Reversion of aortic valve cells calcification by activation of Notch signalling via histone acetylation induction

As shown in Supplementary Fig. 2, the higher senescence level and the increased propensity of sVIC to calcify did not depend on differences in the age of the donor individuals, the so called 'chronological age'. Hence, we explored the possibility that the different senescence levels of the cells in the two pathologic settings are epigenetically programmed,²⁸ such as in vascular stiffening and calcification,²⁹ and reflect the 'biological aging' acceleration.^30,31 To this aim, we assessed the level of global DNA methylation and we measured the methylation level of the C5 and C7 CpG islands within the promoter sequence of ELOVL2 - a gene associated with biological aging.¹⁷ While the analysis of the global DNA methylation revealed an increase in the 5'-Methylcitosine content in sVICs compared to iVICs (5mC; Fig. 2f), not compensated by variations in the level of the product of the 5mC oxidation, the 5'-Hydroxymethylcitosine (5hmC; Fig. 2g),³² the analysis of the hotspots in the ELOVL2 gene promoter revealed a higher variability of methylation in iVICs vs. sVICs (Fig. 2h, i). To exclude that the differences in the methylation level in the cells are due to stochastic variations, we performed linear regression analyses by plotting the data of C7/C5 methylation in each individual cell line from the two pathologic setting. As shown in Fig. 2j, k, the methylation of the two hotspots was consistently correlated, even if it was not correlated to the age of the donor individuals (Supplementary Fig. 3), again indicating a patient-specific programming of the cellular biological aging unrelated to chronological age.

Our attention was then drawn to histone modifications potentially characterizing the differences in senescence observed in sVICs vs. iVICs. In particular, we mapped modifications that are involved in a generalized increase in transcriptional activation (Acetylation of lysine-9 on histone H3 and lysine-16 on histone H4), repression (Methylation of lysine-20 on histone H3), and the so-called 'poisoning' for transcription (tri-methylation or acetylation of lysine-27 on histone H3).³³ Panels l–p in Fig. 2 show the result of this analysis, clearly indicating an excess of activating modifications in iVICs (e.g., H3K9Ac, H4K16Ac) and the presence of an elevated level of specific repressor marks such as H4K20me(3), possibly indicating an enhanced amount of inactive, but with potential to be activated, chromatin. All these effects, again, were independent of the donors' age (Supplementary Fig 4). To verify whether sVICs differ from iVICs for expression of epigenetic-active enzymes and chromatin remodelling factors, we analysed the expression of genes encoding for epigenetic writers/readers involved in chromatin structure/activity using pathway-specific RT-qPCR arrays and RNAs extracted from iVICs and sVICs, both at p3 and age-matched (Supplementary Fig 5). The results of unsupervised hierarchical clustering evidenced a differential expression of 29 genes (Supplementary Table 2), encoding for factors involved in chromatin architecture with specific functions in gene repression (e.g., RING1, PHC1, and PCGF- members of the PRC1 Polycomb repressor complexes³⁴), histone H3 tails mark reading (e.g., PHF1, -2, -6, -7, -21A/B³⁵), and SWI/SNF complex function (e.g., SMARCA4/BRG1, PBRM1³⁶) (Fig. 2q). It was important to note that genes such as MECP2, involved in silencing chromatin containing methylated CpG islands, or HDAC11 encoding for a protein with histone and non-histone deacetylase activity³⁷ were expressed at lower levels in iVICs compared to sVIC. Conversely, iVICs expressed higher NuRD components MBD3 and MTA2 (balanced by a higher expression of MTA1 in sVICs),³⁸ and NSD1 a histone methyltransferase with specificity for H3K36Met(2) involved in the maintenance of active chromatin.³⁹ Altogether, these results suggest a higher degree of active chromatin or chromatin poised for transcription in iVICs.

We next investigated whether the increased epigenetic marks in the chromatin of sVICs treated with SPV106 were associated with variations in the general chromatin accessibility in regions associated with pathways connected to valve disease. To this aim, we performed ATAC-seq⁶⁵ on chromatin extracted from sVICs treated or not with SPV106. After performing alignment of the reads to the GRCh38 human genome by BWA-mem,⁶⁵ genomic regions with high levels of transposition/tagging events (i.e., those with high chromatin accessibility) were determined using the MACS3 peak calling algorithm.⁶⁶ Then, we analysed the loci with significantly higher or lower chromatin accessibility in a pairwise DMSO- vs. SPV106-treated cells comparison, performing a differential analysis with the DaMiRseq⁶⁷ (peak filtering and normalization) and LIMMA⁶⁸ (statistical analysis) R packages, in which a paired test has been implemented. Of the 128,862 loci identified by ATACseq (Supplementary Data 1) a), 32,439 passed stringent quality filtering, namely more than 20 reads in all samples (Supplementary Data 1) b) Of them, 771 bore higher accessible, and 1060 lower accessible chromatin (Supplementary Data 1, c). Unsupervised clustering of the data showed a clear separation between genes with higher or lower chromatin accessibility due to treatment (Fig. 5f), with a percentage of 58.7 gene promoter/enhancers, 7.5 insulators and 7.9 intergenic regions (Supplementary Fig 11) In order to derive insights on the transcriptional pathways potentially affected by the epigenetic treatment, we next analysed the dataset of the genes with significantly different chromatin accessibility with Reactome.⁶⁹ By this, we were able to identify 60 pathways containing genes with significantly increased chromatin accessibility and 74 pathways encompassing genes with less accessible chromatin conformation, suggesting, respectively, activated or repressed conditions. These pathways, listed in Supplementary Tables 5 and 6 are graphically represented in Fig. 5g, h. As shown, they include Notch1- and Rho-related pathways among those with a potentially positive or negative activation, respectively.

The increase in the opening of the chromatin in the promoter region of Notch1 and Sox9, but not of Runx2, was evident by the analysis of the reads distribution (Fig. 5i and data not shown), thus providing a rationale for the higher mRNA levels observed for the two genes in SPV106-treated cells. To confirm that treatment with the epigenetic drug determined structural modifications in the promoters of these genes, we performed Chromatin Immunoprecipitation assay (ChIP) in sVICs treated with DMSO or SPV106 using specific antibodies recognizing acetylated lysine-16 and three-methylated lysine-20 on histone H4 (two of the significant histone marks restored by the treatment, Fig. 5d, e), followed by PCR-based quantification of enriched promoter sequences of the three genes. To do this we chose primers encompassing the region 1000 base pairs (bp) upstream (−1000) of the described transcriptional start site (TSS), as indicated by the ATACseq (Fig. 5i). The results of ChIP-qPCR indicated a positive enrichment of the H4K16Ac (but not of the H4K20me(3)) onto Notch1 and Sox9 promoters at −1000bp (Fig. 5j). Given that the highest H4K16Ac chromatin enrichment occurred on Notch1, we also analysed a wider portion of the promoter using PCR primers pairs encompassing the −1500 to 0 bp (TSS) genomic region. As shown in Supplementary Fig 12, the −1000 region was the only one that resulted significantly enriched by ChIP with H4K16Ac-specific antibody, identifying this site as the most sensitive to the histone H4 hyperacetylation resulting from SPV106 treatment. To finally validate the involvement of Notch signalling in reducing the sVICs senescence mediated by SPV106, we used DAPT, a specific inhibitor of the Notch pathway,⁷⁰ alone or in combination with the epigenetic drug. Results showed that DAPT reverted the positive effect of SPV106 on sVICs senescence, suggesting a direct involvement of Notch in suppressing the sVICs pathologic phenotype. Furthermore, treatment with SPV106 inhibited the transcription of at least one of the tested Notch direct transcriptional targets (HES-5)⁷¹ (Fig. 5k) and reduced the secretion of two of the tested SASP cytokines IL8 and GM-CSF (Fig. 5l, m). By contrast, the drug did not affect the expression of the genes encoding for fibrotic genes TGFβ1 and its receptor TGFβ-R1 (Supplementary Fig 13), or of genes relevant for oxidative stress NQO1, TXNRD1 and PRDX4⁷² (Supplementary Fig 14), showing the specificity of the treatment with SPV106 to reduce sVICs pathologic phenotype by targeting senescence/calcification and, at least in part, SASP phenotype, but not progression toward fibrosis.

Apart from the reduction of calcifications, SPV106 also had a beneficial effect on the overall cardiac function and valve motion, as detected by echocardiographic assessment. For example, the administration of Vitamin D caused a decrease in the ejection fraction, and this decline was prevented by administration of SPV106 (Fig. 6c). More directly on the valve function it was the effect of the drug on the cusp separation⁷⁸ that was decreased by the Vit-D treatment (consistently with a more rigid structure) and preserved in mice receiving SPV106 (Fig. 6c), and the time-to-peak velocity⁷⁹ that was delayed by the Vitamin-D and restored at the control level by the epigenetic drug (Fig. 6d). Interestingly, another parameter linked to aortic valve calcification, the so-called echocardiographic aortic valve back scatter (AVBS), directly referable to calcification,⁸⁰ was also reduced by SPV106 treatment (Fig. 6e). Other clinically relevant valve functional parameters such as, e.g. the Vmax were not affected by the vitamin-D and the vitamin-D/SPV106 treatments, likely due to the early phases of valve calcification detectable by our protocol.

Mechanistically, valve interstitial cells calcification process has been for a long time assimilated to osteoblast-like differentiation, with the identification of common molecular pathways (e.g. the Runx2/BMP pathway).⁸⁴ More recently, new results have highlighted the relevance of mechanical factors such as the cellular stress consequent to strain, and the response to pro-inflammatory signalling.^3,85,86 In keeping with this evidence, nano-analytical methods have established peculiar differences between the mineralization in the bones vs. that of the valves, characterized by secretion of mineralized microparticles and similarities with the vascular calcification.^5,12

Given the implication of senescence in vascular calcification, we hypothesized that VICs obtained from patients with different AoV pathologic settings (insufficiency and stenosis) may differ for the level of cellular senescence and epigenetic setups.^87–90 Data presented in Figs. 1 and 2 support the hypothesis that sVICs calcification is associated with a reduced capacity of the cells to grow (the so called proliferative arrest⁹¹) related to a pre-established senescent status programmed before the beginning of the in vitro cell culture independently of the patient chronological age. This condition was absent in iVICs, which exhibited a consistently shorter doubling time and a reduced β-Gal expression. The higher expression of senescence markers in sVICs of subjects with comparable chronological age to that of iVICs donors, led us to test the possibility that differences in the epigenetic 'clock'⁹² between the two populations may determine the variation in the calcification potential of the cells from the two pathologic settings. A first evidence supporting this conclusion emerged from the differences in the content of methylated DNA, showing around a double level of 5mC in sVICs vs. iVICs independently of donor ages (Fig. 2). Interestingly, the content of 5hmC, the first intermediate generated by TET enzymes during DNA demethylation process,⁹³ was not different in sVICs vs. iVICs. This suggests that the increase in the 5mC in the two pathologic settings might derive from an imbalance in DNA demethylation due to differences in the availability of TCA cycle intermediates such as αKG/succinate, known to function as cofactors in TET-dependent DNA demethylation.⁹⁴ A second indication that the epigenetic clock could be shifted in sVICs compared to iVICs was the difference in the extent of ELOVL2 CpG islands methylation in the two cell types (Fig. 2). Although there was not a significant difference in the absolute amount of the C5 and C7 ELOVL2 CpG methylation sites, we observed a more variable level of CpG methylation in iVICs and a more consistent and less variable epigenetic senescence in sVICs, with lower inter-individual variations. Whether differences in the general DNA methylation of the sVICs vs. iVICs reflects the presence of somatic mutations in epigenetically active enzymes such as DNMT3A and/or TET2, the determinants of the so-called clonal hematopoiesis of indeterminate potential (CHIP) strongly related to calcification of the aortic valve,⁹⁵ is for now only a matter for speculation even though it may establish a rationale for an imbalance in the DNA methylation setting observed in the two classes of valve-resident cells.

The levels of histones post-translational modifications have been correlated to vascular aging and calcific disease. For example, it was found that an age-dependent increase in miR-34a determines the downregulation of SIRT1, a class IV HDAC,⁴⁶ and this is crucially involved in vascular aging and calcification.¹⁶ However, the connection between histones post-translational modifications and valve calcific disease is less clear. In fact, from the relatively few available reports, it emerges that treatment of VICs with inhibitors of different HDAC classes may have beneficial, but also detrimental effects on VICs calcification.⁸⁸ At the same time, in animal settings, the inhibition of p300 histone acetyltransferase was found to reduce VICs calcification by reducing the level of H3K9Ac in VICs derived from healthy valves.⁸⁹

Because of these contrasting results, we decided to profile some histones post-translational modifications controlling active and repressed chromatin. By this approach, we found several differences in these marks, including higher levels of histones H3/H4 acetylation (H3K9Ac, H3K27Ac, and H4K16Ac), and a reciprocally inverted abundance of H3/H4 methylation (H3K27Met(3) and H4K20Me(3)) in iVICs vs. sVICs (Fig. 2). These results are at least in part conflicting with the existing literature produced using animal AoV-derived VICs,^87,88 probably reflecting a different balance of chromatin remodelling enzymes and variations in the expression levels of the genes encoding for the various histone variants in the two human cell types compared with animal-derived VICs often employed in molecular studies.⁹⁶ The existence of a clear difference in the expression of chromatin interacting/remodelling factors was confirmed by the analysis of specific transcripts presented in Fig. 2, whose results revealed variations in the expression level of several components of the PRC1/2 complexes,³⁴ and DNA methylation readers connected with histone methylation/acetylation.⁹⁷

The increased cellular senescence in sVICs was reminiscent of what observed by Vecellio et al. in cardiac fibroblasts from diabetic vs. non-diabetic patients.¹⁹ In particular, in that study, it was found that the prolonged exposure to hyperglycaemia in vivo determined a senescent status of the cells determining a reduced ability to grow, an elevated DNA methylation, and a variation of crucial histone H3 marks, such as H3K9Ac (decreased), H3K9me(3) and H3K27met(3) (increased). Interestingly, in the present study, the reduced levels of histone acetylation coincided with a reduction in GCN5/pCAF histone acetyl-transferase activity, and treating cells with SPV106 reversed the senescent phenotype. In analogy with these findings, treating sVICs with the drug reduced, i) the level of senescence detected by the β-Gal staining in normal/high calcium-containing culture media, ii) the level of calcification, iii) the percentage of cells with nuclear p16, iv) the expression of p21 and, v) the percentage of cells expressing PCNA (Fig. 3). Importantly, the senescence-retarding effects of SPV106 lasted even after the drug withdrawal, suggesting, at least for the culture passages considered, the permanence of an epigenetic memory of the treatment in sVICs (Fig. 3). Conversely, treatment with garcinol, a HAT inhibitor, and ITSA, an HDAC activator, increased the senescence level of iVICs (Fig. 3). Altogether, these results show that altering the level of histone acetylation can accelerate or reduce senescence and consequent calcification in human VICs.

Indications about possible epigenetically-controlled transcriptional pathways affected by SPV106 emerged from the analysis of transcripts differentially expressed in control vs. treated sVICs (Fig. 4). As expected, the pathways that were enriched with the highest significance described functions relative to chromatin modification/remodelling, and were characterized by a general downregulation of genes encoding for components of the chromatin modification and reading machinery. Given the absence of data on genomic regions that could be affected by different occupancy by some of these gene products (i.e., by ChIP-Seq), it is impossible to conclude what the transcriptional targets potentially interested in these changes are. On the other hand, it was interesting to note the implication of survival pathways including PI3K and PTEN genes and p53 activity involving acetylation⁹⁸ and TP53BP1,⁹⁹ and of pathways involved in SUMOylation and cellular senescence with concomitant upregulation of BMI1 and CDKN2A (p16). This collectively suggests that treatment with SPV106 reduces cellular senescence and calcification by increasing cellular longevity,⁵⁴ improving the repair of DNA double-strand breaks,⁵³ or increasing self-renewal.^59,60 A final important prediction emerging from the functional genomics analysis of the transcriptional signatures in SPV106-treated vs. control sVICs was the upregulation of pathways connected to Notch signalling, an essential effector of valve morphogenesis and disease.²⁰ This prediction was validated by RT-qPCR using RNA extracted from cells treated with SPV106 in which an evident upregulation of Notch1 and Sox9 mRNAs and a decrease of Runx2, either at mRNA or protein/localization levels were observed. Furthermore, SPV106 enhanced nuclear transfer of the Notch intracellular domain with potentially wider transcriptional effects^64,100 (Fig. 4). These results are in agreement with the transcriptional level of these genes observed in iVICs, in which expression of Notch1 and Sox9 was higher, and that of Runx2 was lower than in sVICs, confirming the role of Notch in repression of VICs phenotype.⁶² In summary, activation of pCAF/KAT2B by SPV106 elicits a multilevel response in human VICs determining direct transcriptional effects on the control of master genes of VICs calcification under Notch-signalling control.

The transcriptional changes observed in sVICs treated with SPV106 (Fig. 4) were accompanied by an apparent reversion of histones H3/H4 acetylation and methylation marks to levels typical of iVICs (Figs. 2 and 5). Remarkably, these changes were neither associated to a reduction in general DNA methylation nor a specific demethylation of ELOVL2 C2/C7 age-sensitive CpGs (Supplementary Fig. 10). Thus, even if SPV106 did not function as a canonical senolytic drug able to eliminate senescent cells from cultures and tissues¹⁰¹ and/or shifting back the epigenetic clock based on DNA methylation status,⁹² in our experiments it had a senomophic effect in retarding the sVICs senescence/calcification process by promoting a change in the transcriptionally active chromatin status related to H3K9Ac, H4K16Ac and H4K20me(3) histone marks.¹⁰² The genome-wide effects determined by the treatment on chromatin accessibility were confirmed by the ATACseq analysis performed on chromatin of DMSO- and SPV106-treated sVICs (Fig. 5). Apart from modifying chromatin accessibility in almost two thousand loci, the GO analysis of the genes with more open or more closed configurations, and thus bona fide more transcriptionally active or inactive, led to the identification of numerous pathways that were coherently regulated with the pathological status of the cells. For example, several networks encompassing genes with more open chromatin structure in SPV106-treated cells were related to the control of cell cycle/survival (e.g. by TP53 and PTEN) or, again, activation of the Notch-dependent pathway, confirming the transcriptional activation found by the transcriptomic profiling. By contrast, several networks encompassing genes with more closed chromatin configuration were related to Rho-signalling, a positive regulator of valve calcification.¹⁰³ The experimental validation of ATACseq data by ChIP assays confirmed that the proximal promoter regions of Notch1 and Sox9 were associated with chromatin containing the activating H4K16Ac, but not the gene silencing H4K20Me(3), modification.⁹⁶ This result is in line with previous evidences showing the relevance of the epigenetic control of Notch1 for valve pathology,¹⁰⁴ and clearly points to a direct engagement of the Histone H4 acetylation by pCAF/KAT2B activators in suppressing the senescence and the calcification of VICs via activation of NICD. At the moment, it is impossible to conclude on the functional significance of the other histone marks, e.g. the variation in the three-methylation and acetylation of H3K27, two histone modifications with antithetic function on Notch pathway regulation¹⁰⁵ - detected by our analyses, as well as on the negative control of Runx2 by SPV106 treatment. To address this will require specific experiments, such as targeted histone proteomics,¹⁰⁶ and/or evaluation of complementary epigenetically-regulated pathways in control of valve calcification.¹⁰⁷

The ex vivo and in vivo results obtained, respectively, using a bioreactor-based and the Vitamin D-dependent calcification models²³ (Fig. 6) shed light on the ability of the epigenetic treatment to reduce the deterioration of the valve due to hypercalcaemic conditions. In keeping with what reported for other anti-calcification treatments in genetic models of valve disease, e.g. the Klotho knockout mice,¹⁰⁸ the employment of the drug reduced the occurrence of histologically-detectable calcifications and prevented the functional deterioration of the valve. Interestingly, in the ex vivo model, calcification in the valve leaflets was not characterized by abundance of cells with senescence markers such as p16 and p21 (Supplementary Fig. 15). This was expected, considering that the model has been designed to assess the effect of hemodynamic forces on valve calcification, likely occurring through endothelial to mesenchyme transition and not cellular senescence.²³ On the other hand, even in this system, SPV106 determined a reduction of calcification and of cells expressing calcification markers, witnessing the role of epigenetics in control of VICs calcification, also independently of VICs senescent phenotype.

In vivo, the effect of SPV106 was demonstrated by an echocardiographic assessment of the heart function taking as a reference parameters describing the valve's motion (the cusp separation), the transvalvular flow (time to peak)^78,79 and the echocardiographic aortic valve scatter (Fig. 6). Interestingly, treatment with the epigenetic drug also protected the heart from the vitamin-D induced systolic dysfunction as assessed by the recovery of the ejection fraction. Although is not possible to exclude a direct cardioprotective effect of histone hyperacetylation, as observed in myocardial infarction models,¹⁰⁹ we remark on the clinically exciting opportunity to maintain intact the systolic function by preserving the valve structural integrity.¹¹⁰

A second limitation of the study could be represented by our choice to exclude cells from the oldest donors, to match the chronological ages of the two-donor groups and exclude the chronological age as a confounding factor. While this might have caused an under-representation of the most extreme cell phenotypes, reducing the biological differences between the two groups, it helped us to delineate, for the first time and with a higher precision the epigenetic setup of the two pathologies and to find a possible treatment for aortic stenosis, the most prevalent of the two diseases and the one with a higher impact in elderly people.¹¹¹ Given the lack of treatment specificity, direct valve targeting strategies (e.g. cell specific nanoparticle drug delivery) will have, however, to be adopted to avoid side effects of future treatments.

It is finally important to highlight that, while our results are coherent with the reported effect of the SPV106 on inhibition of the cellular senescence in diabetic cardiac fibroblasts and acceleration of cutaneous wounds closure,^19,112 it also contrasts with numerous evidences showing that reduction of histone acetylation has a positive readout on VICs calcific programming.^87,89 A second possibility to explain this contradiction is that the activation of KAT2B/pCAF might have a different readout on histone acetylation/methylation than that of treatment with histone deacetylase inhibitors employed in other reports, and that additional epigenetic modifications, in additions to those revealed in the present study, concur to reduce the senescent/calcific phenotype of sVICs treated with SPV106. More conclusive experiments using chromatin profiling techniques such as, e.g., ChiP-Seq or chromatin proteomics should be done to resolve this relevant issue. A third is that most of the HDACi studies have been performed with animal-derived cells, particularly porcine VICs.⁸⁷ In this regard, a difference in the response of sVICs to treatment with SPV106 compared to HDACi could be contemplated. A final possibility is that the activation of KAT2B/pCAF by SPV106 has a direct effect on the expression or function of relevant genes involved in calcium homeostasis like Klotho or FGF23, as in part demonstrated in other cell types and in vivo in the literature.^113,114 Studies are underway using reference models of valve calcific disease and naturally aged animals to discriminate between these possibilities.

The stenotic and insufficient aortic valves were collected during routine open chest aortic valve replacement procedures. The use of this material was approved by the ethical committee at Centro Cardiologico Monzino, IRCCS (first approval date June 19, 2012) and upon agreement on an informed consent. It also conformed to the Declaration of Helsinki. Primary human VICs (sVICs and iVICs) were isolated as previously described.⁷ The methods employed to perform specific analyses on cellular senescence and DNA/RNA material are specified in the supplementary information.

Animal procedures were performed in conformity with the guidelines from the Directive 2010/63/EU of the European Parliament on protecting animals used for scientific purposes and following experimental protocols approved by the Committee on Animal Resources of Cogentech (#824-2020). Male C57BL/6 J mice were used as already described.^14,16 Further information about the model and the treatment of mice with SPV106, as well as the immunohistochemical and echocardiography methods that were employed to analyse the expression of markers and the valve function are epcficied in the online supplement.

Experiments were performed with hearts of 3–6 months old mice (C57BL/6JRj) according to established protocols.^23,73 The study was approved by the animal welfare committee of the Leiden University Medical Center and conformed with the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes. Details about the experimental procedure to induce valve calcification are provided in the online supplement.

Aortic valve interstitial cells were derived from pathologic aortic valves with insufficiency or stenosis. Cells were extracted following a simple digestion method as already described by us.⁷ Evaluation of cellular senescence was performed according to beta-Gal staining and doubling time while other parameters like expression of senescence markers, calcification, epigenetic status, chromatin accessibility, RNA profiling and bioinformatics are specified in the online supplement.

Data analyses and graphing were performed using GraphPad Prism software (version 9.1.2). The general criteria adopted in the cell biology data analyses was to use unpaired t-test/ANOVA when comparing cells from different donors (e.g., typically the iVICs vs. sVICs comparisons), and to employ pairwise methods to assess the effects of treatments on cells (e.g., the effects of drugs on sVICs vs. DMSO-treated sVICs). The circles overlapped to each plot and bar graphs represent the individual values of independent samples (cells from different donors) and their sum indicates the n of the samples included in the analyses.