IONPs suspensions of size 10, 20, and 30 nm were purchased from Ocean Nanotech, LLC (San Diego, CA). A stock combination mixture (1 mg/mL) of all three sizes was prepared in phosphate-buffered saline (PBS) containing Ca²⁺ and Mg²⁺ (Sigma Aldrich, Steinheim, Germany). Compstatin analog Cp40 [yI[CV(MeW)QDW-Sar-AHRC](NMe)I-NH2] was used to inhibit C3 cleavage (18) and PMX53 was used as a C5aR1 antagonist (19). Eritoran (E5564) was kindly provided by Eisai Co., Ltd. (Tokyo, Japan). The C5 inhibitor eculizumab (Soliris^®) was obtained from Alexion Pharmaceuticals (Boston, MA). Anti-CD14 (r18D11), an IgG2/4 chimeric antibody, was produced in our laboratory as previously described (20). Complement inhibitors anti-C2 (mAb, clone 175-62) and anti–factor D (mAb, clone 166-32) were provided by Genentech (San Francisco, CA). Zymosan A from Saccharomyces cerevisiae was purchased from Sigma-Aldrich. Ultrapure lipopolysaccharide (LPS) from Escherichia coli strain 0111:B4 (smooth type) was obtained from InvivoGen (San Diego, CA).

Human whole blood from healthy volunteers was collected in 4.5 mL cryotubes (Cryo Tube™ Vials; Thermo Scientific, Waltham, MA) containing lepirudin (50 mg/mL), a specific thrombin inhibitor (Refludan; Celgene, Uxbridge, UK), as an anticoagulant. Plasma was isolated by whole blood centrifugation at 3000 x g for 15 minutes, and a normal human plasma pool (NHP) was created by pooling plasma from six donors. All samples were stored at -80°C. To prepare platelet-rich plasma (PRP), whole blood was centrifuged at room temperature, 22°C at 180 x g for 10 minutes. PRP was immediately used in the respective applications.

Human lung microvascular endothelial cells (HLMECs; Cell Applications Inc, San Diego, CA) of passage 4-6 were cultured in endothelial cell growth medium (Sigma-Aldrich) in T25 flasks. When the cells reached 80-90% confluence, they were washed and treated with 0.5 ml TrypLE™ (Thermo Fisher Scientific) for 7 minutes. Afterward, 200 000 or 300 000 HLMVEC cells/mL were seeded in 48-well Nunc plates (Thermo Fisher Scientific) pretreated with 0.1% gelatin solution (Sigma-Aldrich) for 5 minutes. The experiments were performed when the cells had reached confluence, typically after 48-72 hours.

NHP (100 μL) was incubated with different-sized IONPs (10, 20, or 30 nm) in increasing concentrations (2, 20, and 200 μg/mL) or PBS. Zymosan (100 µg/mL) was used as a positive control. The samples were incubated for 30 minutes at 37°C in 1.8 mL cryotubes (Thermo Scientific). Ethylenediamine tetra-acetic acid (EDTA; 20 mM in final concentration) was added to prevent further activation. Finally, all samples were centrifuged at 3000 x g for 15 min at 4°C to remove the IONPs. In separate experiments, the same incubation conditions were used, however, a mixture of IONPs 10-30 nm was prepared and added in NHP at the following concentrations: 5, 10, 50, and 100 µg/mL. In this setup, we also included two inhibitors, Cp40 (20 μM) and eculizumab (100 μg/mL), in samples containing 100 µg/mL of IONPs. In another set of experiments, IONPs (10-30 nm, 100 μg/mL) or PBS were in separate experiments incubated in NHP with or without the anti-C2 mAb (72 μg/mL) and/or anti-factor D mAb (36 μg/mL), or 10 mM Mg²⁺/EGTA, or 10 mM EDTA for 30 minutes at 37°C. EDTA (20 mM final concentration) was added after incubation to prevent further activation. All samples were centrifuged at 3000 x g for 15 min at 4°C.

Whole blood experiments were performed as previously described (21). The addition of 300 μL of blood in 1.8 mL cryotubes was followed by pre-incubation with each of the following inhibitors; Cp40 (20 μM), eculizumab (100 μg/mL), anti-CD14 (15 μg/mL), PMX53 (10 μg/mL) and eritoran (1 μM), or PBS for 5 minutes at 37°C. A mixture of IONPs 10-30 nm (100 µg/mL) or PBS was added, and the samples were incubated for four hours at 37°C. Zymosan (100 µg/mL) and LPS (10 ng/mL), in combination, were used as a positive control for complement activation and cellular activation. After incubation, the tubes were placed on ice, and EDTA (20 mM final concentration) was added. The samples were centrifuged at 3000 x g for 15 min, and plasma was immediately isolated and stored in aliquots.

At 70-90% confluence, the HLMECs were washed with tempered (37°C) PBS to remove cell debris and incubated with whole blood (150 µL) for four hours at 37°C with gentle shaking (50 rpm). For this procedure, whole blood was pre-incubated with the inhibitors described above, followed by the addition of the 10-30 IONPs mixture (100 µg/mL) or PBS. Before the addition, the IONPs were sonicated (UP200S Hielscher Ultrasonics, Teltow, Germany) for 8 minutes at 70% amplitude and 0.6 pulses. A combination of zymosan (100 µg/mL) and LPS (10 ng/mL) was used as a positive control. Separate wells were incubated in cell culture media (negative control) with and without TNF and IL-1β in combination (both at 10 ng/mL; Sigma-Aldrich) to evaluate endothelial cells’ activation. After incubation, the whole blood was removed, and EDTA (20 mM final concentration) was added. The plasma was isolated and stored at -80°C until analysis.

Complement activation markers TCC (soluble C5b-9) and C3bc were quantified by enzyme-linked immunosorbent assays (ELISA), as previously described in detail (22–24). Platelets’ soluble activation markers, platelet factor 4 (PF4), P-selectin (sCD62P), and neutrophil-activating peptide-2 (NAP-2) were measured by ELISAs purchased from R&D systems (Minneapolis, MN). Lactate dehydrogenase (LDH), a marker for cell toxicity, was analyzed by ELISA acquired from Abcam (Cambridge, UK). Cytokines analysis was performed using a 4-plex kit (R&D systems) for whole blood samples or a 27-plex kit (Bio-Rad, Hercules, CA) for whole blood/endothelial cells samples. Both kits included the following cytokines; Tissue Necrotic Factor (TNF), interleukin (IL)-1ß, IL-6, and IL-8, and were performed according to the instructions from the manufacturers. IL-8 was, additionally, measured by ELISA for whole blood samples (R&Ds systems).

Hemolysis was analyzed by recording the absorbance spectra in the 350 – 700 nm wavelength range using Nanodrop™ (Thermo Scientific). The data (λ = 414 nm) refer to the Soret peak of heme.

The data were analyzed using GraphPad (San Diego, CA) Prism version 9 for Mac by ordinary or repeated-measures one-way ANOVA with Dunnett multiple-comparison posttest for the comparison of multiple columns and paired t-tests for the comparison of two columns. A p-value < 0.05 was considered statistically significant.

The study was performed according to the ethical guidelines from the declaration of Helsinki. Sampling of human whole blood from healthy individuals for the preparation of plasma was approved by the ethical committee of the Norwegian Regional Health Authority, ethical permit REK#S-04114, 2010/934. Informed written consent was obtained from each donor.

To evaluate whether IONPs promote complement activation, lepirudin-anticoagulated NHP was incubated with increasing concentrations of IONPs (2-200 μg/mL) of different sizes, 10, 20, or 30 nm, for 30 minutes, and C3bc plasma concentrations were measured. The highest concentration of IONPs, 200 μg/ml, led to significant complement activation (p<0.05-0.01) independently of the size of IONPs ( Figures 1A–C ). IONPs of 2 and 20 μg/mL caused significant complement activation (p<0.05-0.01) only when incubated with 10 nm- or 20 nm- sized IONPs ( Figures 1A, B ). Overall, the different sizes of IONPs activated complement to a similar degree and were in the following set of experiments combined in one mixture. In the next step, IONPs in increasing concentrations of 5-100 μg/mL were incubated with NHP. Two complement inhibitors were included, Cp40 and eculizumab, that block C3- and C5-cleavage, respectively, together with IONPs (100 μg/mL), and sC5b-9 formation was measured additionally to C3bc ( Figures 1D, E ). IONPs increased C3bc and sC5b-9 in a dose-dependent manner, and the most profound effect was observed for the highest concentration of IONPs (100 μg/mL) ( Figures 1D, E ). As expected, both inhibitors abolished sC5b-9 formation while Cp40, but not eculizumab, inhibited C3bc.

To investigate which complement activation pathway that is activated by IONPs, an anti-C2 antibody, inhibiting classical and lectin pathways, and an anti-factor D antibody, inhibiting the alternative pathway, were incubated with IONPs (100 μg/mL) for 30 minutes. Both antibodies inhibited activation to a similar degree ( Figures 1F, G ). After subtracting the levels of C3bc and sC5b-9 in non-incubated plasma (T0), C3bc was reduced by 84% and 81% by the anti-C2 and the anti-factor D respectively ( Figure 1F ). Similarly, sC5b-9 was decreased by 59% by the anti-C2 and by 61% by the anti-factor D ( Figure 1G ). Anti-C2 and anti-factor D in combination completely blocked C3- and terminal pathway activation. Furthermore, addition of 10 mM EGTA, which in the presence of 10 mM Mg²⁺ specifically blocks classical and lectin pathway activation by chelating Ca²⁺, reduced C3bc by 50% and sC5b-9 by 80%.

Human whole blood was incubated with IONPs (100 μg/mL) of combined sizes (10, 20, and 30 nm) for four hours. IONPs significantly increased C3bc levels (p<0.01) by 105% ( Figure 2A ), however, they did not induce a significant increase in sC5b-9 formation ( Figure 2B ). Cp40 resulted in a significant decrease of both C3bc (p<0.001) and sC5b-9 (p<0.01), while eculizumab resulted in a significant reduction of sC5b-9 formation (p<0.01). Anti-CD14, eritoran, and PMX53 did not significantly influence the level of C3bc or sC5b-9.

IONPs led to a robust release of pro-inflammatory cytokines; TNF (p<0.01), IL-1β (p<0.01), IL-6 (p<0.01), and IL-8 (p<0.001) were all significantly increased by IONPs ( Figures 2C–F ). Inhibition with Cp40, eculizumab, and PMX53 significantly reduced IL-8 ( Figure 2C ) by 38% (p<0.01), 42% (p<0.05), and 33% (p<0.05), respectively. Inhibition with Cp40 significantly decreased TNF ( Figure 2D ) production by 22% (p<0.05). Eculizumab lowered TNF levels, but this reduction was not statistically significant (p=0.059). The release of IL-1β ( Figure 2E ) and IL-6 ( Figure 2F ) was also lower in all three complement inhibitor samples, but none of these were statistically significant. Blockade of TLR-signaling with a-CD14 or eritoran did not significantly influence the concentration of any of the analyzed cytokines.

We further studied the inflammatory response induced by IONPs in the human whole blood model, but this time in combination with HLMECs to include the whole blood and endothelial cells interaction. After four hours of incubation, IONPs-induced complement activation, as measured by C3bc ( Figure 3A ) and sC5b-9 ( Figure 3B ), was significantly elevated (p<0.05). Inhibition with Cp40 significantly attenuated both C3bc (p<0.001) and sC5b-9 levels (p<0.01). Eculizumab completely blocked sC5b-9 formation (p<0.01), while PMX53 showed no effect, as expected.

Cytokine release was induced by IONPs in human whole blood accompanied by endothelial cells, as demonstrated by a significant increase in the production of most cytokines, including TNF, IL-1β, IL-6, IL-8, and IL-2, as well as production of IFN-γ. Overall, 21 out of the 27 analyzed markers were significantly increased due to the presence of IONPs ( Figures 3C–N ; Supplementary Figure 1 ). Interestingly, IP-10 and eotaxin were significantly reduced by the IONPs ( Supplementary Figure 1 ). Cp40, eculizumab, and PMX53 significantly attenuated the release of IL-8, IL-2, IL-5, and MIP-1α ( Figure 3 ). In addition, Cp40 and eculizumab significantly also reduced TNF, IL-17A, IL-13 and IL-12, Cp40 alone reduced IFN-γ, VEGF and FGF formation while eculizumab alone significantly decreased IL-6, as well as GM-CSF and MCP-1 ( Figure 3 ; Supplementary Figure 1 ). Additionally, PMX53 showed a significant attenuating effect on FGF, VEGF, and GM-CSF ( Figure 3 ; Supplementary Figure 1 ). No inhibitory effects were observed on IL-1β and IL-1Ra ( Figure 3 ; Supplementary Figure 1 ).

To evaluate the effect of IONPs on endothelial cells’ activation, ICAM-1 and CD62P/E were measured by flow cytometry after four hours of incubation in the presence of IONPs. Expression of both ICAM-1 ( Figure 4A ) and CD62P/E ( Figure 4B ) was significantly elevated (p<0.0.0001 and p<0.05, respectively). Inhibition by Cp40 significantly attenuated ICAM-1 by 21% (p<0.01) and CD62P/E by 33% (p<0.05), respectively. Eculizumab significantly reduced only CD62P/E expression (p<0.05), while PMX53 did not influence endothelial cell activation. Monocytes and granulocytes were isolated from the whole blood after the first 15 minutes of incubation, and the cell surface activation marker CD11b was evaluated by flow cytometry. IONPs induced a substantial CD11b increase (p<0.001) on both monocytes and granulocytes. Monocyte activation was significantly reduced by Cp40 (p<0.05) and PMX53 (p<0.05), but not eculizumab (p=0.051) ( Figure 4C ). Complement inhibition did not significantly attenuate granulocyte activation ( Figure 4D ).

Platelets play an important role in the thromboinflammatory response. Therefore, we evaluated how IONPs affect their activation in whole blood. Soluble platelet activation markers, sCD62P, PF4, and NAP-2, were analyzed in plasma isolated from the whole blood after incubation on endothelial cells for four hours. IONPs significantly increased all three markers (p<0.05-0.001), and no inhibitory effect was observed due to complement inhibition, as measured by C5-, C3-, or C5aR1-blockade ( Figures 4E–G ). In a separate set of experiments, PRP was incubated for 15 minutes with IONPs. Platelet activation, evaluated by measuring surface activation markers CD62P and CD63 by flow cytometry, was increased by IONPs (p<0.05), but it was not significantly reduced by the presence of the complement inhibitors ( Figures 4H, I ). Finally, we measured platelet C3-fragment deposition and found increased values in the presence of IONPs, however, the difference was not statistically significant ( Figure 4J ), most likely due to the biological variation amongst donors, since this increase was significantly (p<0.05) reduced by the C3 inhibitor Cp40,

Since we visually observed increased hemolysis in the samples containing IONPs after incubation in whole blood on endothelial cells, we decided to measure LDH as a marker of cell damage and the total amount of heme as a marker of hemolysis. IONPs significantly elevated LDH release (p<0.001) ( Figure 5A ) and heme concentration ( Figure 5B ) (p<0.0001) in comparison to the control. A positive correlation (r = 0.94) was found between LDH and heme levels ( Figure 5C ). Eculizumab significantly decreased the LDH level ( Figure 5A ) (p<0.05). No significant differences were detected for any of the other inhibitors.