Introduction

Carotid atherosclerosis and abdominal aortic aneurysm (AAA) are distinct, complex, mostly asymptomatic, chronic disorders characterized by unpredictable development of arterial lesions. Atherosclerotic carotid plaque (CPL) and AAA share numerous features (i.e., chronic inflammation, alterations of vascular wall cells and extracellular matrix, and risk of thrombosis), indicative of interrelation1.

Despite the progressive improvement in care management, the incidence of rupture in the lesioned artery leading to negative outcomes including death, remains relevant2,3,4. Compelling evidence in favour of the causal independency but synergy between CPL and AAA diseases suggest a parallel or secondary reciprocal development1,5,6. The European Society for Vascular Surgery guidelines 2024 recommend systematic screening of patients at risk for /with AAA 7 , but not that of carotid prior to AAA surgery (see Recommendation 36).

Inflammatory mediators have been quite extensively studied for their pathogenic involvement either in CPL or AAA development8,9,10,11,12,13,14,15 and the hypothesis of an association between lesion and specific inflammation profile emerged. However, comparative and predictive data about inflammation in single disease vs. co- presentation of CPL and AAA are lacking.

In addition to “classic” inflammatory mediators, namely cyto-chemokines, the P2X purinoceptor 7 (P2X7) has been associated with CPL vulnerability and presence of cells with phenotype intermediate between smooth muscle and macrophage, as well as to macrophagic activity in AAA, both in humans and mice16,17,18,19. Recently, P2X7 protein has been found into the bloodstream, transported by cell-shed microvesicles / microparticles20. The ATP -activated P2X7 contributes to the circulating inflammation in AAA through triggering the formation of NLRP3 inflammasome and the activation of caspase-1 secretome19, or the release from vascular wall cells and macrophages into the bloodstream of IL-1β TNF-α, and proteases21,22, or by indirectly affecting the production and release of cyto-chemokines important in CPL and AAA pathogenesis8,9,10,11,12,23. P2X7 may be modulated by microRNAs (miRs)24 such as miR-150 and miR-186. The former is cardioprotective against ischemic injury in mice25, the latter proposed as biomarker for atherosclerosis26 and overexpressed in ascending aortic aneurysm27.

Here we investigate in patients with CPL at different stages and with/without AAA the interrelationship between P2X7 and selected inflammatory mediators and we delineate the associated molecular pathways. We present proof-of-concept evidence of circulating inflammatory distinctive fingerprints distinguishing patients in relation to CPL and risk of incident clinical AAA. We provide insights on miR-150 and miR-186 in these settings. We also explore the specificity of the impact of serum from patients with CPL and with/without AAA over pro-inflammatory activation of endothelial and smooth muscle cells in relation to P2X7, IL1-β and CXCL1.

Results

Patient population

A total of 41 patients (≥ 10 patient/group), and 9 healthy volunteers were enrolled and investigated. No significant differences among groups in demographic, clinical and hematologic features were found by ANOVA (Table 1).

Table 1 Patient data.
Full size table

Group D differed from groups A and C for a less frequent prescription of Cardioaspirin, and from group A for more frequent use of antiaggregants. Antibiotics were assumed more frequently by patients in group C than in the others (Table 1, Supplementary Table S2).

The development of AAA was under surveillance till aneurysmectomy in 7 out of 20 patients (35%) of groups A and B, and in 2 of them the co-presence of CPL was also recorded. Analogously, 5 out of 10 patients (50%) in group C were under monitoring for known carotid atherosclerosis, and 2 of them suffered a previous cerebral ischemic event.

The groups differed for the distribution (Fig. 1A, B) and stenosis entity (Supplementary Table S2) of the CPL along the artery tree. CPL was undetected by Eco Doppler TSA at common carotid artery levels (COM) in both groups A and B, and absent at right internal carotid artery (ICA) and bulb level in group A, whereas right ICA and both bulbs were the prevalent sites with CPL in group B. Overall, groups C and D showed similar CPL localization, where ICA was bilaterally the more common site of lesion development and COM the infrequent.

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Distribution of carotid plaques and stenosis among groups Plaque localizations along carotid artery branches in patient groups (n ≥ 10/group) shown as pie chart with percentage of patients displaying CPL in the specific branch (A) and as heatmap of stenosis percentage in single patient/branch (B) are presented.

Serum P2X7 and P2X7 artery tissue contents: differences among groups

The serum P2X7 differed (Fig. 2A) in ctrl vs. groups B and D but not vs. group A. The serum content of P2X7 was higher in group A vs. groups B and D. Difference between groups B and C reached only borderline significance. Analysis of the area under the curve from receiver operating characteristic curves of P2X7 validated the results.

Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

P2X7 content in serum and artery tissue: intergroup comparisons. Differences among groups in the P2X7 serum content are plotted (A, left graph). Values are presented as boxes (min to max) and dots indicate single sample values (n total = 60). One-way ANOVA for multiple comparisons with two-stage linear step-up procedure of Benjamini, Krieger and Yekutieli is applied and significant differences shown. Receiver‐operator characteristic (ROC) curves are shown for each group. AUC > 0.7 with significant p values is considered as threshold for good discriminant performance. The P2X7 content in AAA (n = 12) vs. CPL (n = 10) lesion samples is plotted (B). Values are presented as boxes (min to max), and dots indicate single sample values. Relationships between serum and tissue P2X7 content (C), stenosis degree and serum (D) or tissue P2X7 (E) are shown.

P2X7 content did not significantly differ between aneurysmatic aorta (groups A and B, n = 6/group, and atherosclerotic carotid (group C, n = 10; p = 0,0614 for A + B vs. C, Fig. 2B), supporting a local role of the receptor in both the lesioned arteries. An inverse relationship, possibly associable to P2X7 uptake by the carotid, was found between serum and tissue amounts of P2X7 in C group but not in the others (Fig. 2C).

No relationship between P2X7 and AAA diameter was found. Conversely, the serum P2X7 was moderately negatively associated with CPL stenosis by Eco Doppler TSA in patients with critical AAA (group A + B, Fig. 2D). As well, the stenosis degree at CPL bulb was positively related to the aortic tissue content of P2X7 in group B (Fig. 2E). These data possibly accounted for a group-dependent role of P2X7 in the CPL lesion development.

Circulating inflammatory mediators

Circulating inflammatory mediator frameworks may reinforce the intergroup discriminatory role of P2X7 in the development of AAA and/or CPL. To explore this field, the concentrations of 15 cyto/chemokines (namely GM-CSF, IL-1α, IL-1β, IL-6, IL-8, IL-10, IL-12p70, IL-13, CCL2, CCL3, CCL4, CXCL1, CXCL2 and TNF-α) were measured in serum9,13,30. Significant differences among groups were found for GM-CSF, IL-1α, IL-1β, IL-10, IL-12-p70, CXCL1, CCL3 and TNF-α (Fig. 3A), while CCL2 differentiated all patient groups from ctrl (Supplementary figure S2). Despite the observed differences, in our population the levels of CCL2 and TNF-α were in the normal range31. In comparison with groups C and D, the contents of GM-CSF and IL-1β were significantly higher in serum samples from both groups A and B. Moreover, in group A vs. groups C and D the amount of IL-10 and CCL3 were higher, that of CXCL1 lower. Consistently in group B vs. group C the amount of CXCL1 was lower. Besides, in group B vs. group D the IL-1α was lower, and vs. groups C and D the IL-12p70 slightly higher. These differences delineated group-associated fingerprints, depicted by heatmap of mean serum values / patient group (Fig. 3B).

Fig. 3
Fig. 3The alternative text for this image may have been generated using AI.
Full size image

Circulating inflammation mediators: differences among groups Circulating inflammatory mediator levels in all groups are shown (A). Values are presented as boxes (min to max), and dots indicate single sample values (n = 60). Kruskal–Wallis test for multiple comparisons with two-stage linear step-up procedure of Benjamini, Krieger and Yekutieli is applied and significant differences shown. Heatmap of the mean amount of inflammatory mediators visualizes fingerprinting differences among the groups (B). The network of mediator interactions and already known colocalization in homo sapiens by STRING v12 is presented (C). Clustered protein nodes are in the same colour.

In silico analysis of mediator networking

The functional networks interconnecting the mediators associated to each fingerprint were obtained by in silico analysis using STRING v12 engine (Fig. 3C), which also identified a number of functionally enriched GO processes (Supplementary_IN SILICO ANALYSIS FROM STRING.xlsx). The GO processes with false discovery rate < 0.001 were clustered in main families by QuickGO engine, building ancestor charts (Fig. 4). Several GO processes were common to all groups (i.e., immune system processes and inflammatory response), others only to A, B and C groups (i.e., response to LPS and regulation of leukocyte chemotaxis), or to A and B groups (i.e., cellular signalling and metabolism associated to LPS, chemokines, IFNγ, IL-1β and TNF-α, ERK1/2 pathway, Calcium ions transport, and chemotaxis of circulating immune cells).

Fig. 4
Fig. 4The alternative text for this image may have been generated using AI.
Full size image

Ancestor charts for GO processes associated to patient fingerprints Representation of the functional relationship connecting common and exclusive GO processes associated to patient groups as hierarchical ancestor charts by QuickGo engine. In the more complex charts, the terminal domains are delimited and enlarged (red squares).

Moreover, some GO processes were exclusively associated to a single group. Specifically, the positive regulation of of ATP, of response to stimulus, of T lymphocytes, of Calcium signalling and transport, of IL-1β production and prostaglandins secretion, of NK cell and neutrophil chemotaxis, and of angiogenesis are associated to group A ; the regulation of endothelial barrier and of endothelial and smooth muscle cells proliferation, the negative regulation of cytokines and positive of chemokines, immunoglobulins, cell metabolism and the signalling downstream Calcium (including negative calcification) related to group B. One exclusive GO process was associated to group C (i.e., negative regulation of bone resorption) and another to group D (i.e., positive regulation of neutrophil migration). These differences are suggestive of a fine tuning of specific cellular functions in the groups.

Inflammatory mediators in the lesioned arteries and serum-tissue relationships

The circulating mediators could be reflected of affect their tissue content, thus the mediators were measured in AAA and CPL tissues and correlated with their amount into serum.

The lesioned tissues, even if belonging to different arteries, were comparable for the content of CXCL1 and IL-1α in all groups. Conversely, AAA differed from CPL for the highest content of GM-CSF (the cytokine was barely assessable into both serum and CPL from groups C, D) and CCL3, but lowest of IL-10 and TNF-α. In AAA of group B but not of group A the IL-1β content was also higher vs. CPL, hinting to more inflamed AAA in patients with co-presence of severe carotid stenosis (Fig. 5A, Supplementary figure S3). Overall, these findings highlighted the potential of fingerprints for the stratification of patient’ groups.

Fig. 5
Fig. 5The alternative text for this image may have been generated using AI.
Full size image

Inflammatory mediators in artery tissues and correlation with P2X7 content The inflammatory mediator levels in lesioned artery tissues (aorta for groups A and B, carotid for group C; n = 7, n = 7, and n = 10, respectively) are shown (A). Values are presented as boxes (min to max) and dots indicate single sample values. Kruskal–Wallis test for multiple comparisons with two-stage linear step-up procedure of Benjamini, Krieger and Yekutieli is applied and significant differences shown. Linear relations between circulating mediators and stenosis entity in the different groups (C) and relations between inflammatory mediators in tissue and serum samples are shown (D).

Negative linear relationships were found between the entity of stenosis in CPL and serum levels of IL-1β, Il-10 in group C and of CXCL1 in group B and C, compatible with a decreased secretion or increased uptake of the molecules during CPL growth (Fig. 5B). The serum amount of P2X7 was linearly related to serum IL-1β and IL-1α in group A, and to IL-1β in group C (Fig. 5C). In this latter group the content of P2X7 was associated to that of IL-1β and CCL2 into CPL tissues (Fig. 5C). Moreover, the serum P2X7 was related to AAA content of IL-1β and IL-10 in group B. Further weak correlations between serum and tissue were found by Spearman R analysis (Supplementary Table S3), depicting the complex interplay between blood and arterial tissue.

Expression of P2X7 gene, miR-150 and miR-186

The P2X7 functional modulation involves miR-150 and miR-186. P2X7 gene was expressed by AAA and CPL but undetermined in serum, whilst the miRs were found in all tested samples (Fig. 6). Significantly higher serum levels of mir-150 were found both in ctrl and group A vs. group C, and no intragroup difference determined for miR-186 (Fig. 6A). The two circulating miRs appeared linearly interrelated in groups B and C (Fig. 6B). Additionally, circulating miR-186 was negatively associated to the AAA tissue content of GM-CSF in group B (Fig. 6C).

Fig. 6
Fig. 6The alternative text for this image may have been generated using AI.
Full size image

Expression of P2X7 gene and miRs in serum and vascular lesions Expression levels in serum samples of miR-150-5p and miR-186-5p by RT‐qPCR is plotted (A). Values are presented as boxes (min to max) and dots indicate single sample values (n = 60). Linear relations between circulating miR-150-5p and miR-186-5p are shown (B). Nonlinear regression curve fit showing the correlation between tissue GM-CSF and serum miR-186-5p in group B is plotted (C). Tissue levels of P2X7 gene (D, left graph), miR-150 and miR-186 (D), right graph) by RT‐qPCR in AAA (groups A and B, n = 7/group for P2X7 and n = 5 for miRs) vs. CPL (group C, n = 11) are displayed. Values in D panels are presented as boxes (min to max) and dots indicate single sample values. Kruskal–Wallis test for multiple comparisons with two-stage linear step-up procedure of Benjamini, Krieger and Yekutieli is applied in panels A and D left; Mann–Whitney and t-test are applied in panel D right, and significant differences shown.

The P2X7 gene and miR-150 were more expressed in AAA vs. CPL tissues while miR-186 was comparable between lesion types and significantly less expressed than miR-150 in AAA (Fig. 6D). The expression level of miR-150 was significantly higher in AAA tissues vs. serum from patients of groups A and B (p = 0,0151). No further relationships were detected.

Overall, these data on P2X7 gene and target miRs further supported difference among groups and lesion type.

Responsivity of aortic vascular cells conditioned with patient’ serum

AAA and CPL lesions are structurally complex and develop upon multifactorial stimulation, leading in the early stages to the alteration of luminal endothelium homeostasis. The impact of fingerprints over early responses by in vitro primary human aortic endothelial and smooth muscle cells (EC and SMC, respectively) was evaluated. Cell conditioned with human serum from patients of groups A, B , C and not-conditioned cells were compared. The serum samples used for cell conditioning were group-representative for the content of GM-CSF, P2X7, IL-1β and CXCL1 (Fig. 7A, Supplementary figure S3). In EC (Fig. 7B left plot) the serum from group C induced downregulation of P2X7, IL1B, CXCL1 genes with respect to conditioning with serum from groups A and B, and to no conditioning (p = 0,0312 for all genes). Conversely, no significant changes were found in SMC (Fig. 7B right plot), except for a trend vs. IL1B downregulation following serum from group C (p = 0.0622). Worth mentioning that the serum samples from the same 4 patients of group B are responsible of the downregulation of P2X7 and CXCL1 genes in conditioned EC and of the upregulation of P2X7 in conditioned SMC vs. not conditioned cells.

Fig. 7
Fig. 7The alternative text for this image may have been generated using AI.
Full size image

In vitro conditioning of primary aortic endothelial cells: difference in cell response by serum samples from different groups Quantification of inflammatory mediators in serum (n = 6/group) used for cells conditioning are plotted (A). The P2X7, IL1B and CXCL1 gene expression levels by RT‐qPCR in conditioned EC and SMC are plotted (B). Relations between P2X7 gene and CXCL1 in EC or IL1B in SMC (C), and serum content of P2X7 or IL1β in EC (D) are shown. Linear and nonlinear regression curve fits between circulating mediators and gene expression levels in SMC are shows (E). Representative western blot images (horizontal white lines indicate the gel crop) and results from densitometric analysis of the P2X7 isoforms in EC and SMC conditioned with patient serum are shown (F). Expression levels of miR-186-5p in EC and SMC by RT‐qPCR is shown (G). Values in (A), (B), (F) and (G) are presented as boxes (min to max), and dots indicate values from single biological replicates (3 ≤ n ≤ 6). Kruskal–Wallis test for multiple comparisons with two-stage linear step-up procedure of Benjamini, Krieger and Yekutieli is applied in A; Mann–Whitney and unpaired t-test are applied in Band F, and significant differences shown.

P2X7 gene resulted linearly related to expressed CXCL1 in EC, to IL1B in SMC, upon cell-conditioning with serum from group B (Fig. 7C). Moreover, P2X7 gene resulted positively related with serum P2X7 and IL-1β in EC conditioned with serum from group C, and negatively with IL-1β in those conditioned with serum from group A and B (Fig. 7D). CXCL1 gene in conditioned SMC was negatively related with the content of IL-10 and IL-13 in serum from group B, while IL1B expression was associated to the amount of CXCL1, IL-1α IL-12-p70 in serum from group A (Fig. 7E). No relationships were found for SMC conditioned with serum from group C.

In EC, the expression of both 75 kDa and 50 kDa P2X7 isoforms by Western blot was increased by conditioning with serum from groups A and B vs. that from group C (Figure Supplementary figure S4). No difference was found in SMC. These data provided evidence that pro-inflammatory genes including P2X7 may be directly activated by serum from groups A and B (not group C) in human aortic EC but not SMC.

Expression of miR-150 and miR-186 in serum-conditioned vascular cells

The miR-186, putatively involved the aneurysmatic transformation of the aortic wall27, was comparably expressed in both conditioned EC and SMC, without differences intragroup nor vs. no conditioning (Fig. 7G). The intracellular expression of miR-186 was not related to P2X7, IL1B and CXCL1 genes in conditioned cells. The miR-150 was undetermined in both cell types and not induced by conditioning. Overall, a direct early role of serum mediators over miR-150 and miR-186 in aortic cells was not assessed.

Discussion

The co-presentation of carotid atherosclerosis and aneurysmatic lesion of the aorta is frequent, and presence of CPL is a risk factor for AAA development1. The routinary screening programs do not include the imaging-based contemporary evaluation of carotid arteries and aorta, and current guidelines do not recommend carotid screening in patients with AAA7. Consequently, the identification of patients co-presenting CPL and AAA remains an open challenge. Our patients presented CPL, and imaging showed that sites of plaque formation along the carotid tree differed depending on the co-presence of AAA, i.e. CPL were found in internal and common branches in patients with small/without AAA but detected in carotid bulbs in subjects with severe AAA. These data lay along with a study associating the AAA formation to both decreased wall shear stress and different geography of carotid alterations32. The literature reported association between AAA presence and altered carotid mechanical properties33, risk of AAA and carotid dimensions34, increased incidence of aneurysm and arteriomegaly36. The recent ARIC study confirmed the association between risk of clinically detected incident AAA and increase of both intima-media thickness and stiffness in CPL1. Indeed no firm conclusion emerged about the usefulness of a systematically screening for AAA the patients with CPL, and the hypothesis of two processes running in parallel was formulated, suggesting that atherosclerosis is not mandatory for AAA formation despite the common risk factors or genetic predisposition1.

The circulating and arterial inflammatory backgrounds13 are other recognized causal players in the development of CPL and AAA37,38,39,40,41,42,43, poorly explored in patients with CPL at risk of developing AAA.

Our primary focus was on P2X7, a purinergic receptor molecule potentially exploitable for diagnostic and therapeutic purposes in various diseases, included cardiovascular16,44,45. P2X7 was functionally associated to worsening of the inflammatory status through activation of NLRP3 inflammasome in both CPL and AAA19,43,46. Specifically P2X7 was involved in destabilization and increased thrombogenicity of CPL17,46,47, in modulation of macrophage pyroptosis in AAA19. In our study, serum P2X7 is associated with CPL stenosis in patients with large AAA, subtending a role in CPL growth in co-presenting patients. Conversely, in the absence of AAA, the correlation between serum and tissue P2X7 suggests a direct functional role in the interplay blood-artery, contributing to the development of hemodinamically significant CPL.

Low levels of serum P2X7 characterize both patients with the less severe (group D) and those with the worst prognosis (i.e., group B). Since serum P2X7 increased during acute hyperinflammatory process48, a poor response in the early stages of lesion and a long-term adaptation to chronic conditions may explain the apparent paradox. Overall, these data pave the way to largest clinical studies of targeting P2X7 in patients developing CPL.

The circulating inflammatory mediators delineate fingerprints distinctive of the patient groups. Briefly, serum GM-CSF, TNF-α and IL-1β differentiate the patients with co-presence of CPL and severe AAA (groups A, B) from those with absent/small AAA (groups C, D); while IL-10 and CCL3 distinguish between subclinical vs. critical CPL in subjects with severe AAA (group A vs. B). Notably, IL-10 may indicate the success of a systemic carotid -protective attempt in the presence of severe AAA, in agreement with the atheroprotective role of IL-1010,14,15.

In tissues, high levels of P2X7 (gene and protein), CCL3, IL-1β, GM-CSF and miR-150 but low of IL-10 and TNF-α differentiate AAA from CPL (i.e., groups A, B from group C). We show that GM-CSF, a main macrophage regulator49 driving pro-inflammatory or regulatory responses by its targets in microenvironment/dose -dependent manner and involved in AAA development in mice9,13,30, is lacking in patients with small/without AAA but abundant in those with severe AAA, and negatively related to miR-186. The serum miR-186 was proposed as biomarker promoting SMC proliferation and migration in patients with non-critical CPL 24. No confirmation derives from our study and the observed relationships among miR-186, miR-150 and GM-CSF might suggest functional interplay in group B but not in group C.

P2X7 was detected in lesion infiltrating macrophages and macrophage-like cells18,19, widely distributed in the AAA wall but confined in the CPL shoulder50,51,52. Circulating TNF-α and GM-CSF induced P2X7 at immune and endothelial cell level and were modulated by IL-1β18,53,54. In turn, the activated P2X7 triggered metalloprotease activation, vascular cell death or secretion of pro-inflammatory mediators including IL-1β16, depicting a functional loop. In patients with CPL without AAA (group C) lacking of GM-CSF, the linear relationships between P2X7 and IL-1β in serum and tissues, hints to a GM-CSF-independent, synergistic involvement of P2X7 and IL-1β in haemodynamically significant CPL.

Moreover, the serum CXCL1 discriminates group C vs. groups A, B, being globally highest in the absence of severe AAA, and is negatively associated to CPL stenosis degree in specific branches in the presence/absence of severe AAA (groups B and C). In addition, tissue CXCL1 is positively related with serum P2X7 in group B. CXCL1 production by endothelial cells, activated macrophages and neutrophils was reported in CPL and AAA55,56. Consistently with our results, a functional relationship between P2X7 suppression and production of CXCL1 was demonstrated on in vitro endothelial progenitor cells57. Taken together these data hint to regulation of CXCL1 by P2X7 and to functional relation between CXCL1 and CPL (but not AAA) development, deserving further validation.

In silico analysis supports the fine-tuned regulation played by inflammatory fingerprints: although general activation of the immune system is common to all groups, the exclusivity of negative regulation of bone resorption by group C is in agreement with the frequent calcification of advanced CPL, that of GO processes related to T lymphocyte activities in group A with a proatherogenic immune activation, and that of regulation of cellular metabolic pathways and vascular cells proliferation with the widest vascular alterations in group B.

Functional test of the specificity of serum fingerprints in activating cells demonstrates the capability by serum from groups A and B, but incapability by serum from group C to elicit in vitro early pro-inflammatory responses by aortic EC. These findings are in agreement with the absence of AAA in patients of group C and the artery-specificity of the fingerprint action, thus with mechanistic difference in CPL and AAA development. Conversely, serum fingerprints were unable to induce remodelling of SMC, consistently with putative need of cross talk with EC or other vascular wall cells to activate SMC, indirectly exposed to blood in vivo58,59,60.

In addition to pathogenic implications, the patient fingerprinting may represent a diagnostic tool and provide candidates for personalized, targeted approaches. This strategy is endorsed by availability of P2X7 antagonists currently under clinical trials for neuroinflammatory disorders and of drugs modulating GM-CSF, IL-10, IL-1β and TNF-α already applied to treat human diseases16,61,62,63,−64, which could be investigated for repurposing in our settings.

Conclusions

This study demonstrates the association between CPL localization and degree of stenosis, co-presence of CPL and of AAA, and circulating levels of P2X7, GM-CSF, CXCL1, IL-1β, IL-10, IL-12p70, CCL 3, TNF-α and miR-150. Different fingerprints of circulating mediators are associable to patients with CPL and/or AAA at various stages (Supplementary figure S5), possibly serving as stratification tools for improving early diagnosis and preventing complications, allowing reduction of the reliance on invasive or costly imaging methods.

The central role of several fingerprint molecules, namely P2X7, GM-CSF, IL-1β and CXCL1, in modulating arterial responses and lesion progression makes them promising therapeutic candidates for tailored pharmacological treatments.

In silico analyses and in vitro endothelial cells responsivity support different, fingerprint -related molecular mechanism for carotid and aorta alteration, deserving dedicated studies.

Limitations

The present work has some limitations. First at all, tissue samples were available only from several enrolled patients undergoing surgery and not from those at early stages of disease and of healthy volunteers. Tissue samples from both aorta and carotid arteries of the same patient were unavailable, limiting comparisons.

Methods

Patient population

Consecutive male patients (age ≥ 60 years) evaluated with standard medical examination following current guidelines 8 at Division of Vascular Surgery from January 2022 to December 2023, undergoing abdominal and supra-aortic-trunk imaging by Eco-Colour Doppler, angiographic computed tomography or thoracic RX, and either scheduled for aneurysmectomy on unruptured large AAA or for carotid endarterectomy were enrolled.

Based on the presence and dimension of single or double artery lesions, the patients were allocated to different groups:

  1. 1)

    Group A submitted to aneurysmectomy for severe AAA (maximum diameter > 5 cm) and with CPL stenosis < 40%,

  2. 2)

    Group B submitted to aneurysmectomy for severe AAA (maximum diameter > 5 cm) and with CPL stenosis ≥ 40%,

  3. 3)

    Group C without AAA but with CPL scheduled for surgery because of stenosis > 40%,

  4. 4)

    Group D with small/without AAA and with CPL stenosis below critical percentage of stenosis following NASCET criteria.

Exclusion criteria were type I diabetes mellitus, ongoing oncological or autoimmune diseases, recent (within six months) stroke or transient ischemic attack episodes, infectious diseases. Demographic, clinical, imaging and therapy data were collected and organized in anonymized database.

Volunteers matched for age and providing auto-certification of their healthiness served as control group (ctrl).

The population protocol is schematized in flowchart (Supplementary Fig. 1).

The Study Protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the internal review Board of the Ethic Committee at San Raffaele Hospital. All subjects signed an informed consent to participate to the study.

Blood and tissue processing

Blood samples (10 mL) from all patients and ctrl were collected in vacutainer serum tubes and processed to serum, aliquoted and stored at -80 °C till use, avoiding freeze–thaw.

AAA (patients from groups A, B) and CPL tissues harvested during surgery (patients from group C) in sterile vials containing physiological solution with 100 I.U. Penicillin/Streptomycin, were processed within 30 min from the excision. Samples were cut in macroscopically comparable pieces (5 mm x 1 cm side) and snap frozen or immersed in RNA later solution for protein or gene/miR analysis, respectively.

Cell culture and treatment protocol

Primary human aortic endothelial and smooth muscle cells (HAOEC cat# 304-05 s and HAOSMC cat# 354-05a, Cell Applications Inc., San Diego, CA) were cultured in 12 wells plates using specific cell culture medium (cat. # 095211F-500, Human EC Growth Medium and cat. # 095,311–500 Human SMC Growth Medium, Cell Applications Inc., San Diego, CA) supplemented with 2% foetal bovine serum (FBS qualified, USA origin, ThermoFisher, Waltham, MS), following manufacturer’s instructions. At 50% confluence the HAOEC and HAOSMC were conditioned by adding 2% human serum from patients to culture medium and cultured to reach 80–95% confluence, then 5% human serum was added for the last 72 h of culture.

Tissue and cell proteins extraction

Samples from a subset AAA, CPL, HAOEC and HAOSMC were freshly extracted in RIPA buffer supplemented with complete protease/phosphatase inhibitor cocktail (1:100 dilution, Sigma-Aldrich, St. Louis, MO). The protein concentration was measured by Bradford (tissue samples), or BCA (cell samples) protein assay kits (Quick Start™ Bradford Protein Assay, Bio-RAD, Milano, Italy; QPRO-BCA Kit Standard, Cyanagen, Bologna, Italy).

Luminex profiling of cytokines and P2X7 content

Levels of 14 pro-anti-inflammatory mediators (i.e., GM-CSF, CXCL1, CXCL2, IL-1α, IL-1β, IL-10, IL-12p70, IL-13, IL-6, IL-8, CCL2, CCL3, CCL4 and TNF-α) were determined by Luminex Discovery Assays (R&D Systems, Bio-Techne, Minneapolis, MN) in serum and tissue extracts. The content of P2X7 was detected in serum and in tissue extracts by ELISA (CSB-EL017325HU, Cusabio, Houston, TX), measured on Victor 3 V reader (Perkin Elmer, Waltham, MS). A sample repeated in every run served as internal calibrator for inter-plate variability normalization. Two technical replicates were run/sample.

Western blotting

After protein denaturation in 1% 2-mercaptoethanol-containing sample buffer, 25 μg of extract were loaded and resolved on Mini-PROTEAN Protein Gels then transferred onto 0.2-μm nitrocellulose membrane (all from Bio-RAD, Milano, Italy). Non-specific binding was blocked with 5% bovine serum albumin in 0.1% Tween20 in PBS (1 h, RT). The membranes were cropped between 37 and 50 kD, and the upper parts incubated with polyclonal antibody against P2X7 (AB5581, 1:1000 dilution, under gentle shaking overnight at 4 °C, Merck, Darmstadt, Germany), washed and submitted to anti-Rabbit polyclonal antibody-HRP conjugated (A0545, 1:10,000 dilution, 1 h, RT, Sigma-Aldrich, St. Louis, MO). Monoclonal antibody anti-GAPDH-HRP conjugated (MAB-10578-HRP, Immunological Sciences, Roma, Italy, 1:10,000 dilution) applied to the lower part of the membranes served as loading control. Clarity™ western ECL substrate (Bio-RAD, Milano, Italy) allowed visualization of protein expression on Amersham Hyperfilm ECL by Xograph Imaging System (Compact × 4). Optical densitometry was performed with UVITec (Nine Alliance, Cambridge, UK) normalizing bands against intensity of loading control band.

RNA extraction and RT-qPCR on P2X7, miRs and cytokine-encoding genes

RNA, including miR-enriched fraction was isolated from serum with miRNeasy serum/plasma miR isolation kit (Qiagen, Hilden, Germany), from cells and a subset of tissues using miRNeasy mini kit (Qiagen, Hilden, Germany), then quantified at NanoDrop™ 1000 (ThermoFisher Scientific Inc., Waltham, MA). Synthesis of cDNA was carried out either by TaqMan® Advanced miRNA Assay (serum and tissue miRs) or High-Capacity RNA-to-cDNA Kit (tissue mRNA) (all from Invitrogen, Carlsbad, CA). Quantitative real-time PCR was performed using TaqMan® Universal Master Mix II or by TaqMan® Fast Advanced Master Mix and Taqman primer/Fam-labelled probes (Applied Biosystems, Foster City, CA, USA). All the procedures were carried out according to the manufacturers’ instruction. Samples were run in triplicate and target levels were normalized to that of reference probes (Supplementary table S1). The relative expression was determined using the 2^-ΔCt method. Variation of target expressions in human serum -conditioned cells vs. untreated (FBS only) was calculated as fold changes.

In silico analysis

Circulating Inflammatory mediator-associated network was obtained for each group of patients by STRING v12 web engine 28, analysing the mediators putatively upregulated. The GO processes obtained from network analysis were compared and grouped based on their sharing or exclusivity among groups. Common and exclusive ancestor charts were separately obtained from the GO process groups using QuickGO engine 29.

Statistics

The normality of experimental and clinical data distribution was assessed by Shapiro–Wilk normality test. Correlation among clinical and experimental variables was analysed by Spearman test, simple linear regression, or nonlinear fit tests. Differences among variables were calculated by one-way ANOVA or Kruskal–Wallis test for multiple comparisons with two-stage linear step-up procedure of Benjamini, Krieger and Yekutieli, or t-test. Parametric or non -parametric tests are used depending on the normality of data distribution. Probability values < 0.05 were considered significant. Prism 9.5.0 software was used.