Gram-positive Staphylococcus aureus LTA promotes distinct memory-like effects in murine bone marrow neutrophils

: Neutrophils as innate immune cells primarily act as first responders in acute infection and directly maintain inflammatory responses. However, a growing body of evidence suggests that neutrophils also bear the potential to mediate chronic inflammation by exhibiting memory-like features. We recently showed that priming by serial doses of lipopolysaccharide (LPS) from gram-negative bacteria can trigger opposing memory-like responses (exaggerated inflammation, i.e. trained sensitivity or suppression of inflammation, i.e. tolerance) depending on the LPS-dose. We now asked whether this observation could also hold true for lipoteichoic acid (LTA) from gram-positive S. aureus . We found comparable effects of LTA on neutrophil priming as seen for LPS. Low-dose (1 ng/mL) LTA-priming promoted increased production of proinflammatory mediators (i.e., TNF-α, IL-6, ROS), whereas high-dose (10 µg/mL) results in contrary reactions supporting anti-inflammatory responses by increased IL-10 and declined pro-inflammatory capacity. In vitro neutrophil recruitment was similarly regulated by LTA -priming. Investigation of signalling patterns revealed TLR2/MyD88-mediated regulation of NFκB-p65 through intermediate PI3Ks/MAPK. Collectively, our data suggest a previously unknown capacity of neutrophils to be differentially primed by varying doses of LTA, endorsing memory-like features in neutrophils. this observation, activation of protein kinase B (Akt) as a key downstream mediator of phosphoinositide 3-kinases (PI3Ks), as well as extracellular


Introduction
Neutrophils are the most abundant type of polymorphonuclear (PMN) leukocytes and among the first innate immune cells entering the site of infection [1,2]. Formerly, neutrophils have been considered as evolutionary primitive and functionally restricted to pro-inflammatory properties during acute infections [3][4][5]. They continuously survey for any microbial infections and act against them engaging an armory of antimicrobial agents such as cytokines and chemokines, reactive oxygen species (ROS), enzymes (i.e., defensins, cathepsins, lysozymes) as well as they improve their recruitment cascade to encounter the invading pathogens [4][5][6][7]. Once released into the circulation, neutrophils have a short life-span of around 48h after which they get cleared by the liver, spleen or bone marrow -a process regulated by CXC-chemokine receptor 4 (CXCR4), or they get phagocytosed by macrophages [8][9][10][11][12]. Even though neutrophils are too short-lived to primarily shape and regulate long-term responses, recent evidence also from out lab demonstrates their potential to at least induce distinct memory-like inflammatory responses [13][14][15][16].
Innate immune cells can sense different conserved microbial structures termed pathogen-associated molecular patterns (PAMPs) by their germline-encoded pattern recognition receptors (PRRs) [17]. Activation of PRRs results in downstream signalling events culminating in the production of inflammatory mediators [17][18][19][20]. For a long time, innate immune cells have been considered evolutionary primitive due to a lack of immunological memory as an exclusive hallmark of adaptive (acquired) immunity. A rising number of studies suggest however, that innate immune cells (i.e., monocytes, macrophages, microglia, natural killer (NK) cells) as well as tissue-resident stem cells possess a memorylike capacity during secondary infections [21][22][23][24][25][26]. Netea and colleagues were among the first to discover that innate immune cells primed by βglucan or bacillus Calmette-Guerin (BCG) after a secondary challenge triggered the release of pro-inflammatory mediators, whereas cells primed by the principal outer membrane component of gram-negative bacteria, lipopolysaccharide (LPS), were marked by increased antiinflammatory features (i.e., IL-10) and suppressed pro-inflammatory reactions [21,24,[27][28][29][30][31][32]. These findings concluded a PAMP-specific induction of the opposing memory-like inflammatory responses. Later on, several findings disclosed the pathogen dose as well as the maturation state as crucial determinants shaping the induction of these immune memory in innate immune cells [24,28,[33][34][35][36][37]. Epigenetic reprogramming and changes in cell metabolism have been identified to orchestrate both types of memory-like responses [29,[38][39][40][41].
Memory-like ques of neutrophils have been reported previously, showing that the recruitment and capacity to neutralize pathogens can be primed by IL-8, TNF-α, or microbiome-derived metabolites [42][43][44][45][46]. Moreover, primed neutrophils may drive anti-tumoral reactions as well as pronounced antimicrobial activity upon a secondary exposure [14,47]. Recent work from our group showed that LPS-primed neutrophils in vitro promote opposing memory-like inflammatory responses in a dosedependent manner characterized by altered cytokine patterns as well as ROS production resulting with distinct changes in migratory as well as phagocytic activities [13]. It is not known however, whether similar patterns also exist for cell components of gram-positive bacteria, especially for lipoteichoic acid (LTA). LTA is the main cell membrane constituent of gram-positive bacteria and a recognized immunomodulator that plays an important role in bacterial growth and virulence [48]. LTA is a strong activator of immune cells, triggering the release of inflammatory mediators which may result in septic shock and multiorgan failure [49,50]. However, the effects of LTA on the induction of memory-like responses remain elusive.
The current work describes the role of LTA-priming promoting distinct memory-like inflammatory reactions, trained sensitivity and tolerance in a dose-dependent manner in murine bone marrow neutrophils in vitro. In vitro, low-dose LTA-primed neutrophils with a secondary LPS stimulus exhibit elevated levels of pro-inflammatory mediators (trained sensitivity) outlined by elevated TNF-α, IL-6, ROS and IL-1β. In contrast, high-dose LTA-primed neutrophils trigger an immune-suppressive phenotype characterized by diminished pro-inflammatory responses and high IL-10 production. Further, trained neutrophils show enhanced migration, whereas the opposite effect can be seen for tolerant neutrophils in vitro. This study further expands our previous findings on the role of PAMPs in priming neutrophil function. Both LPS and LTA as integral components of bacteria can induce inflammatory and anti-inflammatory phenotypes in neutrophils depending on sequence of exposure and dose. Our hypothesis that neutrophils have the capacity to be primed and exhibit a certain level of immune-plasticity is supported by the current study.

Isolation of murine bone marrow neutrophils
Murine bone marrow neutrophils were isolated from 3-6 months old C57Bl/6J locally inbred mice. All mice were housed in a central animal facility with standard care (12-hour day cycle, standard nutrition an water). All experiments were performed in accordance with the guidelines from Directive 2010/63/EU of the EU Parliament on the protection of animals used for scientific purposes and approved by the Regierungspräsidium Karlsruhe, Germany (Az T-02/20).

Neutrophil stimulation procedure
1.000.000 cells/well were treated immediately after isolation as published before and shown in Figure 1A [13]. Cells were challenged twice, first by priming with increasing doses of S. aureus LTA (100 pg/mL -10µg/mL) (InvivoGen (#tlrl-pslta, Toulouse, France) for 45 minutes at 37°C and 5% CO2, followed by two medium changes. On day two, neutrophils we re-stimulated with the standard dose of 100ng/ml E. coli 055:B5 LPS (InvivoGen (#tlrl-pb5lps, Toulouse, France) and were left at 37°C and 5% CO2 for 4h. The supernatant, protein lysates and RNS were collected. Unstimulated cells (US) and unprimed cells (only exposed to the secondary stimulus with LPS) served as controls.

Protein concentration measurements
Protein concentrations were measured using the Pierce™ 660 nm Kit (#22662, Thermo Fisher Scientific, Waltham, MA, USA) following manufacturer's instructions. Ionic detergent compatibility reagent (IDCR) (#22663, Thermo Fischer Scientific, Waltham, MA, USA) was used in order to increase the detergent compatibility and reduce interference. In brief, assay reagent containing ionic detergent compatibility reagent (IDCR; #22663, Thermo Fischer Scientific, Waltham, MA, USA) was added to standards, samples and blanks, which were previously pipetted on a 96-well plate. Absorbance was measured at 660 nm and protein concentration was then calculated according to manufacturer's instructions.

ROS production
ROS production in murine bone marrow neutrophils was assessed using the DCFDA/H2DCFDA Assay Kit (#ab113851, Abcam, Cambridge, UK) following manufacturer's instructions as described previously [13]. In brief, neutrophils were seeded (100.000 cells/well) into black clear bottom 96-well plates and treated according to experimental protocol. After removal of medium, 20 µM DCFDA solution was added for 30 min at 37°C. After two washing steps, cells were resuspended in supplementary buffer provided in the kit. Intracellular ROS was quantified by fluorescence spectroscopy at 485/ 535 nm using a Wallac Victor3 1420 plate reader (PerkinElmer, Turku, Finland).

In vitro transmigration assay
Transmigration of neutrophils was assessed using an assay kit (#CBA-104, Cell Biolabs, CA, USA) as described previously and according to manufacturer's instructions [13]. Neutrophils from respective experimental conditions were placed into the 96-well membrane chamber, which was then placed in the feeder tray containing 10% FCS-supplemented medium for 5h at 37°C and 5% CO2. Cells were dislodged using the kit cell detachment solution for 30min. 150 µL of cell solution were transferred to a new 96-well plate, and incubated with 50 µL 4x lysis buffer/CyQuant GR dye solution for 20 min at room temperature. Fluorescence was measured with the same settings as for ROS production measurements using the Wallac Victor3 plate reader (results expressed as relative fluorescence units, RFU).

Cell viability and cytotoxicity
MTT assay was used to determine cell viability. Neutrophils (200.000 cells/well) were treated as per experimental protocol in a 96-well plate. After the addition of 10 µL 0,5 mg/mL MTT per well, samples were incubated for 4h, solubilization solution was added followed by incubation overnight all at 37°C and 5% CO2. Absorbance was measured at 570 nm. Data are expressed as relative viability compared to unprimed (UP) (Supplementary Figure 1A). Cytotoxicity of the experimental treatment was assessed using a cell cytotoxicity assay kit (colorimetric; #ab112118, Abcam, Cambridge, UK) according to manufacturer's instructions as described previously [13]. The absorbance was measured at 570 nm and data are expressed as relative cytotoxicity compared to unprimed (UP) (Supplementary Figure 1B).

RNA isolation and real-time PCR
RNA extraction was performed using QIAzol (#79306, Qiagen, Hilden, Germany) as described before [55]. RNA concentration and quality were assessed by Nanodrop (Novix, DE, USA) measurements. cDNA synthesis was performed using a kit from Thermo Fisher (#K1612, Waltham, MA, USA) and RT qPCR was run on a StepOnePlusTM RT PCR System (Applied Biosystems, USA). Gene expression was normalized to GAPDH expression as an internal house-keeping control and deloged as 2 -ΔΔC T [56]. Primer paris are listed below:

Statistical analysis
GraphPad Prism (version 8.0.2, GraphPad Software, CA, USA) was used for graphical display. All statistics were carried out using SigmaPlot (version 12.0, Systat Software GmbH, Erkrath, Germany). Shapiro-Wilk normality test was run prior to statistical analysis. Multi-group comparisons were made by one-way ANOVA followed by Holm-Sidak posthoc analysis. Significance was set at p < 0.05. Data are presented both as individual data points and mean + SEM.

LTA-priming induces dose-dependent inflammatory responses in murine bone marrow neutrophils
To better understand the LTA-mediated inflammatory capacity of neutrophils, we quantified the inflammatory output (TNF-α and IL-6 production) in response to rising concentrations of LTA ( Figure 1B, C).
After re-challenge with a fixed dose of 100ng/mL LPS, we found increased levels of cytokine production in low-dose (1 ng/mL) LTAprimed neutrophils, whereas priming with  1 µg/mL LTA diminished levels of TNF-α and IL-6 production ( Figure 1D, E). We then analyzed the production of other pro-inflammatory mediators like ROS and IL-1β, as well as anti-inflammatory IL-10 cytokine production and its expression (Figure 2). We found that both ROS and IL-1β are significantly increased in low-dose (1 ng/mL) LTAprimed neutrophils displaying trained sensitivity reactions, whereas priming by a higher dose (10 µg/mL) LTA-primed cells showed diminished levels after LPS challenge revealing a level of immune tolerance (Figure 2A, B). In contrast, IL-10 production was not affected by priming with low doses of LTA, however priming with high doses of LTA led to increased levels of IL-10 ( Figure 2C, D). Figure 2. LTA-priming effects on pro-and anti-inflammatory mediators in murine bone marrow neutrophils. LTA-primed neutrophils (1 ng/mL = low-dose; 10 µg/mL= high-dose) were re-stimulated on day 2 with 100 ng/mL LPS as described above. The levels of (A) ROS (n=6) were measured by DCFDA assay and (C) IL-10 (n=8-9) cytokine production assessed by ELISA (normalized to total protein concentration), whereas gene expression of (B) IL-1β (n=4) and (D) IL-10 (n=4) was analyzed by RT-PCR and expression normalized to unprimed, UP, set as 1. Individual data points and mean + SEM are shown. # depicts significant differences to unprimed condition (UP, gray column). p < 0.05 was considered statistically significant.
Cell viability was not affected during the experimental procedure as verified by viability and cytotoxicity assays (Supplementary Figure 1). Taken together our data show that LTA priming of murine bone marrow neutrophils may induce opposing inflammatory responses upon a second challenge by LPS, where low-dose priming promotes trained sensitivity and high-dose priming drives tolerance reactions.

TLR2/PI3K/MAPK activation of p65 regulates LTA-priming effects
Next, we examined the potential signalling proteins which could help understand these antagonistic inflammatory reactions upon LTApriming. Trained sensitivity is characterized by enhanced protein expression of TLR2/MyD88/p65, distinct from the tolerant state of neutrophils outlined by its downregulation (Figure 3A, B, E). In line with this observation, activation of protein kinase B (Akt) as a key downstream mediator of phosphoinositide 3-kinases (PI3Ks), as well as extracellular signal-regulated kinase (ERK) 1/2 as an important member of the mitogen-activated protein kinase (MAPK)-mediated pro-inflammatory pathway were similarly regulated in murine bone marrow neutrophils ( Figure 3C, D). Trained neutrophils exhibited enhanced expression levels of Akt/ Erk1/2, distinct from tolerant neutrophils that showed diminished protein expression of Akt. Figure 3. LTA-priming alters TLR2/PI3K/MAPK/p65 signalling patterns in murine bone marrow neutrophils. LTA-primed neutrophils (1 ng/mL = low-dose; 10 µg/mL= high-dose) were re-stimulated on day 2 with 100 ng/mL LPS as described above. Lysates were collected 4h after LPS re-challenge and the protein expression of (A) TLR2 (n=6), (B) MyD88 (n=5), (C) phospho-Akt (n=6) and (D) phospho-Erk1/2 (n=6) were analyzed by Western blotting and quantified (unprimed neutrophils assigned as 1.0). Gene expression of p65 (E, n=4) was assessed by RT-PCR and normalized to unprimed, UP, set as 1. Individual data points and mean + SEM are shown. # depicts significant differences to unprimed condition (UP, gray column). p < 0.05 was considered statistically significant.
Taken together, our data show that the two opposing states of primed neutrophils correlate with distinct regulation of NFκB-p65 by TLR2/PI3K/MAPK.

LTA-primed inflammatory responses reshape the migratory capacity of murine bone marrow neutrophils in vitro
In order to validate that the cellular processes observed have meaning in regard to cellular immune function, we assessed neutrophil transmigration in vitro. Murine bone marrow neutrophils primed by 1 ng/mL LTA display enhanced migratory effects characterized by increased production of C-X-C motif ligand 1 (CXCL1) chemokine as one of the crucial ligands required for neutrophil trans-endothelial migration ( Figure 4A, B). Antagonistic features were found in high-dose (10 µg/mL) LTA-primed neutrophils, where both in vitro migration and CXCL1 production were dampened ( Figure 4A, B). In parallel, we analyzed the expression of CD11a as an integral component of neutrophil recruitment and found comparable dose-effects ( Figure 4C). Our findings suggest that the cellular processes of dose-dependent neutrophil priming neutrophils correlate with cellular function. LTA-primed neutrophils (1 ng/mL = low-dose; 10 µg/mL= high-dose) were re-stimulated on day 2 with 100 ng/mL LPS as described above. (A) Migration (n=6) of murine neutrophils was assessed by migration assay. Chemokine production of (B) CXCL1 (n=9) was measured by ELISA (normalized to total protein concentration), whereas gene expression of (C) CD11a (n=4) was assessed by RT-PCR and normalized to unprimed, UP, set as 1. Individual data points and mean + SEM are shown. # depicts significant differences to unprimed condition (UP, gray column). p < 0.05 was considered statistically significant.

Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 1 December 2021
In recent years the dichotomy concept of innate and adaptive immunity witnessed a paradigm shift. Innate immune cells like monocytes, macrophages, NK cells and microglia are now recognized to exhibit non-specific memory-like (adaptive) reactions in response to reinfections. Both memory-like states of the innate immune cells, training as well as tolerance, are accompanied by epigenetic modifications occurring in H3 or H4 histones and regulated by different immunometabolic modulations promoting either an increased responsiveness or diminished responses against secondary infections [29,32,[39][40][41][57][58][59]. Different studies revealed pathogen-specificity, dose-dependency, magnitude of priming as well as the maturation state as crucial factors affecting the development of memory-like responses in innate immune cells [23,24,27,28,[31][32][33]60]. However, there are still many fundamental questions to be addressed, for example which spectrum of bacterial, fungal and viral components or other stressing agents may trigger such memory-like responses [61,62].
Terminally differentiated neutrophils represent an essential part of the innate immune system involved in early defense against infectious stressors [8,63]. They possess a variety of antimicrobial agents like reactive oxygen species (ROS), pro-inflammatory cytokines (i.e., TNF-α, IL-6, IL-1β), chemokines (i.e., CXCL1, MCP-1) and hydrolytic enzymes to protect against infectious agents, as well as anti-inflammatory mediators such as IL-10 to promote repairing mechanisms preserving the host from an exaggerated inflammatory response [6,7,[64][65][66]. Recently, Kalafati et al. demonstrated in vivo that trained neutrophils during granulopoiesis (central trained immunity) promote anti-tumoral effects in a ROSdependent fashion [14,26]. In line with this, our recent published work showed that murine neutrophils primed by gram-negative LPS in vitro display opposing memory-like inflammatory responses, trained sensitivity and tolerance in a dose-response manner [13]. Several studies in other myeloid cells also revealed dose-dependent induction of memory-like inflammatory responses imprinted by LPS [24,33,34,67]. However, there have been no studies investigating the role of grampositive LTA-priming in the induction of memory-like inflammatory responses in neutrophils.
The present study firstly describes the role of gram-positive LTA in the induction of mechanisms of immune memory in murine neutrophils. LTA-primed bone marrow neutrophils that have been re-challenged by LPS, displayed a biphasic response, where low-dose LTA priming triggers increased inflammatory responsiveness, and high-dose LTApriming results in tolerance reactions.
To better understand the biological function of these regulatory mechanisms we not only screened for classical pro-inflammatory cytokines such as TNF-α and IL-6, but also addressed associated crucial inflammatory mediators such as ROS and IL-1β. Interestingly, levels of IL-6 greatly exceeded TNF-α production, which may be explained by the capacity of endogenous TNF-α to boost IL-6 production by affecting chromatin modulation of IL-6 regulatory genes [68][69][70]. Trained neutrophils exhibited enhanced levels of ROS production associated with increased expression of IL-1β enhancing neutrophil recruitment. Contrary, tolerant neutrophils suppress the levels of ROS and IL-1β and supported the release of anti-inflammatory IL-10. In line with these findings, IL-1 cytokines (especially IL-1β) are known to promote neutrophil recruitment and, together with ROS, to endorse antimicrobial defence [71][72][73][74]. In a different context, Kalafati et al. outlined the key importance of ROS production by trained neutrophils in vivo displaying anti-tumoral effects [14]. Increased IL-10 levels by tolerant neutrophils are of crucial importance and of dual benefit since IL-10 promotes repairing mechanisms by neutrophils and protects the host against excessive inflammatory reactions [65,71,[75][76][77]. LTA-priming at low doses did not affect IL-10 production which might be due to the lack of B lymphocyte-induced maturation protein-1 (Blimp-1) in vitro, which generally promotes IL-10 production [33]. In should generally be noted that when addressing individual mediators or cytokine networks, it is hard to impossible to replicate in vivo interactions in vitro. This is not an inherent limitation to this study, but it has to be acknowledged that the importance of this novel regulatory mechanism in a living organism remains to be proven.
Several reports have demonstrated the role of pathogen dose provoking opposing memory-like responses in monocytes, macrophages, microglia as well as the hematopoietic stem cells [24,33,34,78,79]. LTA is the main constituent of the cell wall of gram-positive bacteria that acts through its sensing receptor TLR2 driving the downstream activation of NFκB [80]. Although gram-negative LPS and gram-positive LTA act by comparable signalling pathways through the activation of TLRs, they have greater differences between them related to the structural properties that initiate the injury and host responses that follow [81] [13,24,29,67,82,83]. Analogous findings have been reported for the downstream mediator of PI3Ks, Akt, as well as for the Erk1/2 as an important representative of MAPKs family for inflammatory responses.
The essential contribution of TLR/MyD88 pathway induced by microbial agents regulating neutrophil recruitment has been reported previously [84][85][86]. Our data reveal that trained neutrophils by low-dose LTA-priming exhibiting an increased pro-inflammatory status are characterized by increased migratory activities in vitro. The promigratory state is supported by enhanced levels of CXCL1 as well as by increased expression of CD11a (the α chain of LFA-1). CXCL1 is known to be an important chemokine for host protection by recruiting neutrophils to the site of infection improving also the killing activities, whereas CD11a serves as a key molecule in the neutrophil recruitment cascade [4,[87][88][89][90][91]. In contrast, high-dose LTA-primed tolerant neutrophils demonstrated diminished migratory affinities characterized by suppressed levels of CXCL1 and CD11a.
Both, training and tolerance, may serve host protection against bacterial pathogens or against an excessive inflammatory response [16,92]. The biological relevance of our finding on the LTA-mediated priming capacity of neutrophils is however not yet clear. There are obvious universal patterns when comparing the effect of LTA and LPS on neutrophil priming and we speculate that there are far more agents that may induce memory-like effects in neutrophils. In the current work we used a cross-over setup with LTA-priming followed by an LPS-hit, suggesting that the level of regulation in priming is rather downstream of toll-like receptors. In vivo, immune cells are exposed to large quantities of bacterial metabolites and toxins from the microbiota, especially the gut microbiome. It is therefore possible that a finger print of different bacterial ques may directly modulate neutrophil function and plasticity. Translational deductions need to be drawn with caution, as expression patterns, especially the signalling of cytokine and chemokines are known to greatly differ between species and cell populations [93][94][95]. Nevertheless, the identification of a novel way to switch between adaptive inflammatory responses may be of interest to the evolving field of adaptive innate immunity.

Conclusions
Collectively, our data suggest that gram-positive LTA may trigger dose-dependent induction of memory-like responses in murine bone marrow neutrophils in vitro. Low-dose LTA-primed neutrophils trigger trained sensitivity outlined by increased pro-inflammatory mediators, different from tolerant neutrophils marked by diminished proinflammatory reactions and pronounced anti-inflammatory responses. On a cellular level these effects translate to altered neutrophil recruitment in vitro. In light of previous findings at our lab it seems increasingly likely that neutrophils bear limited immune-plasticity and are shaped by their surroundings. The biological relevance of this observation will be the focus of future studies.
Data Availability: The data presented in this study are available on request from the corresponding author upon reasonable request.

Conflicts of Interest:
The authors declare that there is no conflict of interest regarding the publication of this paper. The funders had no role in the preparation of the manuscript or in the decision to publish the results.