key: cord-0254901-b6qx6zrd
authors: Ioannou, Marianna; Hoving, Dennis; Aramburu, Iker Valle; De Vasconcelos, Nathalia M.; Temkin, Mia I.; Wang, Qian; Vernardis, Spyros; Demichev, Vadim; Tsourouktsoglou, Theodora-Dorita; Boeing, Stefan; Goldstone, Robert; David, Sascha; Stahl, Klaus; Bode, Christian; Ralser, Markus; Papayannopoulos, Venizelos
title: SIGNR1 promotes immune dysfunction in systemic candidiasis by modulating neutrophil lifespan via T cell-derived histones and G-CSF
date: 2021-08-18
journal: bioRxiv
DOI: 10.1101/2021.08.09.455510
sha: 1e6529e18e6a79442ec11edcdb482359cae1efab
doc_id: 254901
cord_uid: b6qx6zrd

The mechanisms regulating immune dysfunction during sepsis are poorly understood. Here, we show that neutrophil-derived myeloperoxidase delays the onset of immune dysfunction during systemic candidiasis by controlling microbes captured by splenic marginal zone (MZ) macrophages. In contrast, SIGNR1-mediated microbe capture accelerates MZ colonization and immune dysfunction by triggering T cell death, T cell-dependent chromatin release and the synergistic induction of G-CSF by histones and fungi. Histones and G-CSF promote the prevalence of immature Ly6Glow neutrophils with defective oxidative burst, by selectively shortening the lifespan of mature Ly6Ghigh neutrophils. Consistently, T cell deficiency, or blocking SIGNR1, G-CSF or histones delayed neutrophil dysfunction. Furthermore, histones and G-CSF in the plasma of sepsis patients, shortened neutrophil lifespan and correlated with neutrophil mortality markers associated with a poor prognosis. Hence, the compromise of internal antimicrobial barrier sites drives neutrophil dysfunction by selectively modulating neutrophil lifespan via pathogenic T cell death, extracellular histones, and G-CSF.

in aberrant immune responses. It is estimated that in the year 2017 there were 48.9 million sepsis cases, of which 11 million died worldwide, accounting for approximately 20% of all global deaths (1). Although fungal-induced sepsis accounts for 5% of microbial sepsis, it is by far the most lethal form with mortality rates exceeding 45% and overall fungal infections cause an estimated 1.6 million deaths world-wide (2, 3) . The majority of these infections are caused by invasive candidiasis by Candida albicans, but the emergence of other pathogenic Candida species that are resistant to azoles is of particular concern.

Septic shock is characterized by hyper-inflammation, hypotension, coagulopathy and vascular damage that drive organ failure (4) . How and where systemic cytokines are induced during sepsis is not well defined. Several damage-associated molecular pattern (DAMP) molecules such as cell-free chromatin, high mobility group protein 1 (HMGB1) and S100 proteins have also been implicated in hyper-inflammation and sepsis pathology (5) (6) (7) . Extracellular histones are cytotoxic and pro-inflammatory (5, 6, 8, 9) , but their cellular origin and mechanisms regulating their release are unknown.

Immune dysfunction is also prominent during sepsis and is characterised by loss of cytokine responses, T cell deficiency, delayed neutrophil apoptosis and the prominence of immature neutrophils (10) (11) (12) (13) . Neutrophils from patients with septic shock display defective antimicrobial function exemplified by lower oxidative burst capacity (13) (14) (15) .

These extreme alterations in neutrophil populations persists well after septic shock and the ensuing immune deficiency negatively impacts survival following an episode of septic shock. Moreover, sepsis patients carry low T cell numbers in their spleens and prominent signs of T cell apoptosis that has been linked to Fas ligand or PD-1/PD-L1 signalling or super-antigen-mediated exhaustion (16) (17) (18) (19) (20) . Similarly delayed apoptosis occurs in mature neutrophils via the triggering receptor expressed on myeloid cells L4 (TREML4) receptor but the signals that regulate the specific loss of mature cells as opposed to immature neutrophils remain unknown (21). The signals that drive neutrophil dysfunction and the links with hyperinflammation and immune dysfunction in different immune cell types are poorly understood.

In addition to causing immune deficiency, immune dysfunction may also promote pathology. Neutrophils are critical for controlling invading microbes but are also implicated in tissue destruction and vascular pathology during sepsis (22) and SARS-CoV-2 infection pathology (23). One antimicrobial protein that is specifically expressed in neutrophils is myeloperoxidase (MPO), a granule enzyme that consumes hydrogen peroxide produced by the NADPH oxidase to generate hypochlorite and other halide oxidants (24). Patients with complete MPO deficiencies are prone to recurrent mucosal fungal infections, but rare episodes of systemic fungal infection episodes have also been reported in these patients (25) (26) (27) . MPO deficient mice are more susceptible to pulmonary infection with C. albicans and have increased fungal loads upon systemic challenge (28, 29) . MPO is also required for the release of neutrophils extracellular traps (NETs). NETs control fungi but also promote vascular pathology during sepsis, suggesting that MPO could play beneficial or pathogenic roles during systemic challenge (27, (29) (30) (31) .

Here we show that MPO protects against immune dysfunction by limiting a pathogenic programme that links T cell and neutrophil dysfunction. MPO is critical for controlling infection in the spleen, but over time fungal colonization triggers T cell death that promotes neutrophil dysfunction via the release of cell-free chromatin. Our data demonstrate a pathway that links microbial control to immune dysfunction in the innate and adaptive immune compartments via the critical roles of cell death and inflammatory cytokines.

To investigate whether MPO plays a protective or pathogenic role during systemic infection, we performed survival experiments in WT and MPO-deficient mice in a model of systemic candidiasis. WT mice infected with 1x10 5 WT C. albicans yeast succumbed 7-15 days post-infection, whereas MPO-deficient animals developed severe symptoms within 12 hrs post-infection exhibiting a substantial decrease in body temperature ( Fig.   1A and S1A). In contrast, pulmonary challenge of MPO-deficient animals with a higher fungal load took more than 7 days to lead to serious infection (29). MPO-deficient animals did not develop symptoms upon systemic challenge with the yeast-locked Δhgc1 mutant C. albicans strain. Moreover, MPO-deficient bone marrow (BM) neutrophils failed to control hyphal germination and growth in vitro (Fig. 1B) . In contrast to human peripheral blood neutrophils, these murine BM cells failed to release NETs as they died with condensed nuclei, indicating that hypochlorous acid but not hydrogen peroxide directly suppresses hyphal growth in the absence of NETosis (Fig. S1B) . These experiments suggested that MPO is critical for averting the rapid onset of sepsis.

Given that low MPO expression has been reported in macrophages, we also depleted neutrophils with an anti-Ly6G antibody or macrophages with clodronate liposomes (Clo-L)(32). Neutrophil depletion resulted in acute symptoms arising within 12 hrs, similarly to MPO-deficient animals. By contrast, Clo-L-treated mice developed symptoms only 48 hrs post-infection (Fig. 1A) . Consistently, MPO-deficient mice exhibited elevated circulating IL-1β concentrations (Fig. S1C) and elevated circulating concentrations of cell-free chromatin ( Fig. S1D and S1E ). IL-1β and cell-free histones were undetectable in infected WT mice at this early timepoint. The presence of circulating chromatin in MPO-deficient mice confirmed that NETs are not a major source of circulating histones.

To investigate which organs were afflicted by MPO deficiency we examined the fungal load at 12 hrs post-infection. Systemic candidiasis in mice drives sustained high fungal loads in the kidney. By day 1 post-infection, WT mice carried a low kidney fungal load, whereas MPO-deficient and Clo-L-treated mice exhibited comparably elevated fungal load (Fig. 1C) . However, Clo-L treated animals did not exhibit symptoms of disease as opposed to MPO-deficient animals exhibited significantly higher plasma IL-1β levels and symptoms of severe systemic inflammation such as a hunched position, a scruffy coat, severe hypothermia and lack of responsiveness to stimulation ( Fig. S1A and S1C) . We noted however, that MPO-deficient mice exhibited a significant fungal load in the spleen, suggesting that invasion of this organ may be relevant in the onset of acute symptoms. In 6 symptomatic WT animals, the relative numbers of neutrophils in the spleen peaked rapidly and the total number of cells exceeded the number of neutrophils found in the kidneys ( Fig. 1D and S1F ).

To understand how splenic colonization related to hyperinflammation we analysed pro-inflammatory mediators in the plasma, spleen and kidneys of naïve and infected WT and MPO-deficient animals using luminex-based cytokine arrays. The spleen and blood of MPO-deficient mice exhibited a similar pattern of increase in cytokines and chemokines compared to WT mice (Fig. 1E) . By contrast, kidney cytokine patterns between infected WT and MPO-deficient mice appeared similar. Together, these findings suggested that neutrophils control splenic colonization and suppress spleen-derived inflammation in an MPO-dependent manner.

Next, we investigated fungal colonization by immunofluorescence microscopy.

The kidneys of MPO-deficient and Clo-L-treated mice were colonized throughout the organs (Fig. S1G) whereas C. albicans colonized predominately the spleen marginal zone (MZ) in MPO-deficient animals (Fig. 1F) . The spleen contains several different macrophage populations. MZ macrophages are comprised of the outer layer that expresses the receptors SIGN-related 1 (SIGNR1) and MARCO and the inner layer expresses CD169 (Fig. S1H) . Clo-L administration had little effect on splenic macrophages within the first 24 hrs, but 48 hrs post-treatment eliminated a substantial number of F4/80 + red pulp cells, depleted SIGNR + macrophages completely and had a partial effect on CD169 + cells leaving a substantial number of CD169 + MZ macrophages were still present after Clo-L treatment (Fig. S1I) . WT and Clo-L-treated mice did not bear visible microbial colonies in the spleen 24 hrs post-infection which correlated with the lack of symptoms in these mice (Fig. 1F) .

To determine the precise localization of fungi in the spleen, we stained the spleens of infected WT mice collected from a period ranging from 4-7 days post-infection, which contained asymptomatic mice and mice that developed sepsis symptoms. Fungi colonized the spleen MZ of WT mice and colocalized with SIGNR1 + macrophages in asymptomatic mice with physiological body temperature 37°C (Fig. 1G) . C. albicans and Ly6G+/MPO + neutrophils did not co-localize in the spleen ( Fig. 1G and 1H) . In WT mice that had developed severe symptoms and exhibited a low body temperature, C. albicans had infiltrated the inner CD169 + macrophage layer, suggesting that spreading of the microbes through this barrier may play a role in pathology ( Fig. 1I and 1J) . Taken together, these data indicated that neutrophils were critical to control fungi that were captured by SIGNR1 + macrophages via MPO-derived ROS. The fact that MPO delayed colonization in areas that did not contain neutrophils suggested that neutrophil-derived ROS controlled hyphal growth in a non-cell-autonomous, transcellular manner.

Next, we explored whether the capture of fungi promotes sepsis pathology by targeting the SIGNR1 receptor. The C-type lectin receptor SIGNR1 is predominately expressed in the spleen and lymph nodes and binds C. albicans (33). SIGNR1 + cells could not be detected in the infected kidneys indicating that the antibody would not directly impact immune responses in this organ (Fig. S2A) . Moreover, we confirmed that this antibody did not deplete SIGNR1 + macrophages because we could still detect them in naïve mice with antibodies against MARCO (Fig. S2B) . SIGNR1 staining was not detectable in the spleens of mice treated with anti-SIGNR1, indicating the effective blockade of the receptor (Fig. S2C) . For these experiments, we raised the infection dose to 5x10 5 yeast particles per animal in order to facilitate the detection of any survival benefit against untreated WT mice that are more resistant to infection than MPO knockout animals. With the higher fungal dose, WT mice treated with a control IgG antibody developed severe symptoms 3 days post-infection, whereas SIGNR1 blockade delayed the onset of symptoms and extended survival by 3-fold ( Fig. 2A and S2D ).

SIGNR1 blockade reduced the fungal load in the spleen as well as the kidneys, despite the receptor not being expressed in this organ ( Fig. 2B and S2E) . Moreover, SIGNR1 blockade reduced NET deposition in the kidneys (Fig. S2F) and lowered cytokines such as IL-6 and G-CSF and chemokines in the spleen, blood and kidney 3 days post-infection ( Fig. 2C and   S2G ). To determine which splenocytes produced cytokines we detected cytokine mRNA transcripts by in fluorescence situ hybridization (RNAscope) staining in splenic sections. IL-6 and G-CSF were upregulated in CD169 + macrophages in infected IgG-treated mice and were reduced in anti-SIGNR1 treated mice (Fig. 2D) . Hence, SIGNR1 accelerates the onset of sepsis pathology and exerts its effects systemically in organs lacking cells that express the receptor.

Next, we investigated the impact of SIGNR1 blockade on fungal capture in the spleen. Anti-SIGNR1 treatment prevented the sequestration of C. albicans in the spleen MZ but microbes could be detected in other splenic areas, such as the white pulp. (Fig. 2E and 2F). In addition, we also noticed that Ly6G staining disappeared in infected IgGtreated mice but was present in anti-SIGNR1-treated animals (Fig. 2E) . The prominent 8 MPO staining in the infected spleens of untreated WT mice (Fig. 1I) suggested that neutrophils were still present in large numbers but may have downregulated Ly6G expression. To further test this hypothesis, we analysed neutrophil populations in the blood, spleens and kidneys by flow cytometry, tracking CD11b, MPO and Ly6G. The administration of the anti-SIGNR1 antibody did not alter the total neutrophil numbers in the blood on the day of infection nor did it affect emergency granulopoiesis 24 hrs postinfection (Fig. S3A) . However, 3 days post-infection, control animals contained a large fraction of neutrophils that had high MPO but low Ly6G expression ( Fig. 2G and 2H) . In contrast, mice treated with anti-SIGNR1 maintained the Ly6G high population in the blood, spleen and kidneys (Fig. 2G, 2H and S3B) . The shift towards an elevated Ly6G low / Ly6G high neutrophil ratio in the blood, spleen and kidney was gradual and preceded the onset of sepsis symptoms ( Fig. 2I and S3C and S3D) . We also plotted the change in neutrophil populations against the corresponding body temperature in an experiment where WT mice became gradually sick and were sacrificed at an intermediate timepoint that contained a mixed group of symptomatic and asymptomatic mice. These plots showed that although the total neutrophil population decreased by approximately 30%, there was a near 10-fold decrease in Ly6G high neutrophils and a trend for an increase in Ly6G low cells as body temperature decreased (Fig. 2J) .

Next, we explored whether the shift in neutrophil populations influenced the antimicrobial capacity of neutrophils. We evaluated superoxide production in neutrophils isolated from the spleens of infected control and anti-SIGNR1-treated mice 3 days postinfection. Neutrophils from anti-SIGNR1-treated mice produced a potent ROS burst response to phorbol-myristate-acetate (PMA), whereas neutrophils from control-treated mice exhibited a defective ROS burst (Fig. 2K) . Interestingly, SIGNR1 blockade failed to improve the survival of infected MPO-deficient mice even at 100-fold lower infection doses confirming that the beneficial impact of SIGNR1 blockade is mediated by regulating neutrophil effector function (Fig. S3E) . These data indicated that SIGNR1-mediated fungal capture controls splenic colonization and neutrophil dysfunction.

The presence of cytokines in MZ macrophages raised the question of whether they are undergoing cell death. Hence, we stained infected spleens with terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL). We detected no substantial TUNEL signal in MZ macrophages but there was extensive death of cells in the white pulp of infected control-treated mice (Fig. 3A) . Notably, cell death was absent in SIGNR1-blocked mice and staining with T cell markers demonstrated that the dying cells were predominately T cells that were positive for CD4, CD3 and TCRβ (Fig. 3A, 3B and   S4A ). Interestingly, a substantial number of TUNEL + T cells could be found outside germinal centers (Fig. 3A, 3C and S4A) . We also confirmed the loss of T cells in the spleen by flow cytometry which demonstrated that T cell death impacted T cells non-specifically and afflicted all Vβ-chain populations (Fig. 3D) . Staining these splenic samples with antibodies against the activated form of caspase-3 and capsase-8 confirmed that T cells were dying via apoptosis (Fig. 3E) . We also detected a SIGNR1-dependent increase in T cell death in the inguinal lymph nodes and the thymus although the latter was not statistically significant (Fig. 3F, 3G and S4B) . Interestingly, C. albicans was absent from lymph nodes and the thymus, indicating that T cell death was driven by systemic mediators (Fig. S4C) . Cell death also occurred in the kidneys and was significantly reduced in SIGNR1-blocked mice. However, dying kidney cells were not T cells as TUNEL positive cells were not positive for CD3 (Fig. S4D) . Given the extent of cell death, we examined whether SIGNR1 blockade affected circulating cell-free chromatin.

Consistently, DNA and histones were prominent in the plasma of infected IgG-treated mice but were absent in samples from SIGNR1-treated mice ( Fig. 3H and 3I) . Collectively, these data demonstrate that SIGNR1 regulates T cell death and the release of extracellular chromatin in the circulation.

To evaluate whether T cell death was pathogenic we examined the survival of RAG2-deficient mice that lack T cells and B cells and TCRα-deficient mice that lack αβ-T cells. Both knockout strains exhibited improved survival (Fig. 3J) and carried a lower fungal load in their kidneys 3 days post-infection (Fig. 3K) . Moreover, infected TCRαdeficient mice maintained Ly6G high neutrophils in the periphery indicating that T cells are involved in neutrophil dysfunction ( Fig. 3L and 3M) . Moreover, TUNEL staining was absent from the spleens of infected TCRα-deficient mice and adoptive transfer of T cells restored cell death in the spleen white pulp (Fig. S4E) . Staining with a FLICA poly-caspase activity assay further confirmed the elevated incidence of cell death in T cells and was absent in the spleen of T cell deficient mice (Fig. S4F) . Likewise, circulating chromatin was absent in the plasma of TCRα knockout animals and could be restored by adoptive transfer of T cells ( Fig. 3N and 3O) . Furthermore, T cell deficiency led to a reduction in plasma cytokines, including IL-6 and G-CSF (Fig. 3P) . Together, these observations indicated that dying T cells regulated the release of circulating chromatin, cytokine induction and alterations in neutrophil populations.

Next, we investigated whether cell death plays a role in cytokine induction and neutrophil dysfunction. Macrophage depletion experiments using Clo-L administration induced very high concentrations of circulating chromatin in the absence of infection but did not cause pathology (Fig. 4A) . The spleen is an organ rich in macrophages, therefore we hypothesized that the phenotype we observed in infected Clo-L-treated mice was not only caused by a reduction in macrophages but could also be accelerated by pre-existing chromatin release. Consistent with this hypothesis, Clo-L treatment accelerated the emergence of Ly6G low neutrophils and the reduction in total splenic neutrophil numbers upon infection ( Fig. 4B and S5A) . Therefore, cell death and circulating histones alone were not sufficient to induce pathology in the absence of infection but accelerated cytokine induction and neutrophil alterations upon infection.

Since there was a notable increase in G-CSF that correlated with the emergence of Ly6G low neutrophils, we examined whether G-CSF was differentially induced upon infection in Clo-L treated mice. While Clo-L alone in the absence of infection had little effect on the plasma concentrations of G-CSF, Clo-L pre-treatment amplified circulating G-CSF concentrations upon fungal challenge (Fig. 4C) . These data suggested that circulating histones and C. albicans could synergize in vivo to induce G-CSF. To test this hypothesis, we incubated bone marrow-derived macrophages with recombinant histone H3 alone or in combination with heat-inactivated yeast or hyphae. Histones or hyphae alone were weak G-CSF inducers, but the two signals acted synergistically to induce the cytokine (Fig.   4D ). This synergy depended on the hyphal form, as yeast were unable to induce G-CSF in the presence of histone H3. By contrast, histone H3 was sufficient to induce IL-6 and hyphae had no effect on the induction of this cytokine suggesting a divergence in the pathways that regulate IL-6 and G-CSF.

To test the impact of histones on neutrophils we sought an approach that would clear circulating chromatin enzymatically. We previously found that proteases and endonucleases synergize in chromatin clearance in the sputum of patients with cystic fibrosis (34). Therefore, we tested whether intraperitoneal administration of DNase I could promote chromatin clearance in the plasma that contains serum proteases (5) . We also treated mice with histone-blocking antibodies. DNase I administration delayed sepsis in infected mice (Fig. 4E) and cleared free circulating DNA and histones ( Fig. 4F and 4G) .

Moreover, mice treated with either DNase I or anti-histone antibodies maintained high circulating Ly6G high neutrophil populations ( Fig. 4H and S5B ) and decreased plasma G-CSF and IL-6 concentrations (Fig. 4I, S5C and S6A ) suggesting that plasma chromatin may be affecting neutrophil populations by regulating cytokines. Histone-blocking was more effective than DNAse I treatment and completely suppressed the increase in the Ly6G low /Ly6G high ratio as well as maintained a physiological body temperature (Fig. S6B) .

The link between cell-death-derived chromatin and G-CSF was intriguing since this cytokine is a key neutrophil regulator that amplifies granulopoiesis and mobilizes neutrophils from the bone marrow. G-CSFR-deficient mice are more susceptible to systemic C. albicans challenge despite exhibiting effective emergency granulopoiesis and neutrophil mobilization from the bone marrow (35). However, in our experiments G-CSF correlated with neutrophil dysfunction and was upregulated by SIGNR1 and circulating chromatin. Therefore, we hypothesized that prolonged exposure to G-CSF may promote pathology during systemic infection. To test this hypothesis, we neutralized G-CSF by daily administration of antibodies, starting at 24 -48 hrs post-infection, in order to avoid interfering with any protective functions of G-CSF during the early phases of the infection.

Late anti-G-CSF treatment resulted in a significant delay in the onset of severe symptoms ( Fig. 4J) . Elevated G-CSF plasma concentrations correlated with a decrease in body temperature (Fig. 4K ) and G-CSF neutralization suppressed the emergence of Ly6G low neutrophils ( Fig. 4L and S7A) . We also infected G-CSF-deficient mice with C. albicans and examined the effect on neutrophil populations. Due to peripheral neutropenia these mice are more susceptible to Candida systemic infection and therefore we infected with the low 1x10 3 C. albicans inoculum which induced severe symptoms 2-3 days post-infection.

Given that G-CSF elicits neutrophil egress from the bone marrow we analysed neutrophils in the bone marrow to ensure that any measured differences were not due to lack of neutrophil mobilization. Notably, the Ly6G low /Ly6G high neutrophil ratio in infected symptomatic G-CSF-deficient mice remained low (Fig. 4M) suggesting that G-CSF is required for the observed alterations in neutrophil populations.

The administration of G-CSF in the absence of infection is not known to cause large deleterious effects on granulopoiesis. However, G-CSF is required but not sufficient to upregulate minor populations of low-density granulocytes in murine cancer models (36). Therefore, we hypothesized that G-CSF may synergize with cell death-derived chromatin in regulating neutrophil dysfunction. Moreover, we sought conditions where we could decouple the emergence of Ly6G low neutrophils from infection. To test this hypothesis, we treated mice with PBS liposomes (PBS-L) or Clo-L and injected recombinant G-CSF after 24 hrs. We assessed splenic neutrophil populations by flow cytometry 48 hrs later. At that time, PBS-L and Clo-L-treated groups that had received rG-CSF had comparable G-CSF concentrations in their blood (Fig. 4N) . However, while G-CSF injection in PBS-L-treated mice did not alter the neutrophil ratio substantially, the combination of Clo-L and G-CSF induced a Ly6G low /Ly6G high neutrophil ratio that was comparable to that observed in infected symptomatic mice ( Fig. 4O and S7B) . Pre-treating mice with recombinant G-CSF 24 hrs prior to infection delayed the onset of sepsis indicating that boosting granulopoiesis prior to infection is beneficial and that G-CSF plays distinct roles in the presence and in the absence of circulating chromatin (Fig. S7C ). Together these data 5A ). To compare these cells at the molecular level, we conducted RNA sequencing analysis of Ly6G high and Ly6G low neutrophils sorted by flow cytometry from the spleens of either naïve or infected WT or TCRα-deficient animals and uninfected WT mice treated with G-CSF and Clo-L. PCA analysis indicated that Ly6G high and Ly6G low populations from infected WT symptomatic mice were distinct from one another but also clustered differently from Ly6G high neutrophils isolated from infected asymptomatic TCRα-deficient animals or naïve WT mice (Fig. 5B) . Moreover, Ly6G low neutrophils in G-CSF and Clo-L treated mice were distinct from naïve Ly6G high cells but also exhibited variations from Ly6G low neutrophils found in infected symptomatic mice. Next, we compared these populations based on the major gene expression differences between septic Ly6G low and Ly6G high cells across the entire dataset. Ly6G low neutrophils from infected WT mice had a distinct signature in comparison to Ly6G high populations but also when compared to Ly6G low neutrophils in infected asymptomatic TCRα-deficient mice (Fig. 5C) . Given the asymptomatic state, the TCRα-deficient Ly6G low cells comprised a "healthy" immature population under a state of infection. TCRα-deficient mice exhibited a Ly6G high population that was similar to Ly6G high populations in naïve WT mice. Moreover, WT infected Ly6G low cells were closer to Ly6G low cells from TCRα-deficient mice but also exhibited distinct transcriptional changes that set them apart from immature cells associated with an infected asymptomatic state.

There was a cluster of genes with similar trends amongst the uninfected Ly6G low cells isolated from uninfected G-CSF/Clo-L-treated-mice and the Ly6 low neutrophils derived from infected symptomatic WT mice supporting the hypothesis that G-CSF and chromatin promote shifts in neutrophil populations during infection. We also compared these cell populations for genes associated with the different stages of neutrophil maturation and found that Ly6G low neutrophils exhibited reduced expression in several maturation and post-maturation genes (Fig. 5C) . Comparison of infected WT Ly6G low with Ly6G low neutrophils from TCRα-deficient mice also highlighted changes in a number of genes that included CXCR2, CD14 ITGAM, Csf3r, CD34, CR2 and Arg1, which distinguished this population from TCRα-deficient Ly6G low cells. Infected WT Ly6G low neutrophils also lacked transcripts for C-type-lectin-receptor Clec2d and Clec7a, two C-type lectin receptors that recognize histones and fungal pathogens respectively as well as the complement receptor C5ar2, two Nod-like receptors and IL-1β, which could further compromise their antifungal activity ( Fig. 5C and 5D) . Notably, while Ly6g transcripts were downregulated in Ly6G low cells than in Ly6G high neutrophils from infected WT spleens, the transcripts for the receptor were comparably high in both populations isolated from infected TCRα-deficient mice (Fig. 5E) , although this was not reflected at the protein level. Infected WT Ly6G low cells had additional adaptations that distinguished them from TCRα-deficient Ly6G low neutrophils such as low levels of CXCR2, gasdermin D and E (Gsdmd and Dfna5) and higher Bcl2 transcript levels which may increase cell survival. Moreover, transcripts for the G-CSF receptor Csf3r were downregulated in Ly6G low neutrophils from infected mice or uninfected mice treated with G-CSF and Clo-L.

The downregulation of G-CSFR mRNA could result from a negative feedback response to sustained G-CSF signalling in WT infected animals that was absent in TCRα-deficient mice.

However, the G-CSFR protein could still be detected on the surface of Ly6G high and Ly6G low neutrophils from infected WT mice at comparable intensity as in Ly6G high cells from naïve WT mice indicating that these cells were still sensitive to G-CSF ( Fig. 5F and S7D) . These data confirmed that peripheral neutrophil populations in symptomatic mice were predominately immature cells that were molecularly distinct from peripheral mature and immature neutrophils found in infected mice lacking T cells, indicating that additional changes occur with disease severity.

T cell-derived histones eliminate mature Ly6G high neutrophils but not Ly6G low neutrophils in the bone marrow.

To explore the underlying mechanism behind the change in neutrophil populations, we examined the total numbers of neutrophils and their progenitors in the bone marrow (BM). Symptomatic infection in WT mice was accompanied by decreases in common myeloid progenitors CMPs as well as their derivative granulocyte-monocyte progenitors (GMPs) and megakaryocyte-erythrocyte progenitors (MEPs) (Fig. 6A) . The size and ratio of these populations remained largely unaffected by infection in asymptomatic TCRα-deficient mice with only CMPs exhibiting a slight reduction (Fig.   S8A ). By contrast CMP numbers were reduced by 3-fold in infected WT mice compared to naïve WT controls. Across the two genotypes, the number of CMPs and MEPs was also reduced in infected WT mice in comparison to TCRα-deficient mice, whereas there was no difference between GMPs amongst the two groups. This was an important observation because there was a dramatic decrease in Ly6G high neutrophils in the BM and the spleen of infected WT mice (Fig. 6B) . Ly6G low neutrophils were only reduced by approximately 50% which resulted in these cells becoming the predominant neutrophil population in the BM and in the periphery. By contrast, TCRα-deficient mice maintained a substantial Ly6G high population and their Ly6G low neutrophil numbers remained unchanged in the BM.

We took advantage in the relative difference between MPO and Ly6G expression levels in Ly6G low and Ly6G high neutrophils as determined by RNAseq (Fig. 5C) , in order to distinguish these cells by immunofluorescence microscopy. Consistently, the MPO signal in immature Ly6G low cells in the BM was high compared to mature Ly6G high neutrophils which enabled the specific detection of immature neutrophils by immunofluorescence microscopy. Similarly, Ly6G staining detected specifically mature Ly6G high neutrophils. As in the flow cytometry analysis, Ly6G high neutrophils were completely absent from the BM of infected WT mice but were still present in infected TCRα-deficient BM ( Fig. 6C and   6D ). These findings indicated that neutrophil population changes occurred predominately at the late maturation stage, rather than during early differentiation, with profound loss of mature Ly6G high neutrophils. Importantly, the infection impacted predominately mature neutrophil populations and to a much lesser degree immature granulocytes and their progenitors, since the BM of symptomatic WT and asymptomatic TCRα-deficient contained relatively equal numbers of GMPs and immature neutrophils.

To assess the role of histones in this process, we also examined the impact of histone-blocking antibodies on BM neutrophil populations. A cocktail of anti-H3 and anti-H4 antibodies maintained a significant presence of mature Ly6G high neutrophils in the BM detected by flow cytometry (Fig. 6E and 6F ) and by microscopy ( Fig. 6G and 6H ),

indicating that histones played an important role in eliminating mature neutrophils from the BM.

To investigate how these signals could impact the abundance of distinct neutrophil populations, we sorted Ly6G low and Ly6G high neutrophils from the BM of naïve WT mice as well as Ly6G low from infected symptomatic WT mice and incubated them with varying concentrations of recombinant G-CSF and histone H3 alone or in combination. We monitored cell death by time-lapse immunofluorescence microscopy for 1300 min ( To investigate the clinical relevance of this mechanism in human sepsis we measured the levels of G-CSF and histones in the plasma of sepsis patients with early (<24 hours) and severe (norepinephrine requirement > 0.4μg/kg/min) bacterial septic shock (37). We examined a group of patients with bacterial infections because bacterial sepsis cases are more prominent than fungal-associated sepsis, and although the upstream mechanisms regulating the induction of these cues may vary, we sought to understand whether the pathogenic signals played a relevant role in bacterial sepsis. Most patients exhibited elevated histone H3 and DNA concentrations in their blood when compared to healthy donors (Fig. 7C) . Therefore, we incubated primary human neutrophils from the blood of healthy volunteers with 3% sepsis patient plasma and measured their lifespan over a 2000 min time-course. Sepsis plasma reduced mature human neutrophil lifespan by approximately 30% (Fig. 7D) . This decrease depended on G-CSF and histones since neutralising antibodies against histone H3 and G-CSF significantly blocked lifespan reduction ( Fig. 7E and 7F) . The long duration of these lifespans was inconsistent with the immediate activation of cell death as cells persisted for 9 hrs before gradually starting to die. By contrast apoptosis is a rapid process (38). We also observed changes in neutrophil surface markers analysed by flow cytometry after 1000 min of incubation with sepsis plasma, compared to neutrophils incubated with healthy plasma (Fig. 7G ). Sepsis plasma, or HD plasma complemented with recombinant histone H3 and G-CSF, increased the surface expression of CD11b, CD15, CD16, CD66b, CD10, CD101 and PD-L1. These markers enable the identification of altered neutrophils in sepsis patients and may enable to monitor the degree of lifespan modulation (Fig. 7H) .

To assess the impact of this mechanism on sepsis severity, we generated a proteomic dataset for these patient plasmas and performed and unbiased search for significant correlations of histones and G-CSF with changes in plasma protein abundance.

Interestingly, we found that histone H3 correlated with increased concentrations of cellfree plasma S100A8/9, which are proteins expressed in the cytosol of neutrophils and are released upon neutrophil death. Moreover, elevated plasma S100A8/9 in sepsis patients is associated with increased risk for mortality (38). Our analysis indicated that S100A8/9 levels were elevated in patients with intermediated G-CSF plasma concentrations but were suppressed by very high G-CSG concentrations (Fig. 7I) , likely to be due to improved emergency granulopoiesis which may counter the detrimental effects of G-CSF on neutrophil lifespan (7, 39, 40) . To better understand the impact of both plasma histones and G-CSF, we separated patients into medium and high G-CSF and histone groups and compared their plasma S100A8/9 content. Consistently, there was a significant increase in S100A8/9 in patients with medium G-CSF and high histone H3 plasma concentrations ( Fig. 7J) . Therefore, cell-free histones and G-CSF alter neutrophil lifespan during sepsis in a manner that is consistent with a gradual reduction in mature neutrophil populations over the course of infection.

Our findings uncover a mechanism linking SIGNR1 + macrophages to T cell death and neutrophil dysfunction. Mature neutrophils control fungi captured by SIGNR1 + macrophages via MPO-derived ROS. Progressive fungal colonization of the MZ promotes T cell death and the release of extracellular chromatin that synergizes with hyphae to induce G-CSF in CD169 + MZ macrophages. Sustained G-CSF production and T cell-derived chromatin compromised antimicrobial function by eliminating competent neutrophils and leaving behind immature cells with a defective ROS burst. These findings suggest that the spleen MZ forms an internal anti-microbial barrier, and its disruption accelerates the onset of sepsis pathology. Hence, immune dysfunction in neutrophils and T cells is linked, with neutrophils attenuating T cell death via microbial control and T cell death promoting the release of pathogenic factors that deplete mature immunocompetent neutrophils.

The importance of the spleen in antifungal defence was recently highlighted by the increased susceptibility of splenectomised mice challenged systemically with C.

albicans (41) . Similarly, asplenic patients are at a high risk of bacterial and fungal infection (42) . Our study is consistent with the notion that while a healthy spleen is essential for the effective clearance of circulating pathogens, overwhelming this line of defence is detrimental. Similarly, while SIGNR1 + and CD169 + macrophages play beneficial roles in clearing blood-borne bacterial and viral infections (43) (44) (45) , retroviruses can hijack this process to infect T cells (46) . Similarly, our results indicate that C. albicans can take advantage of SIGNR1-mediated capture to initiate a programme that disrupts neutrophil antimicrobial function. Whether this process applies to bacterial infections is unclear, particularly since MPO deficiency has a minor effect on bacterial sepsis caused by cecal ligation and puncture (47) . However, our experiments with bacterial sepsis patient samples indicate that the downstream signals that influence neutrophil lifespan are relevant in systemic bacterial infections.

Our study sheds light on mechanisms regulating the release of circulating histones in sepsis. Neutrophils are not a major source of histones as shown previously by neutrophil depletion experiments (48) and confirmed by the upregulation of extracellular chromatin in infected MPO-deficient mice. Here, we report for the first time, a cell population whose depletion abrogates plasma chromatin release suggesting that dying T cells are the key regulators and most likely source of cell-free chromatin in the circulation.

While the mechanisms that drive T cell death in the spleen remain undefined, this process may depend on signals from CD169 + macrophages as demonstrated in other models via the action of monocyte-derived Fas ligand (17) . Consistently, the activation of caspase-8 and caspase-3, TUNEL staining, and chromatin fragmentation are all indicative of T cell apoptosis. However, death of other cell types may also promote histone release into the circulation as demonstrated by our macrophage depletion experiments with Clo-L in uninfected mice. Furthermore, our study indicates that endogenous circulating free chromatin is not sufficient to cause acute pathology in the absence of infection. The amount of histones released by Clo-L treatment which did not cause pathology, was superior to the of histone release upon infection. Prior studies that demonstrated direct lethality in response to recombinant histone injections resulted in blood concentrations that were likely to be 100-fold higher than in naïve asymptomatic clodronate-treated animals (5) . Moreover, the immunomodulatory properties of circulating chromatin suggest that caution should be exercised when Clo-L or any cell death-promoting agents are employed as their impact on cell death can alter innate immune production.

Our data indicate that MZ macrophages are important contributors of cytokines during systemic infection. The signals required to trigger various cytokines differ.

Extracellular chromatin is sufficient to induce IL-6 and IL-1β (9) . However, the induction of G-CSF requires the simultaneous presence of histones and fungal signals that signify an uncontrolled infection with a cytotoxic impact. During homeostasis, splenic macrophages regulate granulopoiesis by monitoring the presence of apoptotic circulating neutrophils (49) . We now show that macrophages also link the detection of systemic microbes and DAMPs to the alteration of neutrophil populations.

The selective loss of mature neutrophils has also been demonstrated in murine bacterial sepsis models and linked to TREML4 (21). However, the cell death inducing signals and the basis of the selectivity for mature over immature cells remained unclear.

Our study demonstrates that histones and G-CSF regulate this process selectively and may be relevant in other conditions, given that alterations in the age of neutrophils influences neutrophil function (50) (51) (52) . The ability of G-CSF to shorten the lifespan of mature neutrophils explains how G-CSF can play both beneficial and detrimental roles during systemic infection. Furthermore, these signals induce a cell-surface signature that enables these cells and lifespan-reducing mechanisms to be detected. The different impact these signals exert on the lifespan of mature and immature neutrophils is a paradigm that explains the selective depletion of specific neutrophil populations.

Theoretically, the loss of mature neutrophils should expand the size of hematopoietic niches in the bone marrow (49, 53) . However, all the progenitor numbers were reduced which may account for modest decreases in immature cells. The decrease in progenitors could be caused by the shift towards a ROS-deficient population given that ROS production is required for the expansion of progenitors during emergency granulopoiesis (54) . Moreover, the elevated G-CSF levels may increase mature neutrophil egress in WT mice but can't alone explain the loss of mature cells from the BM as the mature neutrophil numbers were low in the periphery. In addition, S100A8/9 proteins have recently been implicated in the induction of pathogenic neutrophil subsets in severe COVID-19 infection, suggesting by release of these alarmins, neutrophil lifespanreduction could further impact neutrophil populations and disease severity (23).

The transcriptional profile of Ly6G low neutrophils in symptomatic WT mice suggests that they are immature cells with additional infection-associated transcriptional changes, as they exhibit a distinct signature compared to Ly6G low neutrophils in asymptomatic TCRα-deficient mice. Consistently, while the effect of sepsis on neutrophil can be nearly fully recapitulated by G-CSF and Clo-L administration, infection promotes additional changes in mature and immature neutrophils.

The basis for the different sensitivity of lifespan-shortening in mature and immature neutrophils to G-CSF and histones is unclear. Several transcriptional adaptations in the Ly6G low neutrophil population of symptomatic mice may be important for survival. The downregulation of Csf3r transcript may result from a negative feedback in response to hyper-activation of G-CSFR. Ly6G low cells maintained their surface protein expression of G-CSFR but they remain resistant to the lifespan reduction effects of its ligand. The downregulation of Clec2d may provide another mechanism that reduces Ly6G low neutrophil exposure to histones given that this receptor mediates the uptake of extracellular chromatin (55) . Moreover, the downregulation of GSDMD and GSDME transcripts and the upregulation of the anti-apoptotic factor Bcl2 may render Ly6G low neutrophils more resistant to cell death.

In addition, several changes in gene expression may compromise the antimicrobial function of Ly6G low neutrophils in symptomatic mice. A number of important genes for antimicrobial defence appear to be downregulated in these cells, such as Dectin-1 which is critical for fungal recognition and IL-1β which promotes neutrophil recruitment and swarming (56, 57) . The predominant Ly6G low neutrophil population exhibited lower ROS production capacity despite the expression of all major genes involved in NADPH oxidase function. Interestingly, Ly6G low neutrophils exhibited a higher expression of MPO transcript and protein compared to Ly6G high neutrophils allowing these cells to be distinguished from mature neutrophils in the bone marrow by microscopy. Lower ROS production also renders neutrophils more proinflammatory in fungal and bacterial infections which may contribute to sepsis pathology (57) . However, the abundance of NETs in the kidneys of symptomatic mice, suggests that NETosis is not disrupted and may be driven either by the remaining minor mature neutrophil population 20 or immature neutrophils with sufficient ROS burst. On the contrary, anti-SIGNR1 treatment reduced NETosis possibly by direct control of fungi via ROS, which may further reduction in pathology.

The pathogenic dysregulation of neutrophil populations by G-CSF and circulating histones may be relevant in other conditions such as cancer. Tumour-derived G-CSF is required but not sufficient to upregulate a minor population of pro-metastatic, tumourinduced low-density granulocytes (36, 58, 59) . Tumour-induced DAMPs may play a role in driving pro-tumorigenic neutrophil populations and may provide further mechanistic understanding for the therapeutic potential of DNase treatment (60) (61) (62) (63) . Consistently, in our experiments, Ly6G low neutrophils from infected symptomatic mice have increased arginase I expression which inhibits T cell proliferation and promotes tumour growth.

Moreover, neutrophils from healthy human donors upregulated PD-L1 when incubated with sepsis plasma or G-CSF and histones, indicating that these conditions could enhance their pro-tumorigenic capacity. A sub-population of neutrophils with upregulated PD-L1 has also been described in COVID-19 patients and thus this mechanism may drive altered neutrophil populations in severe COVID-19 patients, where circulating chromatin and hyperinflammation have also been reported (64, 65) .

linked and can be targeted by interfering with microbial sequestration in sensitive areas and by eliminating circulating chromatin. The presence of G-CSF and circulating histones in many conditions suggests that they may play a generalized role as drivers of immune dysfunction that can be targeted therapeutically through histone-blocking strategies. directed the study, analysed data and wrote the manuscript.

The high throughput RNA sequencing expression profile data can be accessed at NCBI Gene Expression Omnibus (GSE160301) and the token (ydyjocqyjpczjgt). 

Peripheral venous blood was collected into EDTA tubes, layered on Histopaque 1119 (Sigma-Aldrich) and centrifuged for 20 min at 800x g. The plasma, PBMC and neutrophil layers were collected and neutrophils were washed in Hyclone Hank's Balanced Salt Solution (HBSS) without calcium, magnesium or phenol red (GE Healthcare) supplemented with 10mM HEPES (Invitrogen) 0.1% plasma and further fractionated on a discontinuous Percoll (GE Healthcare) gradient consisting of layers with densities of 1105 g/ml (85%), 1100 g/ml (80%), 1093 g/ml (75%), 1087 g/ml (70%), and 1081 g/ml (65%) by centrifugation for 20 min at 800x g. Neutrophil enriched layers were collected and washed.

We isolated 2x10 5 neutrophils human peripheral blood and seeded them in a U-bottom 96-well plate in HyClone HBSS +Ca, +Mg, -Phenol red (GE Healthcare) supplemented with 10mM HEPES (Invitrogen) and 3% plasma from healthy or septic donors. In vitro stimulation was performed with 500nM of human recombinant Histone 3 (Cayman chemical) and 5ng/ml recombinant human G-CSF (BioLegend). The cells were then incubated for 16 hours at 37 o C and 5%CO2. Cells were then washed, stained and fixed for flow cytometry analysis according to the steps described below. to obtain the neutrophil half-lives. In vitro stimulation was performed with 100nM, 500nM and 1uM of human recombinant Histone 3 (Cayman chemical) and 1ng/ml, 5ng/ml or 50ng/ml recombinant human G-CSF (BioLegend) or mouse G-CSF (BioLegend).

The blocking of H3 and G-CSF was performed by preincubating the different 3% plasma with 0.5 µg/ml of anti-hH3, 1 µg/ml and of anti-hG-CSF or 1.5 µg/ml of anti-IgG at 37°C for 30 mins before adding to the neutrophils.

Wild-type Candida albicans (C. albicans, clinical isolate SC5314) was cultured overnight shaking at 37oC and subcultered to an optical density (A600) of 0.4-0.8 for 4 hours in yeast extract peptone dextrose (YEPD; Sigma) medium. Subcultures were examined for lack of hyphae, washed and resuspended in sterile phosphate-buffered saline (PBS) immediately prior to infection. Mice were injected intravenously with either 1x10 3 , 1x10 4 , 1x10 5 or 5x10 5 C. albicans yeast particles per mouse. The weight and rectal temperature of the mice were recorded prior to infection and daily over the course of infection to track health status. A body temperature below 32 o C, a weight loss superior to 80% of initial weight accompanied by slow movement and non-responsiveness were considered collectively as septic shock and the humane endpoint for the mice. The mice were culled via cervical dislocation or by lethal dose of pentobarbital (600mg/kg) with mepivacaine hydrochloride (20 mg/ml).

Neutrophil depletion was achieved with intraperitoneal injection of 150μg anti-Ly6G Ab (BioXCell) or IgG isotype control (BioXCell) at day -1 and day 0 (day of infection).

Clodronate liposomes or 1mg PBS liposomes as control (Liposoma) at 1 day prior to infection. 

T cells were isolated from naïve spleens with the EasySep T cell isolation kit (STEMCELL Technologies). For each mouse 4x10 6 naïve T cells were injected intravenously via the tail, 2 days prior to infection.

Mice were injected intravenously with 1mg of Clo-L or PBS-L (Liposoma) and the subsequent day intraperitoneally with 2.5μg rG-CSF (BioLegend) or vehicle (PBS).

Analysis of neutrophil populations was performed two days after rG-CSF injection, according to the methods described in flow cytometric analysis.

Wild-type C. albicans was cultured overnight shaking at 37 o C in YEPD and then the following day sub-cultured to an optical density (A600) of 0. 

Freshly extracted organs were embedded in optimal cutting temperature (OCT) compound cryo-embedding media (VWR Chemicals BDH) and flash-frozen in a dry ice/100%ethanol slurry. Frozen sections (8μm thickness) were dried, fixed in 4% paraformaldehyde (PFA; Sigma) and permeabilized with 0.5% Triton X-100 in PBS. RNAscope in situ hybridization assay was performed on frozen sections (8μm) using RNAscope 2.5 LS probes for G-CSF and IL-6 (Mm-CSF3-C2, cat. 400918, Mm-IL6 cat.

according to the manufacturer's instructions. Subsequently, the slides were stained with primary and secondary antibodies following the methods described above.

Pan-caspase labelling was performed using the Poly Caspase Assay Kit (MyBioSource) according to the instructions provided by the manufacturer.

All stained tissue sections were mounted in ProLong Gold (Molecular Probes).

Images were taken using the Leica TCS SP5 inverted confocal microscope (20x, 40x, 63x original magnification) and analysis was performed using Fiji/ImageJ version 2.0.0 software.

Kidneys were chopped up and incubated in a digestion medium containing 0.2mg/ml Liberase TL (Roche) and 0.1mg/ml DNase I (Roche) for 20 minutes while shaking at 37 o C.

The digested kidney tissue was filtered and centrifuged in a Percoll (GE Healthcare) density gradient (40%/70%). Spleens were gently meshed in FACs buffer (PBS containing 3% FCS from Sigma) using a 40μM cell strainer to prepare single cell suspensions. Whole blood was centrifuged in Histopaque-1119 (Sigma) in order to isolate peripheral blood leukocytes. ACK (ammonium-chloride-potassium, GIBCO) lysing buffer was used in all tissue samples to eliminate remaining erythrocytes. 

For neutrophil characterization by nuclear morphology, 10 5 -10 6 sorted neutrophils were cytospun on slides, fixed in methanol and stained with freshly prepared and filtered Giemsa staining buffer, containing Giemsa R (Ral Diagnostics). All slides were mounted in dibutylphthalate polystyrene xylene (DPX) and imaged using the ZEISS Axio Observer Z1.

Mouse spleens were gently meshed using a 40μM cell strainer and erythrocytes were lysed with ACK lysing buffer (GIBCO). Neutrophils were isolated via negative selection with the EasySep mouse neutrophil isolation kit (STEMCELL Technologies). A total of 1x10 5 neutrophils were then incubated with the chemiluminescent probe luminol (Sigma) and horseradish peroxidase (HRP; Sigma) and stimulated with 100nM PMA;

(Sigma). Reactive oxygen species (ROS) production is measured by HRP mediated oxidation of luminol, producing a chemiluminescent signal that is detected with an UV filter on a spectrophotometric microplate reader (Fluostar Omega, BMG labtech).

Laboratories). Proteins were then transferred to a polyvinylidene difluoride (PVDF) membrane (Bio-Rad Laboratories) via semi-dry transfer. Nonspecific binding was blocked with 5% bovine serum albumin (BSA; Fisher Scientific) in tris-buffered saline with 0.1% Tween 20 (TBS-T). The membranes were blotted with anti-histone 3 (Milipore) and detected with HRP-conjugated goat anti-rabbit (Thermo Scientific).

Finally, the membranes were incubated with enhanced chemiluminescent substrate (ECL; Thermo Fisher Scientific) and imaged with a chemiluminescence imaging systems (Bio-Rad) or developed after exposure onto an X-ray film (Kodak) and digitally scanned.

DNA in plasma was quantified using the Quant-iTTM PicoGreen dsDNA assay kit (Thermo Fisher Scientific) following the instructions provided by the manufacturer. The fluorescent signal (excitation at 488nm) was measured using a spectrophotometric microplate reader (Fluostar Omega, BMG labtech). For protein quantification, raw data acquired were processed with DIA-NN 1.7.10 with the "robust LC (high precision)" mode with MS2, MS1 and scan window size set to 20ppm, 12ppm and 8 respectively.

Spleen leukocytes were extracted via methods described in the flow cytometry section.

Following homogenization, samples were enriched using the EasySep mouse neutrophil isolation kit (STEMCELL Technologies) and subsequently stained with antibodies and isolated via flow cytometric-sorting. RNA extraction was performed with the RNeasy mini kit (Qiagen) and the libraries were prepared with the Nugen cDNA synthesis kit.

Sequencing was performed on an Illumina HiSeq 4000 machine. The 'Trim Galore!' utility version 0.4.2 was used to remove sequencing adaptors and to quality trim individual reads with the q-parameter set to 20 (1). Then sequencing reads were aligned to the mouse genome and transcriptome (Ensembl GRCm38 release-89) using RSEM version 1.3.0 in conjunction with the STAR aligner version 2.5.2 (68, 69) . Sequencing quality of individual samples was assessed using FASTQC version 0.11.5 and RNA-SeQC version 1.1.8 (70) . Differential gene expression was determined using the R-bioconductor package DESeq2 version 1.24.0(71). Gene set enrichment analysis (GSEA) was conducted as described in (72) .

Single comparison statistical significance was assessed by an unpaired, two-tailed Mann- 

C. Dubois et al., High plasma level of S100A8/S100A9 and S100A12 at admission indicates a higher risk of death in septic shock patients. Sci Rep 9, 15660 (2019). 

I. Correlation between the fold-increase in plasma histone H3 (left plot) or G-CSF (right plot) and the relative abundance of S100A8 (green) and S100A9 (purple) detected by mass spectrometry in the plasma of sepsis patients. Boxes depict medium and high histone H3 containing plasmas. Boxes depict medium histone H3 or G-CSF-containing plasmas against the high content plasmas.

J. Concentration of S100A8 in medium (left panel) and high (right panel) G-CSF-containing plasmas grouped based on medium (<20X fold over HD) or high (>20X fold over HD)

histone H3 plasma concentrations.

Statistical analysis by unpaired Mann-Whitney t-test (* p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001). 

TCRα -/-WT

Clo-L GCSF High

Inf. Low

Clo-L GCSF Low N High

TCRα -/-WT

Clo-L GCSF High

Inf. Low

Clo-L GCSF Low Figure S4 . Effects of anti-SIGNR1 treatment on cell death in the kidney and lymph nodes.

A. Immunofluorescence micrographs from spleen of WT mice pre-treated control antibody and infected with 5x10 5 WT C. albicans, 3 days post-infection and stained for double stranded DNA breaks (TUNEL), CD3 CD169 and B220. Scale bars: 50 μm B. Immunofluorescence micrographs and quantification of thymus from WT mice, either naïve or infected with 5x10 5 WT C. albicans and treated with control IgG or anti-SIGNR1 antibodies, 72 hrs post-infection and stained for TUNEL and CD3. Scale bars: 100 μm (upper row) and 25 μm (lower row). C. Immunofluorescence confocal micrographs of thymus and inguinal lymph nodes from WT mice, either naïve or infected with 5x10 5 WT C. albicans, 72 hrs post-infection and stained for double stranded DNA breaks (TUNEL), CD3 and C. albicans. Scale bar: 100 μm. D. Immunofluorescence micrographs of kidneys from WT mice pre-treated with either control antibody or an anti-SIGNR1 blocking antibody and infected intravenously with 5x10 5 WT C. albicans, at 72 hrs post-infection. Staining for C. albicans, CD3, TUNEL, DAPI. Scale bar: 50 μm. E. Immunofluorescence micrographs form the spleens of TCRα-deficient mice alone or after adoptive T cell transfer and infected with 5x10 5 WT C. albicans, 72 hrs post-infection. F. Immunofluorescence micrographs from the spleens of naive or infected WT or TCRα-deficient mice infected with 5x10 5 WT C. albicans, 3 days post-infection and stained for CD3 and poly-caspase enzyme activity performed with a FLICA poly-caspase activity assay. Scale bars: 50μm. Statistical analyses by unpaired Mann-Whitney t-test (* p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001). A. Cytokines and chemokines measured by multiplex immunoassay in the plasma of naive WT mice or pre-treated with IgG or anti-H3/anti-H4 antibodies and infected with 5x10 5 C. albicans, 48 hrs post-infection. B. Body temperatures of infected WT mice treated with either an isotype control antibody (IgG), DNase I, or antibodies against histone H3 and H4 from experiment in Fig. 4J and Fig. 6E -H. Statistical analysis by two-way Avova and unpaired Mann-Whitney t-test for single comparison (IL-6) (* p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001). Half-life ratio of Ly6G high and Ly6G low neutrophils from naive mice or Ly6G low neutrophils from infected symptomatic mice supplemented with histone H3 or G-CSF alone or in combination at varying concentrations as indicated in the color scheme over naive (lane 1). C. Difference in half-life between WT naive Ly6G high and Ly6G low neutrophils (black and grey shades) or WT naive Ly6G high and Ly6G low Ly6G low neutrophils from infected symptomatic mice (purple shades) alone or supplemented with histone H3 or G-CSF alone or in combination at the indicated concentrations. 

Global, regional, and national sepsis incidence and mortality, 1990-2017: analysis for the Global Burden of Disease Study

Hidden killers: human fungal infections

Global and Multi-National Prevalence of Fungal Diseases-Estimate Precision

Cytokine storm and sepsis disease pathogenesis

Extracellular histones are major mediators of death in sepsis

Extracellular histones are mediators of death through TLR2 and TLR4 in mouse fatal liver injury

Calprotectin, a new biomarker for diagnosis of acute respiratory infections

Endogenous histones function as alarmins in sterile inflammatory liver injury through Toll-like receptor 9 in mice

Histones, DNA, and Citrullination Promote Neutrophil Extracellular Trap Inflammation by Regulating the Localization and Activation of TLR4

Immunosuppression in patients who die of sepsis and multiple organ failure

Broad defects in the energy metabolism of leukocytes underlie immunoparalysis in sepsis

Persistent neutrophil dysfunction and suppression of acute lung injury in mice following cecal ligation and puncture sepsis

Impairment of polymorphonuclear neutrophil functions precedes nosocomial infections in critically ill patients

Single-cell transcriptome profiling reveals neutrophil heterogeneity in homeostasis and infection

Immature circulating neutrophils in sepsis have impaired phagocytosis and calcium signaling

Caspase inhibitors improve survival in sepsis: a critical role of the lymphocyte

Post-injury immunosuppression and secondary infections are caused by an AIM2 inflammasome-driven signaling cascade

PD-L1 blockade improves survival in experimental sepsis by inhibiting lymphocyte apoptosis and reversing monocyte dysfunction

Different Expression Characteristics of LAG3 and PD-1 in Sepsis and Their Synergistic Effect on T Cell Exhaustion: A New Strategy for Immune Checkpoint Blockade

PD-1 expression by macrophages plays a pathologic role in altering microbial clearance and the innate inflammatory response to sepsis

Calprotectin is superior to procalcitonin as a sepsis marker and predictor of 30-day mortality in intensive care patients

Inflammatory Ly6C(high) Monocytes Protect against Candidiasis through IL-15-Driven NK Cell/Neutrophil Activation

Infections in asplenic patients. Clinical microbiology and infection : the official publication of the European Society of

The C-type lectin SIGN-R1 mediates uptake of the capsular polysaccharide of Streptococcus pneumoniae in the marginal zone of mouse spleen

SIGN-R1 contributes to protection against lethal pneumococcal infection in mice

A Protective Role for the Lectin CD169/Siglec-1 against a Pathogenic Murine Retrovirus

Retroviruses use CD169-mediated trans-infection of permissive lymphocytes to establish infection

Neutrophils employ the myeloperoxidase system to generate antimicrobial brominating and chlorinating oxidants during sepsis

Origin of Circulating Free DNA in Sepsis: Analysis of the CLP Mouse Model

Phagocytosis of apoptotic neutrophils regulates granulopoiesis via IL-23 and IL-17

Neutrophil ageing is regulated by the microbiome

Programmed 'disarming' of the neutrophil proteome reduces the magnitude of inflammation

A Neutrophil Timer Coordinates Immune Defense and Vascular Protection

Rhythmic modulation of the hematopoietic niche through neutrophil clearance

Myeloid cell-derived reactive oxygen species externally regulate the proliferation of myeloid progenitors in emergency granulopoiesis

Immune Sensing of Cell Death through Recognition of Histone Sequences by C-Type Lectin-Receptor-2d Causes Inflammation and Tissue Injury

Dectin-1: a signalling non-TLR pattern-recognition receptor

Reactive Oxygen Species Localization Programs Inflammation to Clear Microbes of Different Size

Priming of neutrophils toward NETosis promotes tumor growth

Cancers predispose neutrophils to release extracellular DNA traps that contribute to cancer-associated thrombosis

Neutrophils escort circulating tumour cells to enable cell cycle progression

DNA of neutrophil extracellular traps promotes cancer metastasis via CCDC25

Extracellular DNA in pancreatic cancer promotes cell invasion and metastasis

Cancer cells induce metastasis-supporting neutrophil extracellular DNA traps

Severe COVID-19 Is Marked by a Dysregulated Myeloid Cell Compartment

Neutrophil extracellular traps in COVID-19

Ultra-High-Throughput Clinical Proteomics Reveals Classifiers of COVID-19 Infection

DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput

STAR: ultrafast universal RNA-seq aligner

RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome

RNA-SeQC: RNA-seq metrics for quality control and process optimization

Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2

Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles

We thank Andrew Aswani for sourcing sepsis patient samples, George Kassiotis, Brigitta Stockinger and Ilaria Malanchi for kindly providing reagents and mice and the Francis Crick Institute core facilities. This research was funded in whole, or in part, by the Wellcome Trust (FC0010129, FC001134) . For the purpose of Open Access, the author has