Inhibition of NETosis by a Nuclear-Penetrating Anti-DNA Autoantibody

Cell-penetrating variants of autoantibodies offer new strategies for delivery of cargo to live cells and for engaging previously undruggable intracellular targets (1, 2). Derivatives of cell-penetrating autoantibodies have progressed to clinical testing in genetic disease and are in preparation for studies in ischemic conditions (3–5). In oncology, nuclear-localizing anti-DNA autoantibodies that interfere with genomic integrity and repair are candidates for development as agents to disrupt the DNA damage response in vulnerable tumor cells (2, 6–16). The recognition that some anti-DNA autoantibodies impact DNA repair in live cells has raised new questions on the roles of anti-DNA autoantibodies in the regulation of immunity (2).

Activated neutrophils form neutrophil extracellular traps (NETs) that contribute to the immune response and the pathophysiology of inflammatory and autoimmune diseases (17, 18). Chromatin decondensation and citrullination of histones are critical events in NETosis, and the importance of the DNA damage response in mechanisms of NETosis was recently reported (19). The effects of nuclear-penetrating anti-DNA autoantibodies that interfere with DNA repair on NET formation and release of neutrophil DNA are unknown, and we hypothesize that such autoantibodies inhibit NETosis.

The 3E10 nuclear-penetrating lupus autoantibody localizes to DNA at damaged tissues and tumors, inhibits DNA repair, increases cancer cell sensitivity to DNA-damaging agents, and as a single agent is synthetically lethal to tumors with defects in homologous recombination (8). An optimized di-scFv fragment of 3E10 (Deoxymab-1 [DX1]) with Fc removed to reduce risks of off-target effects is in preparation for clinical trial testing against DNA repair–deficient tumors (2, 8–10). In the present work we evaluated the effect of DX1 on chromatin decondensation and DNA release as surrogates of NETosis in differentiated PLB-985 cells stimulated with PMA (20–22), as well as the impact of DX1 on NETosis in mouse neutrophils stimulated with PMA or ionomycin (IM). Our findings demonstrate significant suppression of NETosis by the DX1 Ab, which has implications both for potential therapeutic applications of DX1 and for understanding the interplay between anti-DNA autoantibodies, neutrophils, and the immune response.

The 3E10 derivative DX1 (PAT-DX1, Patrys, Melbourne, VIC, Australia) was generated and purified as previously described (10). GSK484 (no. SML1658) was obtained from MilliporeSigma (St. Louis, MO). PLB-985 cells, a subclone of HL-60 human acute myeloid leukemia cells, were obtained by a material transfer agreement with the Leibniz Institute DSMZ–German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany) and used within 1 mo of receipt. Cells were grown in RPMI 1640 + 10% FBS at 37°C with 5% CO₂ and differentiated to granulocyte-like cells by the addition of 1.3% DMSO to the growth medium for 6 d as previously described (23). Unless otherwise specified, all other materials and reagents were obtained from Thermo Fisher Scientific (Waltham, MA).

PLB-985 cells cultured on Shi-fix coverslips (Shikhar Biotech, Khumaltar, Lalitpur, Nepal) were treated for 1 h with control media or media containing 5 μM DX1 and then washed and fixed in 100% chilled ethanol. The presence of DX1 in the cells was evaluated by incubation with Pierce recombinant protein L (1:1000) as previously described (10) (21189, Thermo Fisher Scientific), protein L polyclonal Ab (1:1000) (PA1-72066, Thermo Fisher Scientific), and goat anti-chicken IgY (H+L) alkaline phosphatase secondary Ab (1:1000) (PA1-28799, Thermo Fisher Scientific). Alkaline phosphatase–mediated signal development was visualized under bright-field microscopy (EVOS FL, Thermo Fisher Scientific) as previously described (6).

Differentiated PLB-985 cells were plated on Shi-fix coverslips in 12-well plates at 5 × 10⁴ cells/well. Cells were treated with control media or media containing 10 μM DX1 for 1 h, after which cells were stimulated by the addition of 300 nM PMA for 2 h. Cells were fixed with ethanol followed by addition of DAPI and visualization of chromatin decondensation by fluorescence microscopy (EVOS FL, Thermo Fisher Scientific). The percentage of cells that demonstrated the expected response to PMA stimulation was measured by counting the number of total cells and cells with visualized DAPI signal decondensation in a minimum of 150 cells per treatment condition.

Differentiated PLB-985 cells were cultured in black wall 96-well plates at 5 × 10⁴ cells/well and incubated with control buffer or 10 μM DX1 for 1 h. Cells were then stimulated by the addition of 300 nM PMA for 2 h. SYTOX Green (5 μM) was added to facilitate detection of extracellular DNA. SYTOX Green fluorescence was visualized by fluorescence microscopy (EVOS FL, Thermo Fisher Scientific) and fluorescence intensity was measured by a plate reader (Synergy HT, BioTek Instruments, Winooski, VT), at excitation and emission wavelengths of 485 and 527 nm, respectively.

Differentiated PLB-985 cells were plated at 8 × 10⁵ cells/well in 12-well plates and incubated with control buffer or buffer containing 10 μM DX1 for 1 h. Cells were then stimulated by the addition of 300 nM PMA for 4 h (stimulation duration was increased to 4 h compared with DAPI and SYTOX Green stain and Western blot experiments to allow more time for adherence of DNA–protein complexes to wells). Supernatants were aspirated and wells were gently washed once with PBS. Residual contents adherent to the wells were extracted into PBS, and DNA content was determined by spectrophotometry using the NanoDrop Lite (Thermo Fisher Scientific) in accordance with the manufacturer’s instructions.

Differentiated PLB-985 cells at 8 × 10⁵ cells/well in 12-well plates were incubated with control buffer or buffer containing 10 μM DX1 for 1 h and then treated with control DMSO or stimulated by addition of 300 nM PMA for 2 h. Cells were isolated by centrifugation, lysed in RIPA buffer, and H3Cit content was determined by Western blot using anti-H3Cit primary (AB5103, Abcam, Cambridge, MA) and goat anti-rabbit HRP-conjugated secondary Ab at 1:5000 (AB205718, Abcam). H3Cit band intensities normalized to actin loading control were determined by ImageJ (National Institutes of Health, Bethesda, MD).

All mouse work was conducted under a protocol approved by Monash University’s Animal Ethics Committee. Neutrophils were retrieved from peritoneal cavities of 4% thioglycollate-stimulated male C57BL/6 mice aged 10 wk (n = 8) as previously described (24). Neutrophils plated at 2 × 10⁵ cells/well in 20-well plates were allowed to settle for 30 min and then treated with control media or media containing 10 μM DX1 for 30 min. Neutrophils were then left unstimulated or NETosis was stimulated by addition of PMA (40 μg/ml) or IM (4 μM, Stem Cell Technologies, VIC, Australia) for 4 h. Cells were fixed in periodate-lysine-paraformaldehyde (PLP) overnight, immunostained for markers of NETs including myeloperoxidase (MPO, AF3667, R&D Systems, Minneapolis, MN), H3Cit, and peptidylarginine deiminase 4 (PAD4; AB128086, Abcam), mounted in DAPI ProLong Gold (Molecular Probes, Thermo Fisher Scientific), and NETs were visualized by confocal fluorescence microscopy with a Nikon Ti-E inverted microscope (Nikon Instruments, Melville, NY). Then, 405-, 488-, 561-, and 647-nm lasers were used to specifically excite DAPI, Alexa Fluor 488, Alexa Fluor 594, and Alexa Fluor 647. The percentage of NETs formed in each treatment condition was determined using the ImageJ (National Institutes of Health) analysis feature “cell counter plugin” to count the total number of DAPI-positive cells, as well as the number of H3Cit, MPO, and PAD4 triple-positive cells. The number of H3Cit, MPO, and PAD4 triple-positive cells was divided by the number of DAPI-positive cells to give the percentage of cells identified as NETs.

Statistical analyses were performed in GraphPad Prism version 9.2.0. The p values were determined by a two-tailed Student t test when comparing two groups and by a one-way ANOVA with Tukey’s multiple comparisons test for multiple groups. Significance was considered as p ≤ 0.05. Error bars in the figures represent SEM. The numbers of replicates are indicated in the figures.

Work with isolated mouse or human neutrophils is challenging due to their short functional lifespans, and granulocyte-differentiated HL-60 and their subclone PLB-985 human acute myeloid leukemia cells provide a reliable alternative to isolated neutrophils for NETosis studies (20–23, 25). The nuclear-penetrating anti-DNA autoantibody 3E10 and its derivative DX1 require expression of the nucleoside transporter ENT2 to penetrate target cells. ENT2 expression in HL-60 cells has previously been reported (26), and DX1 was confirmed to penetrate into ∼100% of PLB-985 cells in culture as expected (Fig. 1). We proceeded with evaluating the impact of DX1 on markers of NETosis in these cells.

Neutrophils treated with PMA undergo a cascade triggered by NADPH oxidase (NOX)–mediated production of reactive oxygen species that ultimately results in the release of NETs through NOX-dependent NETosis (27). Differentiated PLB-985/HL-60 cells mirror the activity of neutrophils and respond to PMA stimulation by decondensing DNA and releasing NETs, and their chromatin decondensation can be visualized by DAPI staining (28). Differentiated PLB-985 cells were treated with control media or media containing 10 μM DX1 for 1 h, followed by addition of PMA and subsequent visualization of chromatin status by DAPI stain. The diffuse DAPI stain suggestive of chromatin decondensation was observed in 67.5 ± 1.6% of control cells and 37.8 ± 2.7% of cells pretreated with DX1 (n = 2) (Fig. 2A, 2B).

SYTOX Green is an impermeant fluorescent DNA stain that facilitates visualization of DNA released in NETosis (27, 29). Differentiated PLB-985 cells treated with control or 10 μM DX1 and stimulated with PMA were exposed to SYTOX Green, and fluorescence was visualized under fluorescence microscopy and quantified by a plate reader. DX1 significantly reduced the resulting SYTOX Green signal, with fluorescence reduced to 0.40 ± 0.01 relative to control (p < 0.0001, n = 3) (Fig. 2C, 2D). This result is consistent with inhibition of DNA release mediated by DX1.

The results of the DAPI and SYTOX Green staining studies in (Fig. 2 suggest DX1-mediated inhibition of chromatin decondensation and DNA release by differentiated PLB-985 cells. To further probe this effect, we measured DNA–protein release by differentiated PLB-985 cells after PMA stimulation by resuspending and quantifying DNA–protein complexes adherent to wells by spectrophotometry using a previously described protocol for isolation and quantification of NET-DNA in cell culture (30). This assay is based on the concept that intact cells for the most part remain in suspension, whereas DNA–protein complexes released in NETs are at least partially adherent to wells. This distinction allows for separation of intact cells from adherent NET-DNA by aspiration of the cell culture supernatants, followed by resuspension of adherent DNA–protein complexes and quantification by spectrophotometry.

A control study was conducted to confirm differential response of the undifferentiated and differentiated PLB-985 cells to PMA in this assay. Undifferentiated and differentiated PLB-985 cells were treated with control or 300 nM PMA followed by quantification of the expected NET-DNA adherent to wells by spectrophotometry. Control and PMA-stimulated undifferentiated PLB-985 cells yielded similar DNA concentrations of 0.73 ± 0.03 and 0.93 ± 0.12 ng/μl (p = 0.12, n = 3), respectively, consistent with the expected absence of response by undifferentiated cells to PMA. In contrast, control and PMA-stimulated differentiated PLB-985 cells yielded DNA concentrations of 0.83 ± 0.03 and 4.40 ± 0.27 ng/μl (p = 0.002, n = 3), consistent with the expected release of NET-DNA in response to PMA (Fig. 3A). These findings support the conclusions that differentiated PLB-985 cells exhibited granulocyte-like behavior with response to PMA measurable by the reported spectrophotometry method (30). Furthermore, the absence of an apparent increase in DNA release by undifferentiated PLB-985 cells treated with PMA rules out the possibility that any unexpected nonspecific PMA toxicity made a major contribution to the measured DNA concentrations.

We next verified the ability of this assay to detect the effect of a known NETosis inhibitor on DNA release by differentiated PLB-985 cells stimulated by PMA. The activity of PAD4 in catalyzing histone citrullination is essential to granulocyte NETosis, and the small molecule selective PAD4 inhibitor GSK484 has previously been shown to inhibit NET release (31, 32). Differentiated PLB-985 cells were treated with 0–50 μM GSK484 for 1 h prior to stimulation with PMA, followed by quantification of adherent DNA by spectrophotometry. GSK484 reduced the measured DNA in a dose-dependent manner, consistent with suppression of NETosis in differentiated PLB-985 cells (Fig. 3B).

The differential response of undifferentiated and differentiated PLB-985 cells to PMA and the observed inhibition of response to PMA mediated by the known PAD4 inhibitor GSK484 supported the use of the DNA spectrophotometry assay in evaluating the effect of DX1 on NETosis in differentiated PLB-985 cells. Differentiated PLB-985 cells were treated with media containing control buffer or 10 μM DX1 for 1 h followed by stimulation by addition of PMA and subsequent quantification of DNA release by spectrophotometry as described above. Results were expressed relative to DNA content obtained in unstimulated cells treated with control media. PMA stimulation increased relative DNA content to 4.50 ± 0.14 (p < 0.0001, n = 3), whereas DX1 suppressed the increase in DNA release to 1.80 ± 0.11 (p < 0.0001, n = 3) (Fig. 4). These results suggest inhibition of NETosis by DX1 and are consistent with the findings of the DAPI and SYTOX Green assays.

As discussed above, histone citrullination by PAD4 is a key step in NETosis that is targeted by the selective PAD4 inhibitor GSK484. Differentiated PLB-985 cells treated with control or 10 μM DX1 for 1 h were stimulated by PMA, and cell contents were isolated and analyzed for H3Cit content by Western blot with normalization to actin. Stimulation of control cells with PMA was associated with a reduction in H3Cit content to 0.64 ± 0.06 relative to unstimulated cells, consistent with the expected release of H3Cit in the process of NET formation. In contrast, cells treated with DX1 exhibited an apparent increase in H3Cit content to 1.46 ± 0.20 relative to unstimulated cells (Fig. 5), possibly indicating retention of H3Cit within the cells due to inhibition of NET release by DX1. These data suggest that DX1 does not interfere with citrullination of H3 but rather inhibits the release of NETs and their associated H3Cit content.

Neutrophils were obtained from thioglycollate-treated mice and directly plated onto 20-well plates. Cells were treated with control media or media containing 10 μM DX1 for 30 min, and then left unstimulated (Fig. 6) or stimulated with PMA for induction of NOX-dependent NETosis (Fig. 7) or a calcium ionophore (IM) for induction of NOX-independent NETosis (Fig. 8) (33) for 4 h. Cells were then fixed and immunostained for markers of NETs, including MPO, H3Cit, and PAD4. The percentage of NETs released per 100 cells was determined by visualization on confocal fluorescence microscopy and analyzed by ImageJ (National Institutes of Health). As expected, few NETs were observed in control neutrophils (unexposed to PMA or IM) treated with control or DX1 (3.5 ± 1.0% and 4.0 ± 1.5%, respectively, p = 0.79, n = 4) (Fig. 6), indicating that DX1 does not stimulate NETosis. PMA stimulation yielded NETs in 36.3 ± 11.1% of control neutrophils, but only 8.3 ± 2.7% of DX1-treated neutrophils (p = 0.05, n = 4) (Fig. 7). Similarly, IM stimulation resulted in NETs in 15.5 ± 3.1% of control neutrophils, but only 2.3 ± 0.8% in DX1-treated neutrophils (p < 0.01, n = 4) (Fig. 8). These results demonstrate inhibition of both NOX-dependent and independent mechanisms of NETosis mediated by PMA and IM, respectively, in mouse neutrophils.

Theories on the origins, functions, and relevance of anti-DNA autoantibodies in human disease have evolved greatly over the many years that have passed since these Abs were first identified. DNA-targeting Abs were initially believed incapable of crossing cell membranes and therefore limited to interacting with DNA released by dead or damaged cells. The recognition that some anti-DNA autoantibodies penetrate live cells and localize into nuclei to engage DNA targets and cause meaningful changes in cellular functions such as DNA repair has altered perspectives on the significance of these Abs and their potential to be modulated to ameliorate autoimmune disease or harnessed for therapeutic applications (2, 34).

The 3E10 lupus anti-DNA autoantibody penetrates live cell nuclei without adversely impacting normal cell survival. The mechanism of nuclear penetration by 3E10 is dependent on the presence of local extracellular DNA/nucleosides and cellular expression of the nucleoside transporter ENT2 (6, 7). This mechanism provides the basis for strategies that use 3E10 or its derivatives to target cells in regions enriched in extracellular DNA, such as necrotic tumors or damaged tissues (4, 5, 14). Early efforts focused on use of 3E10 to deliver linked cargo proteins into cells (3–5, 15, 16), and more recent studies have taken advantage of the inhibitory effects of 3E10 on DNA repair in developing new approaches to targeting tumors with defects in homologous recombination (8–12). In the present work we found that DX1, a derivative of 3E10, inhibits both NOX-dependent and independent NETosis. Dual NETosis pathway inhibition by DX1 distinguishes it from other inhibitors of NET formation, such as drugs that target PAD4 (35). These findings add to our understanding of and raise new questions on the interplay between nuclear-penetrating anti-DNA autoantibodies and neutrophil biology.

NETs facilitate innate immunity and contribute to multiple disease processes, including inflammatory diseases such as acute respiratory distress syndrome and COVID-19, thrombotic disorders, autoimmunity, cystic fibrosis, and potentially to malignancy (30, 36–39). Mechanisms of NET-mediated pathophysiology range from damaging effects on epithelial and endothelial cells to promotion of inflammation and autoimmune response through DNA and autoantigen release (18, 40–42). Methods to modulate NET formation are of clinical interest, and systemic administration of DNase I or inhibitors of histone citrullination to promote dissolution of NET-DNA or prevent NET formation has been reported (43, 44). DX1 is in preparation for use against tumors with defects in DNA repair, and the new recognition that DX1 inhibits both NOX-dependent and independent NET release raises the possibility that DX1 may be of benefit in nonmalignant conditions in which suppression of NETosis is needed.

The importance of DNA repair mechanisms in chromatin decondensation in NETosis was recently reported (19), and 3E10 is known to inhibit base excision repair and homologous recombination. It is possible that nuclear-penetrating Abs that interfere with the DNA damage response, such as 3E10/DX1, may suppress NETosis through their inhibition of DNA repair. Additional studies are required to fully elucidate the mechanism by which DX1 blocks NET formation and to evaluate the impact of DX1 and other nuclear-localizing Abs, such as the nucleolytic variants (45), on neutrophil function and NETosis in vivo.

J.E.H. has related intellectual property (IP) and equity interest in and receives research support and consulting fees in the form of equity options from Patrys Ltd., a biotechnology company that has licensed IP that is the basis of this work. J.Z. receives consulting fees in the form of equity options from Patrys Ltd. V.D. and J.A.C. are employees of and have equity interest in Patrys Ltd. X.C., B.J.C., V.D., A.S., J.Z., J.A.C., and J.E.H. are inventors on IP related to DX1 and/or use of DX1 as an inhibitor of NETosis. The other authors have no financial conflicts of interest.