Anaplasma phagocytophilum Induces TLR- and MyD88-Dependent Signaling in In Vitro Generated Murine Neutrophils

Anaplasma phagocytophilum is a tick-transmitted obligate intracellular Gram-negative bacterium that replicates in neutrophils. It elicits febrile disease in humans and in animals. In a mouse model, elimination of A. phagocytophilum required CD4+ T cells, but was independent of IFN-γ and other classical antibacterial effector mechanisms. Further, mice deficient for immune recognition and signaling via Toll-like receptor (TLR) 2, TLR4 or MyD88 were unimpaired in pathogen control. In contrast, animals lacking adaptor molecules of Nod-like receptors (NLR) such as RIP2 or ASC showed delayed clearance of A. phagocytophilum. In the present study, we investigated the contribution of further pattern recognition receptor (PRR) pathways to the control of A. phagocytophilum in vivo. Mice deficient for the NLR NOD2 had elevated bacterial loads in the early phase of infection, but were unimpaired in pathogen elimination. In contrast, animals lacking adaptor proteins of different C-type lectin receptors (CLR) such as DAP12, Fc-receptor γ-chain (FcRγ) and SYK controlled A. phagocytophilum as efficiently as wild-type mice. Further, we investigated which PRR pathways are involved in the sensing of A. phagocytophilum by in vitro generated Hoxb8 murine neutrophils. In vitro, recognition of A. phagocytophilum by murine neutrophils was dependent on TLR- and MyD88 signaling. However, it remained intact in the absence of the NLR NOD1, NOD2 and NALP3 and of the CLR adaptor molecules DAP12 and FcRγ. From these results, we conclude that TLR rather than NLR or CLR are critical for the detection of A. phagocytophilum by neutrophils although in vivo defective TLR-signaling is compensated probably because of the redundancy of the immune system.

The striking tropism of A. phagocytophilum for neutrophils has stimulated the interest in its immunological control which was studied primarily using the laboratory mouse as model organism. C57BL/6 wild-type (WT) mice developed transient bacteremia, but no clinical signs and eliminated A. phagocytophilum within 14 days post infection (p.i.) (von Loewenich et al., 2004). IFN-g-deficient mice had elevated bacterial loads in the early phase of infection, but were unimpaired in pathogen elimination (Akkoyunlu and Fikrig, 2000;Martin et al., 2000;Borjesson et al., 2002;Birkner et al., 2008). In A. phagocytophilum infected animals, NK cells (Birkner et al., 2008;Pedra et al., 2008), NKT cells (Pedra et al., 2008), and CD4 + T cells (Pedra et al., 2007a;Pedra et al., 2007b) have been described as source of IFN-g. NK cell derived IFN-g was probably induced by type I IFN and IL-12 (Birkner et al., 2008). IFN-g production of CD4 + T cells was IL-12 and IL-18-dependent (Pedra et al., 2007a;Pedra et al., 2007b). In contrast to the early phase of infection, bacterial clearance was independent of IFN-g, but strictly required CD4 + T cells (Birkner et al., 2008). In the absence of MHC class II-restricted CD4 + T cells, mice were persistently infected at least until day 77 p.i. (Birkner et al., 2008). However, the effector mechanism underlying the CD4 + T celldependent control of A. phagocytophilum still remains to be determined.
A number of in vivo studies analyzed the pattern recognition receptor (PRR) pathways that are triggered by A. p h a g o c y t o p h i l u m. In m i c e, t h e e li m in a t io n of A . phagocytophilum was independent of Toll-like receptor (TLR) 2, TLR4 and MyD88 (von Loewenich et al., 2004;Pedra et al., 2007b), the main adaptor molecule of the TLR (Takeda and Akira, 2004). Given the apparent lack of TLR involvement in vivo, Nod-like receptor (NLR) pathways (Kim et al., 2016) have been investigated. Mice deficient for RIP2 (RIPK2), the adaptor molecule of the NLR NOD1 and NOD2, showed delayed clearance of A. phagocytophilum lasting until day 20 p.i. (Sukumaran et al., 2012). Other NLR such as NALP3 (NLRP3) and NLRC4 (IPAF) associate with the adaptor molecule ASC (Kim et al., 2016). Mice deficient for NLRC4 or ASC had elevated bacterial loads in the early phase of infection but were able to finally eliminate the pathogen (Pedra et al., 2007a). In contrast, the course of infection in NALP3 -/animals did not differ significantly from WT mice (Pedra et al., 2007a). Thus, longterm persistence as in CD4 + T cell-deficient mice was not observed in animals lacking key components of TLR and NLR signaling.
In vivo, the deletion of a defined PRR might not have an overall effect on pathogen load because of compensatory activity of other PRR. However, the signaling in one particular cell type such as the neutrophil could be affected which can be studied ex vivo or in vitro. In this context, the immortalization of murine hematopoietic progenitor cells by conditional expression of estrogen-regulated Hoxb8 allowed the production of large amounts of neutrophils (Wang et al., 2006). The infection of murine in vitro generated Hoxb8 neutrophils with A. phagocytophilum induced the expression of inducible or type 2 nitric oxide synthase (iNOS or NOS2) mRNA and the secretion of significant amounts of MIP-1a (CCL3), RANTES (CCL5), and TNF (Gussmann et al., 2017). Further, bacterial growth was significantly impaired in IFN-g-stimulated Hoxb8 neutrophils (Gussmann et al., 2017). Thus, neutrophils seem to recognize A. phagocytophilum and to contribute to its control.
In the present study, we first investigated in vivo the role of further PRR pathways including signaling via C-type lectin receptors (CLR) (Brown et al., 2018) for the control of A. phagocytophilum. Secondly, we analyzed in vitro the ability of Hoxb8 murine neutrophils with defect for different PRR pathways to sense A. phagocytophilum. We found here that NOD2 -/mice had elevated bacterial loads in the early phase of infection, but were unimpaired in pathogen elimination. In contrast, animals lacking the CLR adaptor proteins DAP12, Fcreceptor g-chain (FcRg), or SYK controlled A. phagocytophilum as efficiently as WT mice. In vitro, we observed that A. phagocytophilum might be recognized by neutrophils in a TLRand MyD88-dependent manner. However, the NLR NOD1, NOD2 and NALP3 and the CLR adaptor proteins DAP12 and FcRg were not involved in sensing of A. phagocytophilum by neutrophils.

In Vivo Infection Experiments
The A. phagocytophilum Webster strain (Asanovich et al., 1997) was maintained through continuous passage in infected C57BL/6 TCRb -/-TCRd -/mice and used for mouse infection experiments as reported previously (von Loewenich et al., 2004). In general, groups of three mice were sacrificed at different time points after infection. EDTA-anticoagulated blood, spleen and lung were collected from each animal. Samples were subjected individually to DNA preparation and quantitative PCR analysis as reported previously (von Loewenich et al., 2004). The bacterial load was calculated as copies A. phagocytophilum/ copy murine glucose-6-phosphate dehydrogenase × 10 −3 .

Hoxb8 Neutrophils
Female mice were used at the age of 8-12 weeks. Progenitor cells were derived from murine bone marrow. The progenitor cells were retrovirally transduced with estrogen-regulated Hoxb8 and selected for 4 weeks in the presence of stem cell factor (SCF) to generate neutrophil progenitor cell lines (Wang et al., 2006). Polyclonal progenitor cell lines were cultured in Opti-MEM + GlutaMAX medium (Life Technologies, Darmstadt, Germany) supplemented with 10% FCS, 30 µM ß-mercaptoethanol, 1 µM ß-estradiol (Sigma-Aldrich, Taufkirchen, Germany) and 1% supernatant from SCF-producing CHO cells. The SCF producing cell line was kindly provided by Hans Häcker (St. Jude Children's Research Hospital, Memphis, TN, USA). Differentiation was induced by ß-estradiol removal. The differentiation status of the cells was controlled for each experiment microscopically by Giemsa staining and was similar for WT and gene-deficient Hoxb8 neutrophils. MyD88 -/-TRIF -/-Hoxb8 neutrophils were a gift from Hans Häcker.

In Vitro Experiments
The A. phagocytophilum Webster strain (Asanovich et al., 1997) was routinely grown in differentiated Hoxb8 neutrophils and was passaged every 3 to 4 days. To determine the percentage of infected cells, cells were cyto-centrifuged onto glass slides using a Cytospin 4 centrifuge (ThermoFisher Scientific, Langenselbold, Germany) and stained by Diff-Quick (Dade Behring, Marburg, Germany). 200 cells were counted at 1,000-fold magnification. Host-cell free A. phagocytophilum obtained from 3 x 10 7 Hoxb8 neutrophils with an infection rate of 90% was used to infect 1.2 x 10 7 Hoxb8 neutrophils (differentiated for 4 days) in 6 ml medium. To separate A. phagocytophilum from its host cells, the infected Hoxb8 neutrophils were passaged 10 x through a 27 G needle. Subsequently, a differential centrifugation step (10 min 750 g, 10 min 2,300 g) was performed and the pellet used for the infection. Pellets prepared from 3 x 10 7 uninfected Hoxb8 neutrophils served as control stimuli. At the time points 0, 24, 48, 72, and 96 h 500 µl from each set of samples were collected. The pellet was re-suspended in RNAlater (Life Technologies) and stored together with the supernatant at -80°C. Depending on the experiment, cells were stimulated with 200 ng/ml Escherichia coli K12 D31m4 (Re) LPS (List Biologicals, Campbell, CA, USA), 5 µg/ml imiquimod (InvivoGen, Toulouse, France), 5 µM ODN 1826 (InvivoGen), 10 µg/ml iE-DAP (InvivoGen), 10 µg/ml MDP (InvivoGen), or 100 µg/ml HKMT (InvivoGen). Total RNA was prepared using TRIzol (Life Technologies), treated with TURBO DNase (Life Technologies) and reverse transcribed with the High Capacity cDNA Reverse Transcription Kit (Life Technologies). Quantitative PCR was performed on an ABI Prism 7900HT Sequence Detector (Life Technologies) using TaqMan Gene Expression Master Mix (Life Technologies) and the following assays: iNOS (Mm00440485_m1) and HPRT (Mm00446968_m1). To follow the growth of A. phagocytophilum in Hoxb8 neutrophils, the bacterial RNA was quantified using primers 16S RTf2 (GAG AGT TTG ATC CTG GCT CAG AA) and 16S RTr (GCT ATA AAG AAT AAT CCG TTC GAC TTG) and the 16S RT probe (Fam-ACG CTG GCG GCA AGC TTA ACA CAT-BHQ1). Respective mRNA amounts were normalized to murine hypoxanthine guanine phosphoribosyl transferase 1 (HPRT) levels. Relative mRNA expression was calculated using the DDC t -method. Levels of murine MIP-1a (CCL3), RANTES (CCL5), TNF, IL-6, IL-1b, and KC (CXCL1) were measured in the supernatants using CBA Flex Sets (BD Biosciences, Heidelberg, Germany) and a BD LSRFortessa instrument (BD Biosciences). The analysis was performed applying the FCAP array software (BD Biosciences).

Statistical Analysis
Differences between experimental groups were analyzed using the two-tailed Mann-Whitney test. Calculations were done by GraphPad Prism 6.07. A p value <0.05 was considered significant. A correction for multiple testing was not done.

Innate Immune Recognition of A. phagocytophilum In Vivo
In vivo, signaling via MyD88 was not crucial for the elimination of A. phagocytophilum (von Loewenich et al., 2004;Pedra et al., 2007b). We therefore investigated whether MyD88independent TLR-signaling via the TRIF pathway (Ullah et al., 2016) could be involved. C57BL/6 TRIF Lps2/Lps2 (Hoebe et al., 2003) and respective WT mice were infected with A. phagocytophilum. Bacterial loads were determined in blood, spleen and lung at days 3, 7, and 14 p.i. Statistically significant differences between the groups were not found except for day 3 p.i. in blood (p < 0.01) ( Figure 1A). This indicates that TLRsignaling via TRIF is not essential for the control of A. phagocytophilum in vivo.
Mice defective for the intracellular NLR NOD2 (Mukherjee et al., 2019) and for adaptor proteins of different CLR such as DAP12, FcRg and SYK (Brown et al., 2018) were used to study non-TLR PRR pathways. Mice with SYK-deficient myeloid cells were generated by crossing C57BL/6 SYK flox/flox and A B D E C FIGURE 1 | Mice were infected i.p. with A. phagocytophilum. Blood, spleen and lung were collected at the indicated day p.i. and their bacterial load was measured by quantitative PCR. Each data point stands for one individual mouse. The bars represent the median. Data from 2 (A, C, D, E) and 3 (B) independent experiments are shown, respectively. Differences between gene-deficient and control mice were analyzed using the two-tailed Mann-Whitney test.

Müller et al.
Anaplasma phagocytophilum Signals via TLR and MyD88 Frontiers in Cellular and Infection Microbiology | www.frontiersin.org March 2021 | Volume 11 | Article 627630 C57BL/6 LysM cre/cre mice because the constitutive deletion of SYK is lethal in mice (Cheng et al., 1995;Turner et al., 1995). NOD2 -/mice had elevated bacterial loads compared to WT mice in blood (p < 0.01) and lung (p < 0.05) at days 3 and 7 p.i., but were unimpaired in pathogen elimination ( Figure 1B). In contrast, DAP12 -/mice ( Figure 1C), FcRg -/mice ( Figure  1D) and mice with myeloid cells lacking SYK ( Figure 1E) were able to control A. phagocytophilum as efficiently as WT mice at all time points investigated. This indicates that signaling via NOD2 might be involved in immune recognition of A. phagocytophilum in vivo, whereas DAP12, FcRg, and SYK are not.
iNOS mRNA Induction in TLR-5-fold -/-, MyD88 -/-, and MyD88 -/-TRIF -/-Hoxb8 Neutrophils The infection of Hoxb8 neutrophils with A. phagocytophilum has been shown previously to induce iNOS mRNA (Gussmann et al., 2017). Therefore, this parameter was analyzed here. LPS was used as control stimulus. As shown in Figure 3A, iNOS mRNA expression was significantly induced in WT Hoxb8 neutrophils after A. phagocytophilum infection or LPS-stimulation at 24 h to 96 h p.i. compared to uninfected control cells (p < 0.05). In contrast, iNOS mRNA was not up-regulated upon infection or LPSstimulation in TLR-5-fold -/-( Figure 3B), MyD88 -/-( Figure  3C) and MyD88 -/-TRIF -/-( Figure 3D) Hoxb8 neutrophils compared to the respective uninfected or unstimulated control cells. However, the differences in iNOS mRNA expression between WT and gene-deficient A. phagocytophilum infected cells reached statistical significance at most, but not all time points ( Figure 3E). The same was true for LPS-stimulation ( Figure 3F).
Upon infection with A. phagocytophilum, a similar finding was made for TNF ( Figure 4C). For MIP-1a ( Figure 4A) and FIGURE 2 | Increase of A. phagocytophilum 16S rRNA relative to murine HPRT mRNA at different time points p.i. in WT, TLR-5-fold -/-, MyD88 -/-, and MyD88 -/-TRIF -/-Hoxb8 neutrophils. Results were normalized to the respective 0 h value of each sample using the DDC t -method. Mean and SD from five independent experiments are shown. Differences between experimental groups were analyzed using the two-tailed Mann-Whitney test. The following groups were compared: WT and the respective gene-deficient Hoxb8 neutrophils at each time point. *p < 0.05. Figure 4B) production, the differences between WT and gene-deficient Hoxb8 neutrophils were less uniform at the different time points after infection.

Conclusion
Together, these data suggest that in vitro A. phagocytophilum might be recognized by Hoxb8 neutrophils via MyD88dependent TLR-signaling.
Imiquimod induced iNOS mRNA expression in WT Hoxb8 cells ( Figure S3A). The iNOS mRNA expression in TLR7 -/-Hoxb8 neutrophils after imiquimod-stimulation was significantly lower than in WT cells at 24 and 48 h p.i. (p < 0.05) proving the specificity of the effect ( Figure S3B).
In contrast, ODN 1826 did not significantly induce iNOS mRNA expression in WT Hoxb8 cells ( Figure S3A). In line with that finding, there was no statistically significant difference between iNOS mRNA expression in WT and TLR9 -/-Hoxb8 neutrophils after stimulation with ODN 1826 ( Figure S3C). This indicates that TLR9 signaling alone is insufficient for iNOS induction in Hoxb8 WT neutrophils. LPS induced iNOS mRNA expression in WT Hoxb8 cells ( Figure S3A). However, iNOS mRNA expression in TRIF -/-Hoxb8 neutrophils was unaltered compared to WT cells after LPS-stimulation ( Figure S3D). This is explained best by the MyD88-dependent TLR4 signaling which is still active in TRIFdeficient myeloid cells (Takeda and Akira, 2004).
According to these data, MIP-1a production seems to be a suitable readout for TLR7-dependent signaling in Hoxb8 neutrophils. Correspondingly, MIP-1a, TNF and IL-6 were found appropriate to investigate immune stimulation via TLR9. For the TRIF-pathway RANTES was selected.
MIP-1a production upon infection with A. phagocytophilum was unaltered in TLR7 -/-Hoxb8 neutrophils compared to WT cells ( Figure 7A). FIGURE 6 | Relative iNOS mRNA expression normalized to murine HPRT in WT, TLR7 -/-, TLR9 -/and TRIF -/-Hoxb8 neutrophils at different time points after infection with A. phagocytophilum. Results were normalized to the 0 h value of uninfected WT cells using the DDC t -method. Mean and SD from five independent experiments are shown. Differences between experimental groups were analyzed using the two-tailed Mann-Whitney test. The following groups were compared: the respective A. phagocytophilum infected gene-deficient Hoxb8 neutrophils to WT control cells.
FIGURE 5 | Increase of A. phagocytophilum 16S rRNA relative to murine HPRT mRNA at different time points p.i. in WT, TLR7 -/-, TLR9 -/and TRIF -/-Hoxb8 neutrophils. Results were normalized to the respective 0 h value of each sample using the DDC t -method. Mean and SD from five independent experiments are shown. Differences between experimental groups were analyzed using the two-tailed Mann-Whitney test. The following groups were compared: WT and the respective gene-deficient Hoxb8 neutrophils at each time point. *p < 0.05, **p < 0.01.

Conclusion
We had identified lower iNOS mRNA expression and reduced MIP-1a production as readout for TLR7-dependent signaling in Hoxb8 neutrophils, but did not see an effect upon A. phagocytophilum infection (Figures 6, 7A). Therefore in vitro, we did not find evidence for immune recognition of A. phagocytophilum via TLR7.
MIP-1a, TNF, and IL-6 were the three parameters chosen for investigation of TLR9 activation. Significantly lower amounts in the supernatants of the gene-deficient Hoxb8 neutrophils were only found for MIP1-a and only at later time points of 48-96 h p.i. ( Figure 7A) Thus, TLR9-dependent recognition of A. phagocytophilum could not be convincingly proven.
In vivo, TLR7 -/and TLR9 -/mice were able to control A. phagocytophilum as efficiently as WT mice ( Figure S8).

NLR-Dependent Recognition of A. phagocytophilum in Neutrophils In Vitro
We decided to study the NLR pathway using as representatives NOD1, NOD2, and NALP3 (NLRP3), because in vivo mice with defective signaling via TLR were still able to control A. phagocytophilum as efficiently as WT animals (von Loewenich et al., 2004;Pedra et al., 2007b).
Chemokine and Cytokine Secretion of NOD1 -/-, NOD2 -/-, and NALP3 -/-Hoxb8 Neutrophils MIP-1a, RANTES, KC, TNF, IL-6, and IL-1b were measured in the supernatants of gene-deficient and WT control Hoxb8 neutrophils upon stimulation with iE-DAP or MDP. However, iE-DAP and MDP did not induce the expression of any of the chemokines or cytokines tested (data not shown).
The production of MIP-1a ( Figure 9A), RANTES ( Figure  9B) and TNF ( Figure 9C) was not significantly lower in the phagocytophilum infection. MIP-1a, RANTES, TNF and IL-6 were measured in the supernatants using CBA assay. Mean and SD from five independent experiments are shown. Differences between the respective gene-deficient Hoxb8 neutrophils and WT control cells at each time point were analyzed using the two-tailed Mann-Whitney test. *p < 0.05, **p < 0.01.

gene-deficient Hoxb8 neutrophils than in WT control cells upon
A. phagocytophilum infection.

Conclusion
Together these data suggest that at least in Hoxb8 neutrophils signaling via NOD1, NOD2, and NALP3 seems not to be involved in immune recognition of A. phagocytophilum.

CLR-Dependent Recognition of A. phagocytophilum in Neutrophils In Vitro
Next, we investigated whether the adaptor/signaling proteins of different CLR such as DAP12, FcRg and SYK (Brown et al., 2018) were involved in the stimulation of Hoxb8 neutrophils by A. phagocytophilum. SYK-deficient polyclonal progenitor cell lines were severely impaired in growth. Therefore, SYK del/del-Hoxb8 neutrophils could not be produced in significant amounts and were not investigated further.

iNOS mRNA Induction in DAP12 -/and FcRg -/-Hoxb8 Neutrophils
It is not known for all CLR, which molecular pattern they bind. However, mycobacteria are recognized by some of them (Geijtenbeek and Gringhuis, 2009;Brown et al., 2018). Therefore, we used heat-killed mycobacteria (HKMT) as control stimulus and compared it to LPS.

Chemokine and Cytokine Secretion of DAP12 -/and FcRg -/-Hoxb8 Neutrophils
Significant higher amounts of MIP-1a ( Figure S10A), RANTES (Figure S10B), and TNF ( Figure S10C) were produced by WT Hoxb8 neutrophils after HKMT-and LPS-stimulation or A. phagocytophilum infection compared to medium controls at 24-96 h p.i. For IL-6, this was only found for treatment with HKMT or LPS, but not for infection with A. phagocytophilum ( Figure S10D).
In essence, the amount of MIP-1a, RANTES, TNF and IL-6 in the supernatants of DAP12 -/and FcRg -/-Hoxb8 neutrophils was not significantly different to WT control cells after stimulation with HKMT ( Figure S11) or LPS ( Figure S12). However, DAP12 -/-Hoxb8 neutrophils produced significantly more RANTES (Figures S11B, S12B) and IL-6 (Figures S11D, 12D) than WT cells after HKMT-or LPS-treatment at 24-96 h. A similar observation was made earlier for murine DAP12 -/bone marrow derived DC (Chu et al., 2008) and murine DAP12 -/murine bone marrow derived macrophages (Hamerman et al., 2005) that produced significantly higher amounts of IL-6 than WT control cells after LPS-stimulation which was interpreted as negative regulation of the TLRresponse by DAP12. Here, we show that this is also true for Hoxb8 neutrophils.

Conclusion
Together these data suggest that CLR signaling via DAP12 or FcRg seems not to be involved in immune recognition of A. phagocytophilum, because it did not grow significantly better in FIGURE 8 | Increase of A. phagocytophilum 16S rRNA relative to murine HPRT mRNA at different time points p.i. in WT, NOD1 -/-, NOD2 -/-, and NALP3 -/-Hoxb8 neutrophils. Results were normalized to the respective 0 h value of each sample using the DDC t -method. Mean and SD from four independent experiments are shown. Differences between experimental groups were analyzed using the two-tailed Mann-Whitney test. The following groups were compared: WT and the respective gene-deficient Hoxb8 neutrophils at each time point. *p < 0.05 **p < 0.01. the gene-deficient Hoxb8 neutrophils and because the chemokine and cytokine response was not significantly lower than in WT cells.

TLR-Dependent Recognition of A. phagocytophilum
Neutrophil granulocytes express a wide variety of PRR (Thomas and Schroder, 2013). In murine neutrophils, the expression of TLR1, TLR2, TLR4, TLR5, TLR7, and TLR9 has been reported (Thomas and Schroder, 2013). It has been shown primarily in macrophages that the activation of TLR2, TLR3, TLR4, TLR5, and TLR9 mediates iNOS induction and NO production (Abdul-Cader et al., 2016). We found here that iNOS mRNA in murine Hoxb8 neutrophils was induced in a TLR-and MyD88dependent manner after LPS-stimulation and in a TLR7dependent manner after imiquimod-stimulation. In contrast, iNOS expression was unaltered when Hox8 neutrophils were exposed to TLR9 agonist ODN 1826. Previously, it has been shown that the infection of murine bone marrow derived A B C FIGURE 9 | MIP-1a (A), RANTES (B) and TNF (C) production in WT, NOD1 -/-, NOD2 -/-, and NALP3 -/-Hoxb8 neutrophils at different time points after A. phagocytophilum infection. MIP-1a, RANTES and TNF were measured in the supernatants using CBA assay. Mean and SD from five independent experiments are shown. Differences between the respective gene-deficient Hoxb8 neutrophils and WT control cells at each time point were analyzed using the two-tailed Mann-Whitney test. *p < 0.05, **p < 0.01. neutrophils with Leishmania infantum induced iNOS mRNA expression and NO production in a TLR2-dependent manner (Sacramento et al., 2017). Further, zymosan, a cell wall component of yeasts, triggered iNOS protein synthesis and NO release of murine peripheral blood neutrophils via TLR2 and MyD88 (Li et al., 2016). To our knowledge, TLR7-dependent iNOS induction in neutrophils has not been shown before. However in murine monocytes, iNOS expression was partially TLR7-dependent after chikungunya virus infection (McCarthy et al., 2020).
In murine neutrophils, the expression of MIP-1a, RANTES, TNF, and IL-6 has been reported (Tecchio et al., 2014). We showed here that MIP-1a, RANTES and TNF were secreted by murine Hoxb8 neutrophils in a TLR-and MyD88-dependent manner after LPS-stimulation. MIP-1a production was induced via TLR7 after incubation with imiquimod and MIP-1a, TNF and IL-6 production via TLR9 after incubation with ODN 1826. Previously, it has been shown that secretion of TNF (Mohammadi et al., 2016), KC (CXCL1) and MIP-2 (CXCL2) (Lentini et al., 2020) by murine bone marrow derived neutrophils after stimulation with Streptococcus agalactiae was diminished in the absence of the endosomal TLR7, 9 and 13, whereas the lack of the individual TLR had no effect.
Here, we provide evidence that A. phagocytophilum is recognized via TLR-and MyD88-dependent signaling by neutrophils, because iNOS mRNA induction and TNF production were significantly lower in Hoxb8 neutrophils that lacked simultaneously TLR2, 3, 4, 7 and 9 as well as in Hoxb8 Results were normalized to the respective 0 h value of each sample using the DDC t -method. Mean and SD from five independent experiments are shown. Differences between experimental groups were analyzed using the two-tailed Mann-Whitney test. The following groups were compared: WT and the respective gene-deficient Hoxb8 neutrophils at each time point. *p < 0.05. , TNF (C) and IL-6 (D) production in WT, DAP12 -/and FcRg -/-Hoxb8 neutrophils at different time points after infection with A. phagocytophilum. MIP-1a, RANTES, TNF and IL-6 were measured in the supernatants using CBA assay. Mean and SD from five independent experiments are shown. Differences between the respective gene-deficient Hoxb8 neutrophils and WT control cells at each time point were analyzed using the two-tailed Mann-Whitney test. *p < 0.05, **p < 0.01. neutrophils singly defective for MyD88. A. phagocytophilum does not possess the established ligands for TLR3 (dsRNA) and TLR4 (LPS) (Lin and Rikihisa, 2003;Otten et al., 2018). Therefore, TLR7-and TLR9-dependent signaling were estimated as the most likely options and investigated using single gene-deficient cells. However, convincing evidence for recognition of A. phagocytophilum via TLR7 and TLR9 by Hoxb8 neutrophils was not provided. A reason for this could be that A. phagocytophilum expresses ligands for TLR7 and TLR9 and that the effect might be seen only in double deficient cells. In murine peritoneal macrophages, activation of NK-kB after stimulation with A. phagocytophilum was triggered via TLR2, but was independent of TLR4 (Choi et al., 2004 ). Thus, the effect we saw in Hoxb8 neutrophils defective for TLR2, 3, 4, 7 and 9 could be mediated by TLR2.
Ehrlichia chaffeensis and E. muris are obligate intracellular bacteria related to A. phagocytophilum that replicate in monocytes and macrophages (Dumler et al., 2001). After infection with E. chaffeensis, MIP-2 (CXCL2) and TNF mRNA induction in murine bone marrow derived macrophages was MyD88-dependent, but did not require TLR2 or TLR4 (Miura et al., 2011). Similarly, IL-12 production of bone marrow derived DC after stimulation with E. muris was reduced in the absence of MyD88 (Koh et al., 2010). However, the effect was independent of TLR2, 3, 4, 5, 7, 9, 11, IL-1 and IL-18 signaling. Maybe, as discussed above, the lack of one single TLR is compensated by the others. In contrast to these findings, IFN-g expression of murine spleen cells after incubation with E. muris required the antigenpresenting molecule CD1d, but was MyD88-independent (Mattner et al., 2005).
In vitro, the growth of A. phagocytophilum in Hoxb8 neutrophils defective for TLR2, 3, 4, 7, 9 and MyD88 was not essentially altered. This suggests that the pathogen escapes the immune reaction mediated by neutrophils without additional stimulation. Protection of A. phagocytophilum against reactive oxygen species has been shown before (Carlyon et al., 2004;IJdo and Mueller, 2004). However when Hoxb8 neutrophils were stimulated by IFN-g, the growth of A. phagocytophilum was significantly impaired (Gussmann et al., 2017). In vivo, mice deficient for TLR2, T L R 4 a n d M y D 8 8 w e r e f u l l y a b l e t o c o n t r o l A . phagocytophilum (von Loewenich et al., 2004;Pedra et al., 2007b). In contrast, mice deficient for the co-stimulatory molecule CD40 experienced persistent infection and the depletion of CD11c + DC delayed pathogen clearance (Birkner et al., 2008). Therefore, in antigen-presenting cells MyD88-independent PRR pathways might be involved in the recognition of A. phagocytophilum or in vivo the defect is compensated because of the redundancy of the immune system. In contrast to A. phagocytophilum, E. muris infected MyD88 -/mice had elevated bacterial loads in blood and spleen at days 10 and 14 p.i. compared to WT animals probably because of an impaired IFN-g response of CD4 + T cells following reduced IL-12 production by DC (Koh et al., 2010).

NLR-Dependent Recognition of
A. phagocytophilum NOD1 (Jeong et al., 2014), NOD2 (Jeong et al., 2014) and NALP3 (Guarda et al., 2011) protein expression has been shown in murine neutrophils. To our knowledge, no data exist whether NLR stimulation induces iNOS in neutrophils. Here, iNOS mRNA was not upregulated in Hoxb8 neutrophils after stimulation with the NOD agonists iE-DAP or MDP. However, MDP-stimulation of human peripheral blood mononuclear cells led to iNOS mRNA and protein expression in a NOD2dependent manner (Landes et al., 2015).
Here, MIP-1a, RANTES, KC, TNF, IL-6 and IL-1b were not secreted by murine Hoxb8 neutrophils after incubation with iE-DAP or MDP. Previously, it has been shown that murine thioglycollate-elicited peritoneal exudate neutrophils produced MCP-1 (CCL2), KC (CXCL1), TNF and IL-6 after MDP-stimulation in a NOD2-dependent manner (Jeong et al., 2014). This was not the case when NOD1 agonist Tri-DAP was used. Further, the effect was not observed in murine bone marrow derived neutrophils, potentially because of impurity of the thioglycollate-elicited peritoneal exsudate neutrophils. IL-1b secretion of murine bone marrow derived neutrophils was NALP3 (NLRP3)-dependent after combined stimulation with LPS and nigericin (Mankan et al., 2011). Therefore, the control stimulus used here might have been insufficient, although MDP has been described to activate the NALP3 inflammasome (Netea et al., 2010). Here, the production of MIP-1a, RANTES and TNF was unimpaired in NOD1-, NOD2-and NALP3-deficient Hoxb8 neutrophils after A. phagocytophilum infection. This means that these NLR are probably not involved in the recognition of A. phagocytophilum by neutrophils. However, the respective pathways might not be active in Hoxb8 neutrophils because a chemokine and cytokine response after stimulation with NOD agonists was not detected.
In vitro, growth of A. phagocytophilum in NOD1 -/-, NOD2 -/and NALP3 -/-Hoxb8 neutrophils was not better than in WT control cells. However in vivo, NOD2 -/mice had elevated bacterial loads in the early phase of infection compared to WT animals. This finding has to be interpreted with caution because the NOD2-deficient mice were on a C57BL/6J background in contrast to the C57BL/6N control animals. However, it is in line with the previous finding that mice deficient for RIP2 (RIPK2), the adaptor molecule of the NOD1 and NOD2, showed delayed clearance of A. phagocytophilum (Sukumaran et al., 2012). Therefore, it could be that other cell types than neutrophils recognize A. phagocytophilum via NOD2. In contrast, the course of infection in NALP3 -/animals did not differ significantly from WT mice (Pedra et al., 2007a).
Ligand binding leads to production of reactive oxygen species and cytokine secretion (Thomas and Schroder, 2013). To our knowledge, no data exist whether CLR stimulation induces iNOS in neutrophils. However, iNOS mRNA and NO production were induced in murine peritoneal macrophages by the CLR CD23 (FcϵRII) via FcRg-dependent signaling after a-mannanstimulation (Guo et al., 2018). Here, HKMT induced iNOS mRNA in WT, DAP12 -/and FcRg -/-Hoxb8 neutrophils. At least in FcRg -/-Hoxb8 neutrophils, an effect on iNOS induction could have been expected (Geijtenbeek and Gringhuis, 2009), but it might have been masked by TLR2-stimulation through HKMT (Underhill et al., 1999).
In essence, the amount of MIP-1a, RANTES, TNF and IL-6 in the supernatants of DAP12 -/and FcRg -/-Hoxb8 neutrophils was not significantly different to WT control cells after stimulation with HKMT. Other PRR ligands of HKMT might have mitigated an potential effect as discussed above. However, DAP12 -/-Hoxb8 neutrophils produced significantly more RANTES and IL-6 than WT cells after HKMT-or LPStreatment. A similar observation was made earlier for murine DAP12 -/bone marrow derived DC (Chu et al., 2008) and murine bone marrow derived DAP12 -/macrophages (Hamerman et al., 2005) that was interpreted as negative regulation of the TLR-response by DAP12. Here, we show that this is also true for Hoxb8 neutrophils.
iNOS mRNA induction and chemokine and cytokine production after A. phagocytophilum infection were unimpaired in DAP12-and FcRg-deficient Hoxb8 neutrophils. Further, bacterial growth in Hoxb8 neutrophils was not significantly better than in WT control cells. In vivo, bacterial loads in DAP12 -/mice, FcRg -/mice and mice with myeloid cells defective for SYK were able to control A. phagocytophilum as efficiently as WT mice. Thus, CLR signaling via DAP12, FcRg, and SYK might not be involved in the recognition of A. phagocytophilum. However, the in vitro data have to be interpreted with caution because we could not prove whether DAP12-or FcRg-dependent CLR signaling is active in Hoxb8 neutrophils.

CONCLUSION
We showed here that A. phagocytophilum is recognized by murine Hoxb8 neutrophils via TLR-and MyD88-dependent signaling. However, the TLR involved could not be reliably determined. It seems that the absence of TLR7 or TLR9 alone does not significantly impair stimulation of Hoxb8 neutrophils upon A. phagocytophilum infection. Probably, more than one TLR is involved. In vivo, the absence of MyD88 seems to be compensated because of the redundancy of the immune system. Hoxb8 neutrophils defective of the NLR and CLR pathways investigated were unimpaired in recognition of A. phagocytophilum. In other cell types than neutrophils NOD2 might contribute to immune stimulation, because in vivo NOD2 mice showed delayed clearance of A. phagocytophilum in the early phase of infection.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
The animal study was reviewed and approved by the animal welfare committee of the Regierungspräsidium Freiburg (G-06/ 19 and G-11/79).