SARS-CoV-2 infected patients (n = 50), confirmed by RT-PCR, were enrolled at the National Institute for Infectious Diseases (INMI) “Lazzaro Spallanzani” (Rome, Italy). Patients’ characteristics are summarized in Table 1. All patients had severe COVID-19 disease and required intensive care unit admission (ICU). Seventy-eight percent of patients presented one or more bacterial or fungal co-infections such as Acinetobacter spp, Klebsiella spp, Pseudomonas spp, Staphylococcus spp, Streptococcus spp, Aspergillus spp, Candida spp, and others. Sixty-four percent of co-infected patients had bacteremia. Among the patients with coinfections, 15 were infected with both Gram-positive and Gram-negative bacteria, 13 with Gram-negative bacteria, and 11 with Gram-positive bacteria. All co-infected patients received antibiotic or antimycotic treatment, and 2 were also treated with remdesivir. Among the subjects without co-infection, two received an antibiotic treatment. Eighty-four percent had other co-morbidities, such as obesity, diabetes, and cardiovascular diseases. The 36% received anti-COVID-19 vaccination. The presence of any tumor type represented an exclusion criterion. Signed written informed consent was obtained from all patients. Healthy individuals (HD, n=10) were included as controls. The study was approved by the Institutional Review Board of the INMI (approval number: 9/2020). Granulocyte-colony stimulating factor (G-CSF) mobilized donors were enrolled at Bambino Gesù Children’s Hospital, Rome, Italy. They received subcutaneous administration of G-CSF for five days (until apheresis) at a dose of 10− 12 μg/kg/day. Informed consent was obtained by G-CSF mobilized healthy donors to participate in this study, which was approved by the Bambino Gesù Children’s Hospital (Rome, Italy) ethics committees (Prot. n. 132 28/01/2019).

Platelet-rich (PRP) was obtained from heparin-treated whole blood by centrifugation for 10 min at 1100 rpm, at 4°C. was maintained on ice until use.

Ten ml of heparin anti-coagulated whole blood samples were obtained from patients. Density gradient centrifugation (Lympholyte-H, Cederlane, Burlington, USA) was used to isolate peripheral blood mononuclear cells (PBMC). PBMC were resuspended in RPMI 1640 (Corning Incorporated, New York, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (EuroClone, Milan, Italy), 2 mmol/L L-glutamine (Corning Incorporated, New York, USA), penicillin/streptomycin solutions, 100X (Corning Incorporated, New York, USA) and 10 mmol/L HEPES buffer (N-2- hydroxyethylpiperazine-N-2-ethane sulfonic acid, Gibco, USA). PMN-MDSC were isolated from PBMC from COVID-19 patients by magnetic cell isolation technology. Positive selection was performed using CD15+ microbeads following manufacturer’s instruction (Miltenyi Biotec, Bergisch Gladbach, Germany). PMN-MDSC purity was >95% as verified by flow cytometry (data not shown). PMN-MDSC viability was evaluated by trypan blue exclusion. PBMC were also obtained from G-CSF mobilized healthy donors after density gradient centrifugation (Lympholyte-H, Cederlane, Burlington, USA). PMN-MDSC were isolated from PBMC of mobilized donors using CD66b+ microbeads following manufacturer’s instruction (Miltenyi Biotec, Bergisch Gladbach, Germany). PMN-MDSC purity was >98% as verified by flow cytometry (data not shown).

MDSC frequency was evaluated by flow cytometry by staining PBMC using dry pre-formulated antibody panels DuraClone (CD11b-FITC, DRAQ7, HLA-DR-ECD, CD45-KrO, CD14-PC5.5, CD33-PC7, CD80-APC, CD19-APC-alexa750, CD56-APC-alexa750, CD3-APC-alexa750, CD15-Pacific-Blue (Beckman Coulter, Brea,CA,USA) following manufacturer’s procedures. Briefly, 2-5x10^5 PBMC were washed and resuspended in PBS (100 uL) and transferred to the duraclone tube. After 15 min. PBMC were washed and resuspended in 1% paraformaldehyde (PFA). TLR4 expression levels on PMN-MDSC were evaluated by staining PBMCs with different mixtures of monoclonal antibodies: anti-CD45 KrO, anti-CD15 FITC and anti-TLR4 PE (Miltenyi Biotec, Bergisch Gladbach, Germany). After 15 min of incubation, the cells were washed with PBS and fixed with 1% PFA. Acquisition of 100,000 events was performed in the leukocyte-gated population on Cytoflex LX cytometer (Beckman Coulter, Brea, CA, USA) and analyzed by CytExpert software (Beckman Coulter, Brea, CA, USA).

Purified PMN-MDSC (3,5x10⁵) were seeded in poly-D-lysine coated 4-well Chamber Slide (Permanox Slide, Thermo Fisher Scientific, Rochester, New York, NY, USA) in RPMI 1640 supplemented with 2 mmol/L L-glutamine, penicillin/streptomycin solutions, 100X, and 10 mmol/L HEPES buffer.

Cells were stimulated with 10% PRP from patients or HD or recombinant SARS-CoV-2 Spike active trimer (200ng/ml, R&D Systems, USA). Where indicated the TLR4 inhibitor TAK-242 (0.1μM) (Calbiochem) was added 30 minutes before plasma or SARS-CoV-2 Spike protein stimulation. Cells were cultured at 37°C in humidified air with 5% CO₂. After two hours, PMN-MDSC were washed with Hanks’Balanced Salt Solution, and Syto Green solution (25 nM, Life Technologies) was added to stain dsDNA.

After fixation with 4% paraformaldehyde (Sigma-Aldrich, P6148) for 20 min at room temperature the slides were mounted using the ProLong Gold antifade reagent with DAPI (Molecular Probes, Invitrogene, USA).

Image acquisition was carried out using a Leica TCS SP2 confocal microscopy (Leica microsystems). Image processing and analysis were done using Leica Confocal Software (LCS).

Purified PMN-MDSC (2x10⁵ cells/well) were cultured in 96-well flat plates (Corning-Incorporated, New York, NJ, USA), in the above-described medium. Cells were stimulated as mentioned above for 2h at 37°C in humidified air with 5% CO². Extracellular DNA was detected by Pico488 dsDNA quantification kit (Lumiprobe, USA) following the manufacturer’s procedure. Briefly, samples were diluted with the buffer and an equal volume of the Pico488 dye working solution. The final DNA concentration was acquired and analyzed by “Magellan Pro V 7.5” (Tecan Trading AG, Switzerland).

Western blot analyses on cell lysates were performed by washing cells twice with ice-cold PBS (pH 7.4) and lysing them for 30 min on ice with lysis buffer (50 mM Tris pH 7.4, 150 mM NaCl, 0,25% sodium deoxycholate, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 0.5% non-ionic detergent IGEPAL CA-630 (Sigma-Aldrich, Milan, Italy), 1 mM sodium orthovanadate, 20 mM sodium fluoride, 1 µg/mL leupeptin and pepstatin A, 2 μg/mL aprotinin and 1 mM phenylmethylsulfonyl fluoride (PMSF). Whole-cell lysates were centrifuged at 6000× g for 10 min at 4°C. The protein concentration of cell extracts was determined by protein assay dye reagent (Bio-Rad Laboratories). Aliquots of cell extracts containing 30 µg of total proteins were resolved by 8% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred by electroblotting on 0.45 µm pore size nitrocellulose membranes (Amersham™, Merck Life Science S.r.l., Italy). For the immunoassays, membranes were blocked in 3% bovine serum albumin (BSA) fraction V (Biofroxx, Einhausen, Germany) in TTBS/EDTA (10 mM Tris pH 7.4, 100 mM NaCl and 1 mM EDTA, 0.1% Tween 20) for 30 min at room temperature (RT) and then incubated for 1h at RT or overnight at 4°C with specific primary antibodies diluted in 1% BSA/TTBS-EDTA. The antibodies used in immunoblotting were the following: rabbit anti-human ACE2 (Life Span BioSciences), mouse anti-human β-Actin (Santa Cruz Biotechnology), goat Anti-Rabbit antibody (H + L)-HRP conjugate (Bio-Rad Laboratories), and goat anti-mouse antibody (Enzo Life Technologies, Farmingdale, NY, USA).

The immune complexes were detected by enhanced chemiluminescence reaction (ECL Fast Pico; Immunological Sciences, Rome, Italy) and the ChemiDoc XRS (Bio-Rad, Hercules, CA, USA) instrument and the Image Lab software (Bio-Rad) were used to reveal the chemiluminescence signal.

Human Pulmonary Microvascular Endothelial Cells were seeded at 2x10⁴/cm² (ECs, PromoCell Heidelberg, Germany) and cultured in Endothelial Cell Growth medium MV (PromoCell, Heidelberg, Germany), supplemented with SupplementMix (PromoCell, Heidelberg, Germany) in a humidified atmosphere (5% CO₂) at 37°C. The medium was refreshed every 3 days. Once confluence was reached, cells were detached using the Animal Component-Free (ACF) Cell Dissociation Kit (STEMCELL Vancouver, BC, Canada) and seeded on 24-well plates (Corning Incorporated, New York, NY, USA) at a density of 1.1 × 10⁵ cells/mL. When confluence was reached, purified PMN-MDSC were added (3× 10⁵/mL) and cells were stimulated with recombinant SARS-CoV-2 Spike active trimer (200ng/ml, R&D Systems, USA) for 24 hours. ECs viability was evaluated by flow cytometry (Annexin V Apoptosis Detection Kit, eBioscience™).

GraphPad Prism version 9.3.1 for Windows (GraphPad Software) was used to perform statistical analyses. The non-parametric Friedman test with Dunn’s correction or the Mann-Whitney test were used to compare continuous variables. The p-value < 0.05 was considered statistically significant.

Most of the enrolled patients presented bacterial co-infection at the time of enrolment (39 out of 50, 78%), and we wondered whether co-infection could affect PMN-MDSC frequency. Figure 1A shows the gating strategy used to identify circulating MDSC among PBMC. We found that PMN-MDSC frequency was comparable between COVID-19 patients with and without bacterial co-infections (Figure 1B). Moreover, among co-infected patients, we did not find any difference in the PMN-MDSC frequency between patients with and without bacteremia (Figure 1C). These data indicate that, in COVID-19 patients, the presence of bacterial co-infections did not influence PMN-MDSC frequency, suggesting that the main driver of PMN-MDSC expansion was the SARS-CoV-2 infection. Accordingly, the percentage of PMN-MDSC from patients with other co-morbidities did not differ from those without (Figure 1D). We did not find any difference in the PMN-MDSC percentage between males and females (data not shown).

It has been shown that plasma from COVID-19 patients can induce NET by neutrophils, (Middleton et al., 2020) (Kelly et al., 2023) We then evaluated whether PMN-MDSC from COVID-19 patients can release ET upon plasma stimulation. To this aim, purified PMN-MDSC from 16 COVID-19 patients (13 out of 16 with co-infections) were stimulated with 10% of autologous platelets enriched plasma (PRP), and after 2 hours, the release of ET was analyzed by confocal microscopy. We found that PMN-MDSC were able to produce well-organized DNA filaments suggestive of ET upon PRP stimulation (Figure 2A); DNA release was confirmed by treating with DNAse after which DNA filaments were degraded (Figure 2A). Differently, stimulation with PRP from HD was not able to induce ET by PMN-MDSC. In the same experimental conditions, the amount of free DNA in the culture supernatants was quantified by pico488 fluorescence. Figure 2B confirms the results obtained by confocal microscopy, showing an increase in free DNA after stimulation with PRP from COVID-19 patients but not from HD. Red points indicate DNA release from PMN-MDSC isolated from non-co-infected patients. This result suggests that bacterial or fungal co-infection has no major role in inducing ET production.

We then wondered if the capacity to release ET was a peculiar feature of PMN-MDSC from COVID-19 patients. It has been demonstrated that G-CSF stem cell mobilization in human donors induces PMN-MDSC (Luyckx et al., 2012). Mature CD10+ and immature CD10- neutrophils present in G-CSF-treated donors display opposite effects on T cells (Marini et al., 2017; Tumino et al., 2020). Thus, purified PMN-MDSC from G-CSF-mobilized healthy individuals were treated with PRP from patients with COVID-19. Figure 2D shows that PMN-MDSC from HD extrude ET when stimulated with PRP of infected patients, indicating that ET production is not a peculiar feature of PMN-MDSC from SARS-CoV-2 infected patients. To confirm that PRP from COVID-19 patients induced ET formation, PMN-MDSC from G-CSF-mobilized healthy individuals were stimulated with PRP from COVID-19 patients and after 2 hours, the elastase activity associated with ET was evaluated by NETosis assay kit. Supplementary Figure 1 shows that PRP from COVID-19 patients induced the release of ET-associated elastase, confirming the formation of ET by PMN-MDSCs.

In neutrophils, a TLR4-dependent NET formation has been described (Zhang et al., 2021). We wondered whether PRP-induced ET released by PMN-MDSC was TLR-4-mediated. First, we evaluated whether PMN-MDSC from COVID-19 patients express TLR4 and found that this receptor is present on the PMN-MDSC surface (Figure 3A). We then analyzed whether TLR4 inhibition could block ET release by PMN-MDSC. TLR4 inhibition did not affect ET extrusion by PMN-MDSC upon PRP stimulation (Figures 3B, C), indicating that the mechanism driving PMN-MDSC ET release was TLR4-independent.

It has been demonstrated that SARS-CoV-2 Spike protein could be involved in NET extrusion (Youn et al., 2021). We thus evaluated whether the Spike protein could stimulate ET by PMN-MDSC. ET production by purified PMN-MDSC was observed upon Spike stimulation (Figures 4A, B), which was confirmed by DNAse treatment. The release of ET was further confirmed by analyzing the ET-associated elastase using the NETosis assay kit (Supplementary Figure 1). It has been reported that neutrophils express ACE2 and an ACE2-dependent release of NET by neutrophils has been shown during SARS-CoV-2 infection (Veras et al., 2020). We found that PMN-MDSC from COVID-19 patients did not express ACE2 (Figure 4C), excluding the ACE2-mediated ET extrusion. However, Spike protein can bind to other receptors such as TLR4 (Zhao et al., 2021). We then evaluated whether Spike-induced ET by PMN-MDSC was mediated by TLR4. We found that the TLR4 inhibitor TAK-242 inhibited ET release by spike-stimulated PMN-MDSC (Figure 4D), indicating a TLR4-dependent mechanism.

Neutrophils and NET can damage endothelial cells and increase endothelial permeability (Ma et al., 2019). Since the capacity of PMN-MDSC to release ET, we evaluated whether MDSC from COVID-19 patients can induce endothelial cell dysfunction. To this aim, primary human microvascular endothelial cells (EC) were cultured with purified PMN-MDSC from COVID-19 patients and treated with Spike protein to induce ET formation. After 24h, we evaluated EC apoptosis by flow cytometry. Figure 5A shows the EC expression of Annexin V and PI after culture with PMN-MDSC in the indicated conditions. We observed that PMN-MDSC alone were able to induce EC apoptosis, in particular, early apoptosis, being Annexin V+ PI- (Figure 5A). Stimulation with the Spike protein did not affect EC viability compared with PMN-MDSC alone, suggesting that ET are not involved in the PMN-MDSC-induced EC death. Indeed, the treatment with DNAse did not affect the capacity of PMN-MDSC to induce EC death (Figures 5A, B), confirming that PMN-MDSC-induced EC apoptosis was not mediated by ET release.