A total of 75 µl of human serum was diluted in 225 µl phosphate-buffered saline (PBS) and centrifuged at 17,000x g for 20 min at 4°C to isolate microparticles. After centrifugation, the supernatant was discarded and the pellet was resuspended in 300 µl PBS. The samples were then stained using ASC-B3-Alexa-647 (sc-514414, Santa Cruz) for 1 h at 37°C. Afterward, 120 µl of the samples was measured using a flow cytometer BDFacsLyric (BDscience). ASC-GFP specks derived from THP1-ASC-GFP cells were labelled with an ASC-B3-Alexa-647 antibody and added to human serum to serve as a positive control for gating the samples. These were detected in the 488-nm channel of the flow cytometer (Figure 1A) and the identified population was then marked accordingly in the 640-nm channel (Figure 1B). The resulting gate was used to identify extracellular ASC specks in the serum of the human samples (Figures 1C, D). Results were then calculated as ASC specks/µl by dividing the amount of ASC specks in the gate by the volume of the probe measured accounting for dilution during the centrifugation steps. The validation of the FACS measurements was based on an established cell line—THP1-ASC-GFP (Invivogen)—for the isolation of ASC specks (Proske and Morgan, 2001). The measured particles in human serum resemble the isolated specks in granularity, size, and fluorescence intensity. Intra- and interassays on the serum of the volunteers without prior exercise yielded coefficients of variation of 21.5 and 17.2%, respectively (Figures 1E,F).

Blood samples were collected from participants of the Berlin Beat of Running study, a prospective, observational, and investigator-initiated study conducted during the 38th Berlin marathon (clinicaltrials.gov, NCT01428778) (Haeusler et al., 2012). Inclusion criteria were completed for registration for the marathon run: age 35–60 years, a marathon history of at least two marathon runs within the last five years and self-reported running for at least 40 km per week. Exclusion criteria included known cardiac diseases or atrial fibrillation, known or newly detected brain disease, contraindications for (brain) magnetic resonance imaging, severe liver or kidney disease, hyperthyroidism, pregnancy, or lactation. The study protocol is in accordance with the Declaration of Helsinki and was approved by the local Ethics Committee of the Charité-Universitätsmedizin Berlin, Germany (EA4/042/ 11). All participants gave their informed consent before their inclusion in the study.

The baseline visit (t0) took place three days before the marathon. Sociodemographics, cardiovascular risk factors, and the level of physical activity were documented. Participants underwent an assessment of vital parameters (heart rate and blood pressure) and venous blood sampling (Sarstedt Monovette). Within 30 min after the end of the marathon run (t1), vital parameters were assessed and the second blood sampling was done. The third blood sampling with the aforementioned parameters and assessment of vital parameters was done up to 58 h after the race (t2) (Haeusler et al., 2012; Herm et al., 2017).

Primary human coronary artery endothelial cells (HCAEC) were purchased from Promocell (C-12221) and maintained in EGM™ Endothelial Cell Growth Media (CC-3121) and Single Quots™ Supplements (CC-4133). Human umbilical vein cells (EA.hy926) were purchased from ATCC (CRL-2922) and maintained in Dulbecco’s Modified Eagle Medium (DMEM) high glucose (Gibco). Mutant NLRP3-YFP p.D303N human embryonic kidney 293 (HEK) cells were used to isolate overactive YFP-labelled NLRP3-inflammasome oligomers as previously described (Martín-Sánchez et al., 2015). THP1-ASC-GFP cells were maintained in RPMI1640. This cell line was cultured in Dulbecco’s Modified Eagle Medium F-12 (DMEM/F-12) growth media in the presence of Geneticin G418 (2 mg/ml) to maintain mutation (Baroja-Mazo et al., 2014). Internalization of NLRP3-YFP particles in HCAEC was determined by immunofluorescence staining with anti-YFP and anti-rabbit Alexa-488 antibodies. Alexa Fluor™ 555 Phalloidin and DAPI were used for F-actin and nucleus staining, respectively.

Fluorescent p.D303N NLRP3-YFP inflammasome oligomers were isolated and purified from 1 × 10⁷ mutant NLRP3-YFP (p.D303N) HEK cells and resuspended in 100 µl non-denaturating CHAPS buffer following an established protocol (Martín-Sánchez et al., 2015). Fluorescent ASC-GFP specks were isolated and purified from THP1-ASC-GFP cells (Invivogen) and resuspended in 100 µl non-denaturating CHAPS buffer following the same protocol with the addition of a primer with LPS at 1 µg/ ml for 3 h and activation by Nigericin at 5 µM for 1 h before isolation.

HCAEC were treated with NLRP3-YFP particles for 24 h. Cells were fixed with 4% paraformaldehyde, permeabilized and incubated with an anti-YFP (GFP) antibody (#6556, Abcam), anti-YFP Alexa555 antibody (A-31851, Invitrogen), anti-ICAM-1 (#MA5407, Invitrogen), IgG isotype control (#RIgG, #MigG, Dianova), goat anti-mouse Cy3 (#115-165-146, Dianova), donkey anti-rabbit Alexa-488 (A-21206, Invitrogen), or goat anti-rabbit Alexa-647 secondary antibody (A-21245, Invitrogen) to analyze the internalization of extracellular NLRP3-YFP oligomers or ICAM-1 expression. DAPI, Hoechst, and F-actin Phalloidin Alexa 555/ Alexa647 were used for nucleus and F-actin staining, respectively (Gaul et al., 2021). Cells were visualized on a ZEISS Axiovert 200 M fluorescent microscope. Z-stacking (12 subsequently optical slices of 0.8 µm thickness) and 3D reconstruction were performed with AxioVision LE Rel 4.4 software. A minimum of five randomly selected fields were used. To analyze ICAM-1 expression, the evaluation of 8–10 immunofluorescence images per condition and experiment was performed by ImageJR software. First, a gray scale image was used to set a threshold from which the ICAM1 signal was considered positive. This value was applied to all images. Then, the product of the mean intensity and the area of the ICAM1 signal (integrated density) were determined. This determined value was normalized to the cell number, which was counted on the basis of the intact cell nuclei. An ImageStreamX Mk II Imaging Flow Cytometer was also used for internalization studies.

Steroid hormones in serum, which include cortisol, cortisone, testosterone, androstenedione, 17-hydroxyprogesterone (17-OHP), dehydroepiandrosterone sulfate (DHEAS), progesterone, and estradiol, were simultaneously quantified using liquid chromatography-tandem mass spectrometry (LC–MS/MS) as described (Bae et al., 2019).

Monocyte adhesion was assessed by calcein red-orange staining (7.5 µM) of THP1 monocytes and fluorometric analysis (Ex571/Em596) of monocytes on HCAEC. For this purpose, HCAEC were treated with extracellular NLRP3-YFP particles for 24 h and then incubated with stained THP1 monocytes for 4 h. Adherent cells were analyzed fluorometrically and normalized to untreated cells.

Analyses were performed with the Graph Pad Prism (version 7), ImageJ 1.52a, and FlowJo v10. The significance level was set at p < 0.05. Multiple groups were analyzed using a mixed-effect analysis and the Tukey post-hoc test. Two groups were analyzed by a two-sided Student´s t-test. Data are expressed as mean ± standard error of the mean (SEM) if not stated otherwise. A correlation analysis was performed using JASP 0.14.1 (JASP Team, 2020). CorelDraw2018 was used to create the artwork. Statistical outliers were identified using the ROUT method with Q = 1%.

Baseline characteristics are shown in Table 1. Performing a marathon run resulted in a significant 1.5-fold increase in circulating ASC specks immediately after the run compared with the baseline assessment (p < 0.05). The absolute number of ASC specks increased from 11 ± 7 ASC specks/µl to 15 ± 9 ASC specks/µl serum. After the marathon, the concentration decreased toward baseline levels to 11 ± 6 ASC specks/µl (p < 0.01) (Figure 2A).

An exploratory analysis of quartiles of baseline ASC speck concentrations at time t0 revealed that the exercise-induced regulation correlates with the baseline levels. Athletes with low baseline ASC speck levels showed a marked upregulation at t1 that persisted until t2. The middle group of athletes had a distinct increase in circulating ASC specks after the run that recovered until t2. In contrast, individuals with baseline ASC specks in the highest quartile showed no additional upregulation after the marathon. (Figures 2B–D).

Marathon running induces the upregulation of stress hormones and cellular microparticles (Bae et al., 2019; Schwarz et al., 2018). The ASC specks showed a significant positive correlation with stress hormones such as cortisone (R = 0.22), cortisol (R = 0.19), aldosterone (R = 0.24), and progesterone (R = 0.24) (Figures 3A,B) (Bae et al., 2019).

In addition, there was a significant negative correlation with the number of circulating lymphocytes that are known to decrease after severe endurance exercise (R = 0.27) (Santos et al., 2013).

In contrast to the stress hormones, there was no correlation of the concentration of extracellular ASC specks with circulating microparticles. These included endothelial, platelet, and leukocyte microparticles (Figure 3C) (Schwarz et al., 2018). The lack of correlation between the ASC specks and the different types of microparticles is an indicator of the specificity of the circulating ASC particles as released by inflammasome activation rather than more general cell damage.

To test whether extracellular NLRP3-YFP inflammasome particles are taken up by human endothelial cells, immunofluorescent staining of internalized YFP-tagged NLRP3 inflammasomes and cell organelle staining were performed (Figure 4A). Internalization of extracellular NLRP3-YFP inflammasome particles into HCAEC was analyzed by nucleus (DAPI) and cytoskeleton (F-Actin Phalloidin) staining and the detection of intracellular YFP-positive signal. The intracellular NLRP3-YFP particles are located in the region around the nucleus (Figures 4B–D).

Treatment of HCAEC with NLRP3-YFP particles (3 NLRP3-YFP particles/cell for 24 h) increased the expression of the cell adhesion molecule ICAM-1 measured by immunofluorescent staining by 6.9-fold (p < 0.05) (Figures 5A,B).

Additionally, monocyte adhesion on endothelial cells was increased by the extracellular NLRP3-YFP particles by 1.4-fold (p < 0.05) (Figure 5C).

Extracellular inflammasome particles are generated and released when cells undergo pyroptosis. The number of these particles in the blood reflects pyroptotic cell damage. The upregulation of the ASC specks showed an inverse correlation with baseline levels, the lower the baseline, the higher the marathon-induced upregulation. These data suggest a maximum saturation of the exercise-induced number of circulating inflammasome particles. Further studies are needed to address whether lower baseline ASC speck levels relate to individual predisposition or the amount of exercise performed by different athletes in the days before the marathon run. The variance of baseline ASC speck levels may be attributed to different lifestyles, pre-existing conditions or nutrition, and the microbiome in the gut (Youm et al., 2015; Ding et al., 2019). The number of ASC specks in the serum does not correlate with endothelial, platelet, and leukocyte microparticles as a correlate of cell damage. This observation supports the importance of ASC specks as an independent factor. The mechanical damage of myocytes induces the release of typical inflammasome activators such as extracellular ATP or reactive oxygen species which then results in a cascade of inflammasome activation not only in the surrounding muscles but also vascular cells (Bailey et al., 2004; Bauernfeind et al., 2011; Heid et al., 2013; Panicucci et al., 2020). The cell damage of myocytes is documented by the release of typical proteins, miRNAs, and the clinical presentation of muscle soreness after excessive bouts of endurance exercise (Uhlemann et al., 2014; Bernat-Adell et al., 2021). Therefore, the increased ASC specks after marathon running may originate from skeletal muscle cells that are subjected to high physical strain (Pincheira et al., 2021).

An increase of the circulation of ASC in plasma of cyclists post-exercise using Western blot has already been shown; we see similar results in marathon runners and also a similar variance in our baseline values (Nieman et al., 2020). It is also known that exercise results in an increased expression of NLRP3 indicating that it is the relevant sensor protein to these ASC oligomers (Li et al., 2015; Khakroo Abkenar et al., 2019).

The increase in the circulation of ASC particles after the marathon is in agreement with the observation of upregulation of circulating ASC in the plasma of cyclists (Nieman et al., 2020) and an increased expression of NLRP3 post-exercise (Li et al., 2015; Khakroo Abkenar et al., 2019). We used the flow cytometry method to measure extracellular ASC specks. Compared with Western blot (Nieman et al., 2020), flow cytometry has the advantage that the particles are defined more precisely by spiking experiments and that flow cytometry allows a high turnover with highly standardized, parallel, and quantitative measurements.

The study shows for the first time that endothelial cells internalize inflammasome particles leading to inflammation exemplified by the increase of ICAM-1 expression and the increase of monocyte adhesion. These data identify extracellular inflammasome particles as a mechanism that mediates inflammation from the site of damaged tissue to the endothelium. Previous findings of our group demonstrated that extracellular NLRP3 particles are also ingested by human coronary artery smooth muscle cells and macrophages where they induce pro-inflammatory and atherogenic signaling by increasing the extracellular matrix production, promoting the secretion of pro-atherogenic and inflammatory cytokines and cell migration (Gaul et al., 2021).

Extensive exercise with a high mechanical muscular load can mediate systemic inflammation via the release of extracellular inflammasome complexes. We and others propose that these pyroptosis-derived NLRP3 inflammasome particles exert pro-atherogenic effects (Gaul et al., 2021; Sun et al., 2021).

According to the “extreme exercise hypothesis,” long-term excessive training is associated with an increase in the overall mortality (Eijsvogels et al., 2018). Marathon runners show a higher prevalence of coronary artery calcification scores ≥100 Agatston units compared with the general population (36 vs. 22%) indicating coronary artery atherosclerosis (Möhlenkamp et al., 2008). Champion athletes have more coronary calcifications compared with sedentary men (44.3 vs 22.2%) (Merghani et al., 2017). One critical factor for cardiovascular risk is endothelial cell physiology (Souilhol et al., 2018). Therefore, the internalization of circulating pro-inflammatory inflammasome particles into endothelial cells and smooth muscle cells is likely relevant in the progression of cardiovascular diseases.

Interestingly, athletes have more stable calcified plaque phenotypes than controls and fewer mixed plaques (Merghani et al., 2017). This could indicate different underlying pathomechanisms of atherosclerosis in athletes and the general population. The role of inflammation and immune cells in the development of various plaque types is incompletely understood. Circulating inflammasome particles may be an important component in this process by increasing monocyte adhesion to endothelial cells (Waring et al., 2021).

There are limitations to our study. Our study population included only non-professional athletes who completed high-level endurance exercises, so transferring the data to other intensities or forms, such as strength exercise, is not possible. Nevertheless, marathon runs at competitive levels are typical for extremely intensive endurance exercise and are practiced worldwide by an increasing number of individuals. The studied population is large for a marathon study but addresses a selected and homogenous population. Therefore, the results may not apply to the general population. Measurements were performed in human serum. Other tissues are not available. Because the serum was stored for several years, the proteins may have been degraded (Cuhadar et al., 2013). Since we are only measuring one relevant protein and this should be changed in the same way in all samples, this should not change the dynamic of the ASC specks. This is also the reason why cytokines could not be measured in this study, since cytokines are extremely dependent on the storage time and conditions (Jager et al., 2009).