedial tears and aortic wall dissection

The pathology observed for advanced TADs induced by Act E+BAPN is distinct from TAAs created with the same dosage²⁷, but was very similar to what we documented for advanced TADs induced by chronic AngII infusion⁵. To further characterize the aortic phenotype, we evaluated the pathology occurring at the early stage of TAD formation. First, we evaluated integrity of the endothelial barrier with Evans blue staining and En face microscopy as described previously⁸. TADs were collected on d3 from WT mice exposed to Act E+BAPN or Inact E +BAPN and processed for bright field or scanning electron microscopy (SEM). For the Act E+BAPN group, fresh intramural hematoma (presence of blood within the aortic wall) was noted in 2 out of the 6 specimens. Isolated blue areas with or without intimomedial tears were found in the anterior and posterior aortic wall. Intimomedial tears generally oriented with the direction of blood flow, while differing in length (30–420 μm), depth (partial to full media), and frequency (6–11 tears/sample) among the specimens. Tears were always present in areas with fresh intramural hematomas, but most tears were not associated with this event (Figure 2A and 2B). Scanning electron microcopy confirmed intimomedial tears in TADs. Adherent leukocytes and thrombi were located on the surface of the media lacking intimal coverage (Figure 2C). Elastic fiber breaks and dissemination of the aortic wall were readily detectable in cross sections of the d3 TADs under fluorescence microscopy (Figure 2D and 2E). Pathologies described above were not observed in aortas exposed to Inact E +BAPN (data not shown). These results demonstrate that Act E+BAPN causes development of TADs, instead of TAAs, in the ascending aorta.

Figure 2.

Pathology identified in TADs at d3. TADs were collected from WT mice exposed to Act E+BAPN (n=6) or Inact E+BAPN (n=6). A, En face imaging of TADs (n=3/treatment). Yellow arrows indicate branches of the aortic arch. The black arrow depicts the direction of blood flow. The white star markers an area with a fresh intramural hematoma. Regions with intimomedial tears or impaired endothelial barrier function were stained blue due to Evans blue extravasation (white arrows). B, magnified view of the boxed area in A, highlighting an intimomedial tear outlined by Evans blue staining. Blue arrows point to the edges of the intimomedial tear. C, Scanning electron microscopy of the luminal surface of TADs (n=3/treatment). Blue arrows point to the edge of the intimomedial tear while the red star marks a thrombus. The magnified view of the boxed area is shown on the right. The black arrows point to adherent white blood cells while the red star marks a thrombus containing platelets (small white bodies) and amorphous fibrin. D and E, fluorescence imaging of d3 TADs subjected to Evans blue staining in vivo. Evans blue penetrated the full thickness of the media, with the stained elastic laminae emitting red fluorescence. The open arrow in D points to an intimomedial tear. In E, a blood clot (red star) trapped in the medial layer emits bright green fluorescence. White arrows indicate the layer of external elastic lamina. F and G, Prussian blue staining of d28 TADs. Ferric iron deposits from degraded red blood cells appear blue.

Following the creation of this TAD model, we used it in multiple studies. Gross examination was performed for all TADs that were collected at various timepoints (n=41, Table 1). Aortic enlargement and peri-aortic adhesions were observed in all TADs. An interesting finding, however, is the timing of the onset of aortic dissections. Fresh hematoma was noted in 7 out of the 9 TADs collected on d7, but rarely detected at d14 or d28. Absorbed hematoma (defined as yellow coloration of the aortic wall) was absent in the first week of surgery, but grossly evident only in some TADs collected on d14 or d28. To detect resolved hematomas, we performed Prussian staining on d28 TADs. Positive staining was observed in 4 out of 8, 4 out of 7, and 2 out of 6 TADs induced by Act E+BAPN, Act E, and Inact E +BAPN, respectively, while not detected in any of the aortas exposed to Inact E (Figure 2F and 2G), suggesting that Prussian staining could not catch all ever-existing hematomas in TADs.

Table 1:

Gross pathology detected in WT TADs induced by Act E+BAPN

Pathology	d7		d14		d28
	presence	absence	presence	absence	presence	absence
Tears	9	0	9	0	not evaluated
Fresh IH	7	2	0	9	0	32
Absorbed IH	0	9	1	8	4	28

WT: wild type; IH: intramural hematoma

Act E+BAPN perpetuates inflammatory response in the aortic wall

An interesting finding on aortas challenged with various conditions is that Act E+BAPN and Inact E both stimulated inflammation in the aortic wall, but only Act E+BAPN induced TAD formation. To explore mechanisms responsible for the phenotypic disparity, we performed bulk RNAseq for aortas exposed to those conditions and compared their gene expression profiles. TADs induced by Act E+BAPN or Inact E (water was taken as the vehicle control for BAPN) were collected at d7 and d14 while ascending aortas of age-matched naïve mice were included as a baseline reference (n=6 per group). Principle component analysis showed that TADs cluster together by stimulation (Figure S5), indicating distinct gene signatures among groups subjected to different treatments. Time-dependent clustering was evident for aortas exposed to Inact E, but not for TADs (Figure S5), suggesting that the overall gene expression profile does not differ between TADs collected on d7 and d14.

To explore treatment- and time-dependent regulation of pathways and events, we performed Ingenuity Pathway Analysis (IPA) with differentially expressed genes (DEGs). The first set of analyses was carried out to evaluate early (d7) regulation of gene expression by Inact E and Act E+BAPN, and then we analyzed temporal changes of gene expression from d7 to d14. With an FDR of “10” and a significance level of 0.05, 9,255 and 11,106 genes were differentially expressed in aortas exposed to Inact E (Figure 3A) and Act E+BAPN (Figure 3C), respectively, at d7, compared with naïve aortas. IPA predicted prominent activation of “pathogen induced cytokine storm signaling pathway” in aortas treated with Inact E (Figure 3B) or Act E+BAPN (Figure 3D). Other canonical pathways related to activation of monocytes and Th1 T-cells were also activated in the aortic wall under either condition (Figure 3B and 3D). Potent mediators common to acute inflammatory response, such as IFNG, TNF, and IL1B, were predicted to be significant upstream regulators for the differential gene expression provoked by Inact E or Act E+BAPN. In support of the predictions, IFNG, TNF, and IL1B were significantly upregulated by Inact E (Figure 3A, red dots) as well as Act E+BAPN (Figure 3D, red dots).

Figure 3.

TADs failed to resolve the acute inflammatory response triggered by Act E+BAPN. A-D, genes (A and C) and pathways (B and D) induced by Inact E (A and B) and Act E+BAPN (C and D) at d7. Naïve ascending aortas served as the baseline reference. n=6/group. Volcano plots illustrate differentially expressed genes (DEGs) induced by Inact E (A) and Act E+BAPN (C). Upstream regulators and downstream pathways contributing to DEGs were identified using IPA and summarized in B and D. Notably, downstream pathways activated by Inact E or Act E+BAPN share commonalities, indicating the prominence of a similar acute inflammatory response under those conditions. E-H, temporal changes in gene expression (E and G) and downstream pathways (F and H) in aortas exposed to Inact E (E and F) or Act E+BAPN (G and H). In aortas exposed to Inact E only, a large number of genes showed significant downregulation from d7 to d14 (E), leading to the suppression of all inflammatory pathways that had been activated by d7 (F). In contrast, in Act E+BAPN TADs, only a few genes displayed notable changes in expression over the same period (G). These changes are largely associated with the inhibition of SRF and TP53, which are key regulators of SMC physiology. Importantly, no significant changes in inflammatory pathways were observed. Dashed red lines: significance threshold (P-value); Brown: pathway activation; Blue: pathway inhibition; Red dot: actual expression levels of genes predicted to be significant upstream regulators by IPA.

Next, we investigated the temporal (i.e. d7 vs. d14) changes of gene expression in aortas challenged by Inact E or Act E+BAPN. 4,828 genes displayed significant time-dependent expression in aortas exposed to Inact E (Figure 3E). Among these genes, 2,714 genes were significantly downregulated (Figure 3E). In contrast, only 137 genes displayed significant temporal expression in aortas treated with Act E+BAPN (Figure 3G). IPA predicted an overwhelming inhibition of the inflammatory pathways activated at d7 for aortas exposed to Inact E (Figure 3F), while none of the inflammatory pathways activated at d7 in Act E+BAPN treated TADs was quenched by d14 (Figure 3H). SRF and TP53 were predicted to be inhibited (Figure 3H). However, these regulators are critical for SMCs to survive and maintain their contractile phenotype³⁰. These results suggest that resolution of the acute inflammation might have contributed to the distinct phenotype observed for the two groups at d28 (Figure 1).

Both innate and adaptive immune responses are activated and intensified by Act E+BAPN in TADs

To determine specific molecular pathways responsible for TAD development induced by Act E+BAPN, we compared the gene expression profile and analyzed DEGs between Act E+BAPN and Inact E aortas. At d7, 1,457 DEGs were identified, with 801 and 656 genes being up- and down-regulated, respectively, in Act E+BAPN group compared with Inact E group (Figure 4A). IPA predicted activation of a set of upstream regulators to account for the DEGs (Figure 4B). Most of them were expressed at a significantly higher level in Act E+BAPN TADs than in Inact E aortas (Figure 4A, red dots). It is worth noting that IL17A and IL23A, both critical players of IL17 signaling³¹, were predicted to be significant upstream regulators. TNF and IL1B, although both were upregulated in aortas exposed to Inact E (Figure 3B) or in Act E+BAPN (Figure 3D), were upregulated to a significantly higher level in TADs compared with aortas treated with Inact E (Figure 4A, red dots). With this gene signature, IPA predicted activation of pathways specialized in “recruitment of myeloid cells”, including mononuclear cells, neutrophils, and phagocytes, in TADs (Figure 4B). By d14, 6390 DEGs were identified (Figure 4C), due to the resolution and the substantiation of the acute inflammation in Inact E and Act E+BAPN aortas, respectively (Figure 4C). Regulators of adaptive immunity were upregulated (Figure 4C, red dots) and predicted to promote Th1 (IFNG and IL12A) and Th2 (TSLP) responses (Figure 4D). Meanwhile, macrophages were regulated to differentiate and perform phagocytosis (Figure 4D), events commonly occur during chronic inflammation.

Figure 4.

Act E+BAPN triggered a progressive immune response and enhanced pyroptosis signaling pathways in TADs. Comparisons were made between Act E+BAPN and Inact E treatments, both of which triggered acute inflammation but led to distinct aortic phenotype. A-D, differential gene expression and pathway activation at d7 (A and B) and d14 (C and D). Expression levels of the predicted upstream regulators are highlighted in red in the volcano plots. Red lines: significance threshold (P-value); Brown and blue: upstream regulators and pathways predicted to be activated and inhibited in Act E+BAPN TADs, respectively. Note the activation of neutrophil-driven acute inflammation at the early stage (B) and T cell-mediated adaptive immune response at the more advanced stage (D). E, A comparison of the relative pathway activity between d7 and d14. The heatmap covers only top pathways with a Z score ≥ 4.7. “Pyroptosis signaling pathway” emerged as a highly ranked cellular event (lower red dashed box).

Next, we analyzed enrichment of DEGs in KEGG pathways using DAVID functional annotation analysis. Because the maximum number of genes is limited to 3,000, we only included DEGs with a fold change ≥ 2 and P ≤ 0.005. The enriched KEGG pathways were exported and plotted as dot plots (Figure 5A and 5B). Many of the enriched pathways were identified in IPA analysis (Figure 4E). However, additional signaling networks were uncovered. Among them, those related to stress response, cell growth, and cell differentiation, such as cAMP, cGMP-PKG, PI3K-AKT, MAPK, WNT, and regulation of actin cytoskeleton, were enriched on d7 (Figure 5A) and d14 (Figure 5B), indicating contribution of inflammation-independent mechanisms to TAD formation. This finding is particularly relevant to TAD formation since inflammation is not always present in human TADs³². Nevertheless, critical signaling for activation and differentiation of T- and B-cells, such as T- and B-cell receptor signaling, was not enriched at d7, but was at d14, suggesting activation of adaptive immunity in TADs. This finding is consistent with a previous study that CD4⁺ T-cells of smokers with TADs are autoimmune to elastin fragments³³.

Figure 5.

KEGG pathway analysis identified key signaling networks in developing TADs. DEGs from d7 and d14 TADs were analyzed using DAVID to discern KEGG signaling networks associated with the activation of canonical pathways predicted by IPA. A, KEGG pathways enriched in d7 TADs. In addition to pathways typically associated proinflammatory responses, signaling networks related to stress response, cell growth, and differentiation, such as cGMP-PKG, cAMP, PI3K-AKT, MAPK, and WNT, were significantly enriched. B, KEGG pathways enriched in d14 TADs. While confirming activation of T-cell mediated adaptive immunity as predicted by IPA, the analysis also demonstrated enrichment of the same set of “non-inflammatory” pathways identified at d7, indicating persistent activation of these pathways in TADs over time.

Finally, we performed IPA to assess the relative activity of the pathways in TADs between d7 and d14. A total of 288 canonical pathways had a significant Z score (i.e. Z ≥ 2.0), of which only 8 pathways were inhibited (Supplementary Table 1). Among the activated pathways, most are related to activation and differentiation of innate and adaptive immune cells. Many of the top-ranked pathways, such as neutrophil degranulation³⁴ and S100³⁵, have been evaluated for their role in TAD formation. “Pyroptosis Signaling Pathway” was ranked among the top 10 pathways specializing in regulating specific cellular events. Its activation was substantiated from d7 to d14 (Figure 4E). Given the proinflammatory nature of pyroptosis, this pathway might have played an important role in TAD development induced by Act E+BAPN. Several studies have evaluated its role in AAA development^15,16,36. However, its effect on TAD formation had not been assessed previously. Therefore, we addressed this issue with our novel TAD model.

Cell death is an active event in developing TADs

TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) assays were performed to label dying cells in aortas exposed to different conditions for four weeks. TUNEL positive cells (TUNEL⁺) were located and distributed across all layers of the aortic wall in aortas exposed to Act E+BAPN, Act E, or Inact E +BAPN. While there appeared to be more TUNEL⁺ cells in the adventitial layer, a spatial pattern could not be established for any of the conditions due to the small number of TUNEL⁺ cells detected in TADs (Figure 6A). However, a significantly greater number of TUNEL⁺ cells were detected in aortas challenged with Act E+BAPN than those subjected to Act E, or Inact E+BAPN, while hardly any were noted in aortas treated with Inact E (Figure 6B). This differential degree of cell death correlated with the histological finding that aortas exposed to Inact E only were free of intimomedial tears (Figure 1C). Immunofluorescence double labeling assays revealed that a few TUNEL⁺ cells located in the medial layer of TADs were also positive for α-actin (Figure S6), indicating occurrence of SMC death during TAD development.

Figure 6.

Act E+BAPN caused greater cell death than other conditions in d28 TADs. A, representative images of TUNEL staining in TADs induced by the indicated treatments. Magnified views of the boxed areas in the first row are shown in the second row. Blue: DAPI nuclear counterstain; red: positive TUNEL staining; purple: TUNEL⁺ nuclei. L: lumen; M: media; Adv: adventitia. White dashed lines indicate internal and external elastic laminae. B, quantification of TUNEL⁺ nuclei (n=6–7). Data were analyzed using two-way ANOVA.

Cell death mediated by GSDMD contributes to TAD development

Pyroptosis is executed by multiple members of the gasdermin family¹³. We focused our study on GSDMD as suggested by studies of AAA development^16,37. TADs were created in WT and Gsdmd null (Gsdmd^−/−) mice and followed for four-weeks. During the follow-up, Gsdmd^−/− TADs exhibited a significantly slower rate of dilation than WT TADs, while TADs of either genotype displayed a significant progressive enlargement over time. By d28, Gsdmd^−/− and WT TADs increased their lumen diameter by 31% and 68%, respectively (Figure 7A). Consistently, an overall improvement in structural integrity was observed for Gsdmd^−/− TADs compared with WT TADs (Figure 7B). Morphometric analysis revealed significantly less medial loss and adventitial fibrosis for Gsdmd^−/− TADs compared with the WT controls (Figure 7C and 7D). Consistent with the attenuation of the aortopathy, fewer TUNEL⁺ cells (Figure 7E), CD4⁺ (Figure 7F), and CD8⁺ (Figure 7G) T-cells were detected in Gsdmd^−/− than in WT TADs, whereas differences of the number of CD19⁺ B-cells (Figure 7H) and CD68⁺ macrophages (Figure 7I) were insignificant between WT and Gsdmd−/− TADs. It was noted that the composition of major immune cell subsets in the blood is similar between Gsdmd^−/− and WT mice at the baseline level (Figure S7). These results suggest that GSDMD-mediated cell death is a significant driver of TAD development and the protective effect of GSDMD deficiency likely results from modulation of a T-cell mediated immune response. This conclusion is also supported by the IPA prediction that Th1 and Th2 pathways are activated in TADs (Figure 4D).

Figure 7.

Genetic deficiency of GSDMD attenuated TAD formation. A: aortic dilation of TADs created with WT (C57) and Gsdmd null (Gsdmd^−/−) mice (male, 10 to 14 weeks, n=15 per group). Data were expressed as the percentage increase in lumen diameter from baseline (pre-TAD induction) and analyzed using two-Way Repeated measurement ANOVA. *P=0.028, Gsdmd^−/− vs. C57 at d28. B, representative histology of TADs with the indicated genotype. C-D, morphometric evaluation with the indicated parameters. E, semi-quantification of TUNEL⁺ cells. F-I, semi quantification of immunohistochemistry assays for cells with the indicated markers. For F-I, n=8–9 per genotype. Data were analyzed using unpaired t-tests.

To evaluate cleavage of GSDMDs in developing TADs, we performed western blotting assays using TADs collected at d14 (n=4 per genotype), a timepoint when differential dilation was more pronounced between Gsdmd^−/− and WT TADs. Initial assays using antibodies from published studies either produced bands corresponding to full length GSDMD in Gsdmd^−/− TADs, or only detected full length GSDMDs in WT TADs (data not shown). We then tested several antibodies from different vendors. To validate our WB protocol, THP-1 cells treated with Lipopolysaccharides (LPS) and nigericin to induce pyroptosis were used as a positive control, while Gsdmd^−/− TADs served as a negative control. Mouse AAAs induced by angiotensin II, which have been demonstrated to produce intense WB bands specific for cleaved GSDMD were also included in the validation assays. Using these controls, we identified an antibody (Cell Signaling, 39754) that demonstrated high specificity for mouse GSDMD (Figure S8A). Consistent with the vendor’s claims, we confirmed that this antibody also recognizes both full length and cleaved human GSDMDs (Figure S8B). However, despite using the validated antibody and WB protocol, we were unable to detect cleaved GSDMDs in mouse AAAs or TADs (Figure S8C).

Inflammatory cell death programs are activated in human TADs

Previous studies from other laboratories demonstrated activation of NLRP3 inflammasome and caspase-1 in human TADs³⁸. GSDMD, a substrate of Caspase-1, undergoes cleavage to release pore-forming N-terminal fragments, causing cell death¹³. Using an antibody specifically targeting the N-terminal fragment (36425, Cell Signaling), cleaved GSDMDs were detected in the medial layer of human TADs (Figure 8A). All TADs (n=9) displayed positive signals, though the extent and distribution of positivity varied significantly among samples. In contrast, control aortas(n=8) consistently showed no staining (Figure 8B), indicating that GSDMD-mediated cell death programs are activated in human TADs. To quantify full length and cleaved GSDMD, WB assays were performed using antibodies that recognize both full length and cleaved forms (39754, Cell Signaling) or only the cleaved form (36425, Cell Signaling). While comparable levels of full-length GSDMD were observed in TADs and control aortas, cleaved GSDMD was not detected in aortic samples of either group (Figure 8C). However, cleaved GSDMD was readily detected in THP-1 cells undergoing pyroptosis (Figure 8D). These findings suggest that the level of cleaved GSDMD in TADs is below the detection threshold of standard WB assays using total protein extractions from the tissue.

Figure 8.

Cleaved GSDMDs were detected by IHC but not WB assays in human TADs. A, IHC assays using a specific rabbit antibody specific for N-terminal fragments of cleaved GSDMDs. Brown: positive staining; Blue: nuclear counterstain; B, negative control with non-immunized rabbit IgGs. C, two-color WB assays for GSDMD and GAPDH expression in TADs and healthy control (Ctrl) aortas. Bands corresponding to cleaved GSDMD (31 kDa, indicated with the dashed line box) were not detected. Data were analyzed using unpaired t-test. D, single-color WB assays for GSDMD expression in the indicated samples. THP-1 cells undergoing pyroptosis were included as positive control. Cleaved GSDMD was readily detected in pyroptotic THP-1 cells but not in TADs (boxed area), suggesting that the level of cleaved GSDMD in TADs is below the detection threshold of standard WB assays.

Discussion

In the present study, we created a mouse model able to recapitulate several critical pathologies of human TADs. This model features an early (< 1 week) onset of intimomedial tears and dissection, followed by medial loss and adventitial fibrosis. Mechanistically, myeloid cells, particularly neutrophils, are likely the driver of the early pathology, while medial loss and progressive dilation during the late stage is likely to be perpetuated by chronic inflammation involving activation of both innate and adaptive immune responses. Upstream regulators, such as cAMP, cGMP-PKG, PI3K-AKT, and GSDMD, which influence cell growth, differentiation, and death, are implicated in these processes. Supporting these findings, we demonstrated that GSDMD deficiency attenuated medial loss, adventitial fibrosis, and dilation of TADs. This attenuation is associated with reduced cell death and T-cell accumulation.

Due to the undefined etiology of TADs, artificial challenges remain the only option for modeling TAD pathogenesis. Topical application of elastase in the descending aorta of mice induces aortic aneurysms when BAPN is not included in the induction^39,40. A similar aortic phenotype can also be induced with topical application of calcium chloride⁴¹. However, these models only recapitulate pathologies of TAAs and do not develop phenotypic traits typical for aortic dissections. The widely used AngII model develops TADs⁴² that may rupture at a rate greater than 50%⁴³. One of the major limitations of this model is that it develops multiple isolated dissecting aneurysms, with locations varying from the ascending to the abdominal aorta. Several studies^38,44,45 including those from our laboratory^5,46 showed that TADs and AAAs develop in the same animal and rupture at a similar rate, which differs from human TADs that affect only the ascending (Type A) or spare the ascending (Type B) aorta². TADs may be induced with BAPN. However, this model works only with immature animals^34,47,48, which differs from the epidemiology of human TADs that affect, most frequently, the aging subpopulation². Other models, such as SMC specific deletion of Tgfbr2⁷ or Tgfbr1^8,9,18, induces TADs, but are constrained by time-intensive breeding processes and off-target effects inherent to genetic modifications^10,49. In contrast, the TAD model created in the present study addresses these limitations. It develops TADs at complete penetration. Additionally, it eliminates complications and mortality associated with the open-chest surgeries reported in other studies⁵⁰.

Quantification of the severity of TADs remains a challenge for studies using mouse models. For the AngII model, rate of rupture, incidence of dissection, aortic dilation, and complication of TADs have been taken into the equation^38,42,51. This set of parameters represents the “gold-standard” for estimation of severity of TADs in the AngII model⁴⁴. When using this model, we noticed that one of the hall-mark pathologies of the TADs is the partial or complete loss of the medial layer⁵. Although medial loss is not a typical pathology of human TADs, its significance for evaluating medial degeneration cannot be overlooked. Significant medial loss in experimental TADs have also been documented others^7,52. In previous studies, we proposed a five-scale scoring system incorporating medial loss as an additional parameter to evaluate severity of TADs induced in other models^5,8. In the present study, we demonstrated that medial loss is a useful parameter for quantification of TADs induced by Act E+BAPN.

To meet the challenge of evaluating TAD severity, several studies have demonstrated advantages of using sophisticated imaging to detect and quantify the onset and progression of aortic dissection. These imaging techniques offer robust tools to detect pathologies that are not easily detected in the routine histological evaluation and to quantify the progressive spreading of dissections^53,54. However, their requirement of special imaging expertise and sophisticated imaging instruments limit widespread adoption. There is a need for simple, sensitive, reproducible, and clinically meaningful systems to evaluate TAD severity in experimental settings.

Immune injury appears to be the primary driver of TAD formation in the Act E+BAPN model. The importance of inflammation in TAD formation has been widely accepted⁵⁵. However, most studies focused on general inflammatory markers such as cytokines, proteases, and macrophage accumulation. A few studies explored the effect of immune cells, such various subsets of T-cells, B-cells, and macrophages in the development of abdominal aortic aneurysms (AAAs) and proposed a hypothesis that Th2 and Th17 immunity is deleterious to AAA development^56,57. In contrast, TAAs^58–60 and TADs^33,61 are associated with enhanced Th1 responses. In the present study, we showed that general inflammation with enhanced expression of TNFα, IL1β, IL6, and IFNG does not consequently lead to TAD formation. It is a unique type of inflammatory response specializing in activation and recruitment of neutrophils that correlated with the onset of TADs. This finding is consistent with the study by Kurihara et al, showing an influx of neutrophils promoted acute dissection of the thoracic aorta³⁴.

Innate immune response triggered by DAMPs (damage associated molecular patterns) contributes to TAD formation. DAMPs are released by stressed or dying cells and serve as alarmins to trigger danger signals, activating protective mechanisms to maintain tissue homeostasis. However, danger signals can be overshot and become pathogenic when uncontrolled⁶². Previous studies from other laboratories have demonstrated that DAMPs, such as self-DNA⁵¹ and S100A12^35,63, promote in TAD formation. In the present study, we demonstrated that pyroptosis pathways and danger signals were activated at the early stage (≤ 1 week) in both TADs (i.e. challenged with Act E+BAPN) and aortas not undergoing dissection (i.e. exposed to Inact E), indicating that these events alone are not sufficient to induce TAD formation. At the advanced stage, inflammation in aortas exposed to Inact E was nearly resolved, while TADs displayed intensified pyroptosis pathways along with activation of Th1 as well as Th2 immune responses in TADs. This sustained inflammatory response correlated with a progressive medial degeneration and aortic dilation, suggesting that pyroptosis may contribute more to subsequent remodeling than the initial onset of TAD formation.

Pyroptosis is executed by GSDMs, with the cleaved N-terminal fragments forming pours in the cell membrane, leading to membrane rupture and cell death¹³. Using the TAD model created in the present study, we demonstrated that genetic deficiency of GSDMD attenuated TAD development by reducing medial loss and aortic dilation at the advanced stage. The attenuation is associated with a reduction in TUNEL⁺ cells and decreased T-cell accumulation in TADs, suggesting a role for GSDMD-mediated inflammatory cell death in modulating T-cell response and TAD remodeling. Among the dying cells in TADs, only a small portion were identified as SMCs. While SMC death might contribute minimally to the pool of DAMPs, its impact on SMC depletion could still be substantial, considering the snapshot nature of TUNEL assays and the rapid clearance of dead cells in tissue. Previous studies have shown that inhibition of GSDMD protects against AAA formation via both cleavage-dependent¹⁶ and cleavage-independent¹⁵ mechanisms. Our results demonstrated that GSDMD is cleaved in human TADs but not in healthy control aortas, suggesting a potential role for cleaved GSDMD in TAD formation.

The present study has several limitations. First, TADs induced by Act E+BAPN rupture at a very low rate in four weeks, making this model less suitable to studies focusing on aortic rupture. Although dissections occurred in >70% (7 of 9) TADs in the first week, they tend to progress into chronic medial destruction, resembling the phenotype of unruptured TADs induced by AngII. Furthermore, our model contracts with previous reports showing that Act E with or without BAPN causes formation of aortic aneurysms, rather than dissection, in the abdominal aorta^28,29 or descending thoracic aorta²⁷. This discrepancy might stem from anatomic differences of aortas in response to exogenous stimulation, a hypothesis supported by a recent study demonstrating that BAPN causes distinct pathology in ascending versus descending aortas⁴⁸.

Another limitation is the incomplete understanding of the mechanisms underlying the protective effects of Gsdmd^−/− on TAD formation. Our results suggest that a unique type of inflammatory response, particularly the influx of neutrophils, drives the onset of TAD formation. While this observation aligns with findings from studies using the AngII model³⁴, whether the same mechanisms such as neutrophil MMP9 activity contribute to early dissection in our model remains unclear. The impact of GSDMD deficiency on TAD formation is more pronounced at the advanced stage, but further studies are needed to explore the activation of GSDMD in specific cell-types, the dynamics of GSDMD-mediated cell death, and the role of immune cell subsets during TAD development. It also remains unclear whether the medial loss results from dissection (i.e. physical removal of the medial layer), chronic death of SMCs, and/or other events. Furthermore, GSDMD mediates not only several inflammatory cell death programs^13,14, but also cell death-independent functions⁶⁴. For instance, Gao et al. demonstrated that SMC-specific GSDMD promotes AAA formation by elevating endoplasmic reticulum stress rather than inducing pyroptosis¹⁵. Although our results indicate that Gsdmd^−/− attenuates cell death, further studies are required to identify specific cell death programs activated in TADs and to investigate their implications in TAD development. Moreover, it is worth noting that although the WT and Gsdmd^−/− mice used in this study were maintained on a C57BL/6J background, they were not littermates. To confirm the protective effects of Gsdmd^−/−, future studies should be conducted under more stringent genetic conditions.

Highlights.

• Topical elastase application to the ascending aorta, combined with systemic β-aminopropionitrile (BAPN) administration, induces intimomedial tears and aortic dissection. This sharply contrasts with the abdominal aorta, where the same treatment results in aortic aneurysm formation.

• A minimally invasive procedure has been developed to precisely deliver elastase to the outer surface of the ascending aorta without requiring thoracotomy. Using this approach, a novel mouse model of ascending aortic dissection was established.

• The onset of TADs in this model is driven by a unique inflammatory mechanism, with inflammatory cell death as a key upstream regulator. This was further confirmed by experiments demonstrating that genetic deficiency of gasdermin D protects against TAD formation.

Nonstandard Abbreviations and Acronyms

TAD

Type A aortic dissection

Act E

Active elastase

BAPN

β-aminopropionitrile

GSDMD

of gasdermin D

Inact E

Heat inactivated elastase

TAA

Thoracic aortic aneurysm

DEGs

Differentially expressed genes

IPA

Ingenuity Pathway Analysis

AAA

Abdominal aortic aneurysm