**3. Results**

In the development of atherosclerosis, the endothelial dysfunction is one of the beginning steps or a permissive status of the endothelial layer for the onset of the pathology. Treatment of HUVEC cells with TNF-α can mimic the systemic inflammatory status of the endothelium [30,31].

In order to confirm the inflammatory condition of HUVEC cells, we analyzed nitric oxide synthases (NOSs) expression. Nitric oxide (NO) is important to maintain normal vascular functions and endothelial integrity. As expected, the endothelial isoform NOS3 was the most expressed form in HUVEC (Figure 1A), and the expression levels of NOS3 and NOS1 were significantly decreased after TNF-α stimulation, while NOS2 showed a non-significant tendency to decrease (Figure 1B). These data agree with the literature in which in vitro studies confirm the defect in the NO production in isolated atherosclerotic blood vessels [7,32].

Migration recorded through 8 h and vitality at 24 and 48 h were not affected by the cytokine (Supplementary Figure S1).

The extracellular matrix expressed by endothelial cells is commonly referred to as glycocalyx and has an important role in controlling shear stress from laminar flow through mechano-transduction mechanisms [14] and inflammation, thus controlling cell adhesion, motility, and proliferation [15–17].

The HA production has also an important role in the maintenance of cell homeostasis and in activation of different signal transduction pathways [33]. In HUVEC cells, HAS3 mRNA was the most abundant (Figure 2A), whereas HAS1 messenger was not detected (data not shown). Interestingly, after TNF-α stimulation, HAS2 increased expression while HAS3 was decreased (Figure 2B). In order to evaluate the glycocalyx of the HUVEC, we quantified the glycosaminoglycans from the membrane and from the medium with no significant differences (Supplementary Figure S3). The pericellular coat surrounding the endothelial cells was measured and showed a significant increase after TNF-α stimulation, which was mainly constituted of HA as demonstrated by enzymatic digestion (Figure 2C).

**2021**, , x FOR PEER REVIEW 6 of 14

α α **Figure 1.** Effect of TNF-α on NO synthetic enzymes in HUVEC. (**A**) relative expression of NOSs (neuronal NOS1, inducible NOS2, and endothelial NOS3) in HUVEC. (**B**) NOSs expression in HUVEC untreated (control) and treated with TNF-α (0.1 µg/mL) for 24 h. Data are mean ± S.E.M. of three independent experiment, \*\*\* *p* < 0.001. α

α

**Figure 2.** Effect of TNF-α on Hyaluronan synthesis in HUVEC. (**A**) HASs expression profile in HUVEC. The reference gene used for normalization was β-actin and the normalizer HAS2 expression. (**B**) Relative expression of HAS2 and HAS3 after TNF-α stimulation (24 h). The reference gene used for normalization was β-actin and the normalizer untreated samples. Data are mean ± S.E.M. of four independent experiments, \*\*\* *p* < 0.001. (**C**). Particle exclusion assay performed on HUVEC untreated (control) and under TNF-α stimulation for 24 h. To clarify the HA composition of the pericellular matrix, we digested HA with 2 U/mL of Hyaluronate Lyase from Streptomyces hyalurolyticus (HYAL) before the addition of erythrocytes. Original magnification 40×. Values represent the measure of the single cell pericellular area, and the red bars are the mean of three independent experiments, \*\*\* *p* < 0.001 and \*\* *p* < 0.01.

Due to the lining of the vessels, ECM is also important in recruitment and activation of immune cells and of platelets from the blood [34–36]. Among all the HSPGs, Syndecans

are a family of four transmembrane proteoglycans acting as co-receptors interacting with different molecules including growth factors, matrix components, and cytokines that are present in glycocalyx [11]. In HUVEC, the main Syndecans expressed are the -3 and -4 isoforms (Figure 3A), but only Syndecan-4 increased during TNF-α stimulation from 24 up to 48 h (Figure 3B). The core protein of the proteoglycan was evaluated in the cell extraction and shown in Western blot and turned out to be increased, even if not significantly (Figure 3C). The Western blot shows three different bands positive to antibody recognition. The three different bands at around 27, 37, and 45 kDa can be the proteoglycan with different GAG chains [37], Syndecan-4 bound to growth factor or matrikines, and/or its homo- or hetero-oligomerization forms [38]. α

α

β α

**2021**, , x FOR PEER REVIEW 7 of 14

β

α

As reported, the NDST1 has the capacity to bind to EXT2, and EXT1 and EXT2 expressions affect the N-sulfation degree, suggesting that the overexpression of all the three enzymes happens simultaneously, as shown in Supplementary Figure S3.

**Figure 3.** TNF-α influence on Syndecans expression. (**A**) Syndecans expression profile in HUVEC. The reference gene used for normalization was β-actin and the normalizer the Syndecan-1 expression level. (**B**) Syndecans isoforms expressions in HUVEC control and after 24- and 48-h of TNF-α stimulation. The reference gene used for normalization was β-actin and the normalizer untreated samples. Values represent mean ± S.E.M. (n = 3), \*\* *p* < 0.01. (**C**) Western blot analysis of Syndecan-4 (SDC4) protein in HUVEC control and treated 24 h with TNF-α. Bar chart represents normalized mean ± S.E.M. of two independent experiments and the figure is a representative SDS-PAGE.

The GAG moiety of the HUVEC HSPGs was analyzed by enzymatic digestion followed by HPLC analysis, and the disaccharide percentages are reported in Table 1. The main drastic difference seems related to the increment of N-sulfation on glucosamine residue. The higher amount of N-sulfation correlates well with the increment of expression of the enzyme NDST1, heparan sulfate N-deacetylase/N-sulfotransferase 1 (Supplementary Figure S3), that catalyzes both the N-deacetylation and the N-sulfation of glucosamine.

β

β

α

α

**2021**, , x FOR PEER REVIEW 8 of 14

α

**Table 1.** HPLC analysis of the main HS/HE disaccharides. GAGs were isolated from plasma membrane and from culture medium of HUVEC control and TNF-α treated (24 h). To obtain HS/HE disaccharides, we digested GAGs with heparinases. After AMAC derivatization, the disaccharides were analyzed by means of HPLC. Data are expressed as % area of each HS disaccharide/ % area total. The N-sulfation in bold (NS) is catalyzed by NDST1. Values are mean ± SD of three independent experiments \*\* *p* < 0.01. UA: uronic acid; GlcNAc: N-acetyl; GlcNS: N-sulphonyl glucosamine: S: sulphate group [39]. α


The major event in atherosclerosis onset is the accumulation in the sub-endothelium of lipids driven by the lipoprotein LDL [3]; nevertheless, the data about the events causing the LDL particle transcytosis and accumulations within the tunica intima are still scant.

To test whether the endothelial permeability is altered under the inflammatory condition, we incubated a continued layer of HUVEC in a transwell system with FITC-dextran using a permeable membrane with a cut-off unable to let the cells pass. As reported in Figure 4A, the presence of the HUVEC layer (control) blocks the free passage of the fluorescent dextran, and the same cell under the inflammatory condition of TNF-α increases the blocking by a significant, even if small, amount. α

**Figure 4.** Transwell permeability assay. (**A**) FITC-dextran flow through HUVEC monolayer. Confluent HUVEC cells in the upper chamber of a transwell system +/− TNF-α were added with 1 mg/mL of dextran conjugated with FITC. After 24 h, the medium of the lower chamber was collected and FITC fluorescence was measured, \* *p* < 0.05 and \*\* *p* < 0.01. (**B**). FITC-dextran flow through siRNA control (siØ ctr) or siRNA against SDC4 HUVEC monolayers; data are mean ± S.E.M. and n = 3.

Since Syndecan-4 exerts various effects on the endothelial glycocalyx, with particular regard to TNF-α induced endothelial modifications [40], and has a pivotal role in the dynamics of focal adhesion [41] and in the formation of networks at gap junctions [42], we investigated the HUVEC permeability of SDC4-silenced cells (Figure 4B) (silencing efficiency 80%, data not shown). The abrogation of the proteoglycan does not alter the dextran passage, even considering the high silencing levels, thus indicating a complex metabolism and turnover for the proteoglycan as well as multiple control levels for the layer permeability.
