Transcriptional firing helps to drive NETosis

Abstract

Neutrophils are short-lived innate immune cells. These cells respond quickly to stimuli, and die within minutes to hours; the relevance of DNA transcription in dying neutrophils remains an enigma for several decades. Here we show that the transcriptional activity reflects the degree of DNA decondensation occurring in both NADPH oxidase 2 (Nox)-dependent and Nox-independent neutrophil extracellular trap (NET) formation or NETosis. Transcriptomics analyses show that transcription starts at multiple loci in all chromosomes earlier in the rapid Nox-independent NETosis (induced by calcium ionophore A23187) than Nox-dependent NETosis (induced by PMA). NETosis-specific kinase cascades differentially activate transcription of different sets of genes. Inhibitors of transcription, but not translation, suppress both types of NETosis. In particular, promoter melting step is important to drive NETosis (induced by PMA, E. coli LPS, A23187, Streptomyces conglobatus ionomycin). Extensive citrullination of histones in multiple loci occurs only during calcium-mediated NETosis, suggesting that citrullination of histone contributes to the rapid DNA decondensation seen in Nox-independent NETosis. Furthermore, blocking transcription suppresses both types of NETosis, without affecting the reactive oxygen species production that is necessary for antimicrobial functions. Therefore, we assign a new function for transcription in neutrophils: Transcriptional firing, regulated by NETosis-specific kinases, helps to drive NETosis.

Figure 1. Degree of transcription reflects the extent of nuclear decondensation during NETosis.

Transcriptomics was conducted for control, PMA and A23187 conditions at 30- and 60-min time points (n = 3 for each condition). (A) Principal component analyses of the transcriptomics data set show that the expression profiles of all 6 conditions are different from each other. Each small ball represents one replicate. (B) Frequency distribution of transcripts shows that different degree of transcription occurs under all four different conditions (Chi-square test, p < 0.05). Note that the magnitude of frequencies reflects the level of transcription in each condition. (C) Hierarchical cluster analyses of the transcriptomics data show the relatedness of each data set. See Supplementary Fig. S1 for nuclear decondensation during NETosis.

Figure 2. Genome-wide transcription occurs during both Nox-dependent and -independent NETosis.

(A–D) Locations of transcripts, that show 1.5-fold increase, are indicated with red tick marks along each chromosome. Loci of both coding and non-coding transcripts are shown. (E–H) Venn diagrams showing the numbers of common and unique transcripts detected during PMA- and A23187-mediated NETosis at 30- and 60-min time points. See Supplementary Fig. S3 for percentages.

Figure 3. Different sets of coding genes are expressed differentially during Nox-dependent and -independent NETosis.

(A–D) Probability distributions of highly expressed coding genes (1.5-fold difference). (E–H) Venn diagrams show the numbers of common and unique coding genes with p < 0.05 that are transcribed in PMA mediated and A23187-mediated NETosis at 30- and 60-min time points. See Supplementary Fig. S4 for percentages and Table S2 for a complete gene list.

Figure 4. Transcription network analysis indicates that different sets of kinases could drive the transcription during both types of NETosis.

(A–D) Network analysis at 30- and 60-min time points for both PMA-mediated and A23187-mediated NETosis showing the activation of different transcription factors by various kinase cascades. Transcripts used in this analysis include all the coding genes that showed 1.5-fold increase with a p-value of <0.05 shown in Fig. 3 and Table S2. Values at the bottom of the network maps represent the numbers of transcripts predicted to be transcribed by phosphorylation of specific kinase cascades and transcription factors. See Table S3 for transcription factors and number of regulated genes.

Figure 5. Sytox green assays show that inhibition of initial steps of transcription suppresses both types of NETosis.

(A) Effect of G-C rich promoter melting inhibitor Actinomycin D (0, 0.625, 1.25, 2.5, 5.0 μg/ml) on PMA-mediated NETosis. (B) As of (A) but for A23187-mediated NETosis. (C) Effect of DNA topoisomerase 1 inhibitor Camptothecin 11 (0, 0.625, 1.25, 2.5, 5.0 nM) on PMA-mediated NETosis. (D) As of (C) but for A23187-mediated NETosis. (E) Effect of mRNA elongation inhibitor, that prevents the phosphorylation of CDK 7 and 9 to limit RNA polymerase movement on DNA (0, 25, 50, 100, 200 μM DRB) on PMA-mediated NETosis. (F) As of (E) but for A23187-mediated NETosis. *Indicates that NETosis with no inhibitor significantly differs from the rest of the conditions (p < 0.05; One-Way ANOVA with Dunnett’s post-test conducted at each time points; n = 4). See Supplementary Fig. S5 for additional information.

Figure 6. Confocal microscopy images show that transcription inhibitor Actinomycin D suppresses PMA-mediated and A23187-mediated NETosis.

(A) Unstimulated control neutrophils show typical polymorphonuclear morphology. Myeloperoxidase (MPO) is visible in the cytoplasm. MPO co-localizes to NET DNA generated by PMA-mediated NETosis. A limited amount of citrullinated histone 3 (CitH3) immunostaining is visible on some of the NETs. By contrast, intense immunostaining of CitH3 is visible on decondensed nuclei during A23187-induced NETosis; nicely dispersed immunolabeling of CitH3 is visible on the NETs released from these neutrophils. MPO colocalizes to NETs. (B) As of (A) but with Actinomycin D (ActD, 5 μg/ml). PMA- and A23187-treated neutrophils did not show NETosis, and the nuclear morphology of these cells remains the same as that of the unstimulated control neutrophils. Nuclear morphology of these cells also do not show signs of substantial apoptosis. Only a limited amount of CitH3 staining is detectable in any of the ActD treated conditions. Blue, DAPI staining for DNA; Green, MPO; Red, CitH3. Scale bar, 25 μm. See Supplementary Fig. S6 for additional information.

Figure 7. Actinomycin D suppresses NETosis induced by biologically relevant Nox-dependent (LPS) and -independent (ionomycin) agonists.

(A,B) Neutrophils were induced by media control (-ve control), LPS (25 μg/ml) and ionomycin (5 μM) in the presence or absense of ActD (0, 0.625, 1.25, 2.5, 5.0 μg/ml). The Sytox green plate reader data show the suppression of NETosis by ActD in a dose-dependent manner (p < 0.05; One-Way ANOVA with Dunnett’s post-test conducted at each time points; n = 4). (C) Confocal imaging of the MPO and CitH3 immunostained cells at 2.5 h show NETosis (NET DNA colocalize with MPO). These ActD treated neutrophils do not show signs of considerable degree of NETosis (nuclear decondensation) or apoptosis (nuclear condensation). Furthermore, ionomycin-mediated NETosis show extensive citrullination of histones (CitH3) on decondensing nuclei, and on the NETs. Blue, DAPI staining for DNA; Green, MPO; Red, CitH3. Scale bar, 25 μm. See Supplementary Fig. S7A and S7B for additional information.

Figure 8. Transcriptional firing model of NETosis.

Upon activation with appropriate agonists, neutrophils generate reactive oxygen species, and activate distinct kinase cascades during the early steps of NETosis. Some of the kinases in these cascades phosphorylate and activate specific sets of transcription factors, which help to initiate transcription at specific promoters. Citrullination of histone (CitH3) at promoter regions could help to extensively decondense chromatin, particularly in rapid Nox-independent (intracellular calcium-dependent) NETosis. Genome-wide ranscriptional firing at these loci helps to promote chromatin decondensation necessary for NETosis. During the progression of NETosis, myeloperoxidase (MPO) and other granular components (e.g., elastases and other proteases that cleave chromatin) are deposited on chromatin, which is released as NETs.