The Properties of Proinflammatory Ly6Chi Monocytes Are Differentially Shaped by Parasitic and Bacterial Liver Infections

In the past, proinflammatory CD11b+Ly6Chi monocytes were predominantly considered as a uniform population. However, recent investigations suggests that this population is far more diverse than previously thought. For example, in mouse models of Entamoeba (E.) histolytica and Listeria (L.) monocytogenes liver infections, it was shown that their absence had opposite effects. In the former model, it ameliorated parasite-dependent liver injury, whereas in the listeria model it exacerbated liver pathology. Here, we analyzed Ly6Chi monocytes from the liver of both infection models at transcriptome, protein, and functional levels. Paralleled by E. histolytica- and L. monocytogenes-specific differences in recruitment-relevant chemokines, both infections induced accumulation of Ly6C+ monocytes at infection sites. Transcriptomic analysis revealed a high similarity between monocytes from naïve and parasite-infected mice and a clear proinflammatory phenotype of listeria-induced monocytes. This was further reflected by the upregulation of M2-related transcription factors (e.g., Mafb, Nr4a1, Fos) and higher CD14 expression by Ly6Chi monocytes in the E. histolytica infection model. In contrast, monocytes from the listeria infection model expressed M1-related transcription factors (e.g., Irf2, Mndal, Ifi204) and showed higher expression of CD38, CD74, and CD86, as well as higher ROS production. Taken together, proinflammatory Ly6Chi monocytes vary considerably depending on the causative pathogen. By using markers identified in the study, Ly6Chi monocytes can be further subdivided into different populations.


Introduction
Monocytes are a type of mononuclear phagocyte that, as cells of the innate immune system, are part of the initial immune response to invading pathogens [1]. Under homeostatic conditions, monocytes patrol the blood stream, replenish macrophage pools in tissues, and are recruited rapidly to sites of infection and inflammation [2,3]. Egress from the bone marrow is mediated mainly by C-C chemokine receptor 2 (CCR2), which binds to CCL2 secreted by cells in injured or infected tissue [4,5].
Once in the tissues, these cells shape the inflammatory milieu via expression of pro-or anti-inflammatory cytokines, phagocytic activity, and antigen presentation; they can also differentiate into macrophages [3,6]. However, monocytes can trigger immunopathology when inadequately controlled [7,8]. Murine monocytes are mostly identified as CD11b + Ly6C + Ly6G − cells and are commonly divided into two major subsets: proinflammatory CD11b + Ly6C hi and anti-inflammatory CD11b + Ly6C lo monocytes [9,10]. They can be further subdivided according to the expression of CCR2 and CX3CR1: Ly6C hi CCR2 hi CX3CR1 int (proinflammatory) and Ly6C lo CX3CR1 hi CCR2 -(anti-inflammatory), although the usage of CX3CR1 is presently questioned [11]. Fate mapping and single cell approaches revealed the priming of Ly6C hi monocytes towards a neutrophil-like monocyte (nMO) or dendritic cell-like monocyte (dcMO) phenotype under steady state conditions, and to Cxcl10 + and Saa3 + monocytes in pathogenic conditions [12][13][14]. In two murine models of liver infection, the absence of Ly6C hi monocytes results in opposite disease outcomes. In the murine model of hepatic amebiasis, intrahepatic infection with the protozoan parasite E. histolytica results (as in humans) in focal liver destruction (18). This type of liver damage, also termed amebic liver abscess (ALA), is almost abolished in mice with a Ccr2 knockout (Ccr2 -/-) or in mice in which monocyte were immunodepleted, suggesting an immunopathologic role for inflammatory monocytes (14). This liver damage also depends on inflammatory factors and can be inhibited by a specific blockade of TNF-α or by general immunosuppression [15,16]. By contrast, in the murine model of L. monocytogenes infection, monocytes play a protective role, as indicated by the higher bacterial load and increased granuloma formation in the liver of Ccr2 -/mice (18).
The aim of this study was to characterize liver Ly6C hi monocyte subsets from both infection models with regard to their histological localization within infected liver areas, their transcriptome and surface marker expression, and their production of reactive oxygen species (ROS). This study revealed functionally distinct inflammatory Ly6C hi monocytes in these two infection models and identified a novel combination of a transcription factor and surface markers that allow to distinguish proinflammatory monocyte subsets during inflammatory states.

Mice
All murine studies complied with relevant ethical regulations for animal testing and research. Animal experiments were performed in accordance with the German animal protection laws and were reviewed by the federal health authorities of the State of Hamburg in accordance with the ARRIVE guidelines (N082/2018). C57BL/6J (WT) and Ccr2 -/- [17] were bred and kept in individually ventilated cages under specific pathogen-free conditions at the animal facility at the Bernhard Nocht Institute for Tropical Medicine, Cd38 -/- [18] mice were kept in the facility of the University Medical Center Hamburg-Eppendorf. The mice were kept with a day/night cycle of 12 h, humidity of 50-60%, and a temperature of 21 • C. Mice were euthanized with CO 2 with a replacement rate of 20-30% of the cage volume per minute, followed by cervical dislocation and cardiac puncture.

Infection of Mice with E. histolytica and L. monocytogenes
Male C57BL/6 mice (aged 8-12 weeks) were used for infections. Briefly, 2 × 10 5 trophozoites of the highly pathogenic clone B2, generated from cell line B (HM-1:IMSS), were suspended in 20 µL of incomplete TYI-S-33 medium and injected into the left liver lobe, as described previously [19]. Abscess size was calculated as % abscessed left liver lobe. Mice were infected with 2 × 10 4 L. monocytogenes strain EGD or 1 × 10 7 L. monocytogenes ∆actA in 200 µL PBS via the lateral tail vein. Bacterial inoculi were controlled by plating serial dilutions on tryptic soy broth agar plates at 37 • C.

RNA Sequencing and Data Analysis
Hepatic monocytes of E. histolyticaor L. monocytogenes-infected C57BL/6 mice were sorted on the indicated days p.i., with a purity of 80-85% and surface staining was performed using CD11b (APC-Cy7; Ly6C (FITC) and Ly6G (APC). Cells were then sorted into collection tubes containing 2 mL of RNAprotect cell reagent (Qiagen). RNA was isolated using the RNeasy Plus Micro Kit (Qiagen) and RNA integrity was analyzed using an Agilent 6000 Pico Kit and an Agilent 2100 Bioanalyzer (Agilent). Samples used for transcriptome sequencing fulfilled the following criteria: total RNA ≥ 200 ng (≥20 ng/µL); RNA integrity number (RIN) ≥7.0, 28S/18S ≥ 1.0. RNA library preparation and sequencing were performed by BGI Genomics, China. The data comprised paired-end short reads. All raw data were aligned to the mouse reference genome GRCm38 Ensembl 85 and the corresponding Gencode annotation using STAR [20] (version 2.5.2a). Differential expression analysis was performed in R (version 3.3.3; R Foundation for Statistical Computing, Austria) using DESeq2 (version 1.13.8) [21]. Reads from different lanes per replicate were combined after checking for the absence of batch effects on a PCA plot. We performed differential gene expression analysis between both infection models within the Ly6C hi and Ly6C lo monocytes respectively using the day after infection as additional co-variate in the DESeq design. To test for differential expression across all three time points (naive, d3 p.i, d5 p.i.), we used a likelihood ratio test for each monocyte group using time point, infection model, and an interaction term for both variables as full design and compared it against the reduced model with the interaction term removed. A threshold of 0.05 for Benjamini-Hochberg corrected p values was used to determine significance by the "statistical overrepresentation test". Gene set analysis was performed using PANTHER GO-slim (Version 16) [22], with p adj < 0.05. Heatmaps were created using http://heatmapper.ca (accessed on 1 July 2019) combining time points 3 and 5. Volcano plots were made using GraphPad Prism V8.4.3.

Different Recruitment and
Localization of Ly6C hi CD11b + Monocytes in the Liver following Infection with E. histolytica or L. monocytogenes Proinflammatory Ly6C hi monocytes exhibit opposite functions in murine models for E. histolytica and L. monocytogenes liver infection. In the former, their absence resulted in the amelioration of parasite-dependent liver damage, whereas in the listeria model it exacerbated liver pathology [16,23]. To better understand the dynamics of monocyte recruitment, we examined hepatic protein concentrations of chemokines involved in these processes such as CCL2 and CCL3. We found a significant increase in CCL2 levels at d3, and an increase in CCL3 levels at d3 and d5 following parasitic infection ( Figure 1A). During L. monocytogenes infection, CCL2 levels were higher and increased already at d1 post infection (p.i.), while CCL3 levels were lower than during parasitic infection or in naïve mice ( Figure 1A). Additional cytokine analysis revealed significantly elevated expression of IL-1β, IL-10, and IL-13 during E. histolytica infection, as well as increased expression of TNF-α and IFN-γ during L. monocytogenes infection ( Figure S1A). an increase in CCL3 levels at d3 and d5 following parasitic infection ( Figure 1A). During L. monocytogenes infection, CCL2 levels were higher and increased already at d1 post infection (p.i.), while CCL3 levels were lower than during parasitic infection or in naïve mice ( Figure 1A). Additional cytokine analysis revealed significantly elevated expression of IL-1β, IL-10, and IL-13 during E. histolytica infection, as well as increased expression of TNFα and IFN-γ during L. monocytogenes infection ( Figure S1A). To determine monocyte recruitment, we isolated leukocytes from infected livers and measured the percentage of CD11b + Ly6C hi and CD11b + Ly6C lo monocytes by flow cytometry. At d3 p.i., the CD11b + Ly6C hi monocyte populations increased in both infection models, whereas the CD11b + Ly6C lo monocyte population increased only following E. histolytica infection ( Figure 1B; absolute numbers see Figure S1B,C). Staining of paraffin-embedded liver sections from both models (d3p.i.) with hematoxylin and eosin (H&E), anti-CD11b, and anti-Ly6C revealed accumulation of CD11b + and Ly6C + cells in a dense margin around the central amebic abscess ( Figure 1C) while CD11b + and Ly6C + cells accumulated in the center of typical L. monocytogenes-induced granulomas ( Figure 1C; controls see Figure S1D).
In both infection models, the increase in CCL2 expression led to an increase in the proportion of Ly6C hi monocytes in the liver, but with different localization in the affected tissue. The protective effect of monocytes from the listeria model already suggests heterogeneity of Ly6C hi monocytes between the two infections.

Monocytes from Both Infection Models Show Significant Differences in Gene Expression
To gain a deeper understanding of the phenotype of the Ly6C hi monocyte subset during hepatic amebiasis and listeriosis, 3 and 5 days p.i., CD11b + Ly6G -Ly6C hi as well as Ly6C lo monocytes were sorted from infected livers by flow cytometry. The RNA from both populations was extracted and subjected to RNA sequencing ( Figure 2A). Genes showing a significant difference in expression (adjusted p value < 0.05) in monocytes obtained from the two infection models at d3 p.i. were included in the analysis. 5486 genes were differentially expressed in Ly6C hi monocytes from E. histolytica-and L. Genes showing a significant difference in expression (adjusted p value < 0.05) in monocytes obtained from the two infection models at d3 p.i. were included in the analysis. 5486 genes were differentially expressed in Ly6C hi monocytes from E. histolyticaand L. monocytogenes-infected mice, 54 genes were differentially expressed in Ly6C lo cells from both models, and 194 genes were differentially expressed in both Ly6C hi and Ly6C lo monocytes in both infection models ( Figure 2B). PANTHER GO analysis of genes differentially expressed in Ly6C hi monocytes revealed that a small percentage of genes was included in the GO terms "immune systems process" (GO:0002376) and "response to stimulus" (GO:0050896) ( Figure 2B). By contrast, we also observed the differential expression of genes associated with GO terms related to immune responses in anti-inflammatory Ly6C lo monocytes (see Figure S2A-D for a detailed list of highly regulated genes and GO terms).
Principal component analysis of Ly6C hi and Ly6C lo monocytes revealed clustering into different groups ( Figure 2C). The day post infection did not affect the grouping, however there was a clear difference between Ly6C hi monocytes from L. monocytogenes-infected mice (cluster A) and Ly6C hi monocytes from E. histolytica-infected mice (cluster B): the latter clustered together with Ly6C hi monocytes from naïve mice. Additional differences between Ly6C hi and Ly6C lo from the E. histolytica and the listeria model are also depicted by a volcano-plot and a heat map (see Figure S2E,F). For example, differences include genes involved in proinflammatory IFN-γ related signaling (i.e., Iigp1, Gbp2, and Gbp8) in Ly6C hi monocytes from L. monocytogenes-infected mice, but the upregulation of genes involved in anti-inflammatory, phagocytic, or metabolic processes (Cx3cr1, Mfge8, Hpgd) in Ly6C hi monocytes from E. histolytica-infected mice. Overall, proinflammatory Ly6C hi monocyte in both infection models differed significantly at the transcriptional level. Ly6C hi monocytes from L. monocytogenes infected mice presented an upregulated expression of a large number of genes, including genes with a potential function in their antibacterial response. In contrast, changes in the gene expression of Ly6C hi monocytes from E. histolytica infected mice were less pronounced and a large part of their expression profile was shared with Ly6C hi monocytes from naïve mice. By focusing on transcription factors with putative relevance to the polarization of monocytes towards classically activated M1 or alternatively activated M2 macrophages, we found that during infection with L. monocytogenes, the activation and development of proinflammatory monocytes is characterized by factors, such as Irf1, Irf2, Ifi204, Batf2, Mndal, and Irf7 [24][25][26][27][28] (Figure 3A). By contrast, during E. histolytica infection, Ly6C hi monocytes are characterized by the upregulation of transcription factors Mafb, Hes1, Fos, and Tsc22d3 ( Figure 3A), which contribute to an anti-inflammatory and regenerative phenotype [29][30][31][32][33]. Time-course analysis of the expression data shows that, in addition to other genes, a selection of the above factors exhibits significantly different expression patterns between the two infection models, starting as early as d3 p.i. and remaining different until d5 p.i. (Figure 3B).
Fate mapping and transfer approaches have revealed the marked plasticity of proinflammatory monocytes and identified distinct routes of monocyte polarization. Such studies suggest the existence of novel monocyte subsets, such as "Ly6C hi to Ly6C lo "-converting monocytes, nMO, dcMO, and Cxcl10 + and Saa3 + monocytes [12][13][14]. According to this classification, the Ly6C hi monocytes triggered by E. histolytica infection would appear to belong to the Ly6C hi to Ly6C lo -converting monocyte subset. By contrast, with the exception of Csf1 and Ly6c2 (Ly6C), the respective genes were downregulated in monocytes from the L. monocytogenes infection model ( Figure 3C) [35]. Relevant genes related to nMO development were upregulated in monocytes from L. monocytogenes-infected mice ( Figure 3C). When we considered the genes that define dcMO, we found an intermediate picture. MHC-II related genes were upregulated in monocytes from L. monocytogenes-infected mice. However, some other hallmark genes of dcMO (i.e., Flt3, Pid1, and Hpgd) were strongly downregulated. Moreover, with the exception of il1b, signature genes of Cxcl10 + and Saa3 + monocytes were also upregulated in Ly6C hi monocytes from L. monocytogenes-infected mice, indicating a broad repertoire of putative new Ly6C hi monocyte subsets during this type of infection ( Figure 3C). However, several other relevant genes involved in proinflammatory or antiinflammatory immune processes are additionally upregulated in monocytes following E. histolytica infection. Among these are Cd14, Trem2, a negative immune regulator and marker for M2 polarization [36], as well as Arg1 and Arg2, further indicating the transition from pro-to anti-inflammatory monocytes ( Figure 3D). (D) Heat map of selected genes upregulated during ALA. All heatmaps were designed using the online tool "heatmapper" [34].
Fate mapping and transfer approaches have revealed the marked plasticity of proinflammatory monocytes and identified distinct routes of monocyte polarization. Such studies suggest the existence of novel monocyte subsets, such as "Ly6C hi to Ly6C lo "-converting monocytes, nMO, dcMO, and Cxcl10 + and Saa3 + monocytes [12][13][14]. According to this classification, the Ly6C hi monocytes triggered by E. histolytica infection would appear to belong to the Ly6C hi to Ly6C lo -converting monocyte subset. By contrast, with the exception of Csf1 and Ly6c2 (Ly6C), the respective genes were downregulated in monocytes from the L. monocytogenes infection model ( Figure 3C) [35]. Relevant genes related to nMO development were upregulated in monocytes from L. monocytogenes-infected mice ( Figure 3C). (D) Heat map of selected genes upregulated during ALA. All heatmaps were designed using the online tool "heatmapper" [34].
In summary, we were able to assign monocytes from both infection models to recently suggested subgroups with a more proinflammatory and activated phenotype in the L. monocytogenes model (nMO; dcMO; Cxcl10 + and Saa3 + monocytes) and a scarcely activated phenotype in the parasite model characterizing Ly6C hi to Ly6C lo converting monocytes.

Surface Marker Expression Implies Pathogen-Dependent Subsets of Proinflammatory Monocytes
Next, we analyzed the differential expression of genes encoding surface markers that may be useful for further subdivision of Ly6C hi monocytes. We found significant upregulation of genes encoding Ly6c2, Cd38, and Cd74 in Ly6C hi monocytes from L. monocytogenes-infected mice, whereas higher expression of Cd14 was characteristic for Ly6C hi monocytes from E. histolytica-infected mice ( Figure 4A,B). The expression of these genes at the protein level on Ly6C hi monocytes from both infection models was validated by flow cytometry. Although not significantly regulated at the transcriptional level, we included the analysis of the co-stimulatory receptor CD86 in the panel to further describe M1 polarization of inflammatory monocyte [37]. As in monocytes from naïve mice and in agreement with the transcriptomic data, we found that the percentage of CD14-expressing Ly6C hi monocytes was higher and remained higher from d1 p.i. on following E. histolytica infection than in Ly6C hi monocytes from L. monocytogenes-infected mice, the latter initially decreased but increased from d5 of infection ( Figure 4C). This picture was mirrored by lower MFIs for CD14 on Ly6C hi monocytes derived from the L. monocytogenes model than on monocytes from the E. histolytica model ( Figure 4D). When compared with that in uninfected animals, expression of CD38 increased significantly in both models shortly after infection. However, expression was significantly stronger on monocytes from the L. monocytogenes infection model ( Figure 4D). Initially, the percentage of CD74-expressing monocytes remained the same as that in naïve mice following E. histolytica infection, but decreased on d5 p.i. (Figure 4C). During L. monocytogenes infection, expression and MFI of CD74 increased on d1 p.i., but then decreased to the level observed in uninfected animals as infection progressed ( Figure 4C,D). The expression and MFI level of CD86 also decreased significantly over time during infection with L. monocytogenes, and to a lesser extent this was also true for monocytes in the E. histolytica model ( Figure 4C) (expression on monocytes from spleen, blood, bone marrow see Figure S3A-C).
Overall, the results of the transcriptome analysis of surface marker expression, with the exception of CD74, are reflected at the protein level in vivo. Furthermore, they suggest that CD14 in combination with CD38 may be additional putative markers for a Ly6C hi monocyte subset that is very different from the conventional proinflammatory Ly6C hi monocyte subset.
To further differentiate Ly6C hi monocytes, we selected molecules that were shown by transcriptome analysis to be highly expressed by Ly6C hi monocytes after E. histolytica infection (Mafb) or L. monocytogenes infection (Irf2) ( Figure 3B). However, MAFB1 protein was excluded from further analysis since less than 1% of Ly6C hi monocytes expressed the protein (data not shown). As seen for mRNA, protein expression of IRF2 was significantly stronger in Ly6C hi monocytes after L. monocytogenes infection than in monocytes from naïve or E. histolytica-infected mice ( Figure 5A). On day 3 after infection, when the most severe symptoms in both models appear, the combination of antibodies against IRF2 and CD14 resulted in the detection of a significantly increased monocyte population after E. histolytica infection. When IRF2 detection was combined with the detection of CD38 and CD86, Ly6C hi monocytes after infection with L. monocytogenes were significantly different from those of naive or E. histolytica-infected animals, whereas the combination of IRF2 detection with CD74 did not reveal significantly different monocyte subpopulations between the two infection models ( Figure 5B).  Overall, the results of the transcriptome analysis of surface marker expression, with the exception of CD74, are reflected at the protein level in vivo. Furthermore, they suggest that CD14 in combination with CD38 may be additional putative markers for a Ly6C hi  CD14 resulted in the detection of a significantly increased monocyte population after E. histolytica infection. When IRF2 detection was combined with the detection of CD38 and CD86, Ly6C hi monocytes after infection with L. monocytogenes were significantly different from those of naive or E. histolytica-infected animals, whereas the combination of IRF2 detection with CD74 did not reveal significantly different monocyte subpopulations between the two infection models ( Figure 5B). (B) Gating strategy (exemplary for CD14 + gated cells) based on the FMO control to determine CD14 + , CD38 + , CD74 + , and CD86 + monocytes within the IRF2 + Ly6C hi monocyte population from naïve mice and from both infection models on day 3 post infection. (C) UMAP plots of Ly6C hi monocytes (including CD14 + , CD86 + , CD74 + , CD38 + ) from naïve mice and from E. histolyticaand L. monocytogenes-infected mice (down sampled to 90,000 cells per sample). Six samples per source material were included in the analysis. (D) Cluster heatmap table showing surface marker and IRF2 expression by Ly6C hi monocytes, cluster events, and integration of cluster events in the UMAP plot (* p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001; Mann-Whitney U test).
After analysis of the distribution of Ly6C hi monocyte subpopulations by UMAP, initially considering only surface markers, clear demarcation of populations by L. monocytogenes and E. histolytica infection and by naïve animals was seen, with overlapping populations of the latter ( Figure 5C). When Irf2 was included, 6 distinct clusters were identified ( Figure 5D). A distinct IRF2-positive monocyte population and a greater heterogeneity in the area of monocytes from the listeria model and a clear delineation of monocyte populations from naive and E. histolytica-infected animals could be visualized in the UMAP analysis after clustering with FlowSOM and Cluster Explorer analysis ( Figure 5D).
In summary, we found that the combination of antibodies against IRF2 with antibodies against CD14, CD38, or CD86 helped to distinguish proinflammatory Ly6C hi monocytes from both infection models, supporting the results of the transcriptome study ( Figure 3C) in that these monocytes are already in a transitional stage to anti-inflammatory monocytes.

CD38 + Ly6C hi Monocytes Produce ROS and Contribute to Monocyte-Dependent Immunopathology during Hepatic Amebiasis
In addition to the production of proinflammatory cytokines and chemical mediators, activated Ly6C hi monocytes also express ROS [38]. Based on the higher mRNA expression of genes involved in ROS production and NADPH oxidase (i.e., Sod1, Sod2, Ncf1, Ncf 4 as well as Nox2 and Noxred 1) by Ly6C hi monocytes from L. monocytogenes-infected compared with E. histolytica-infected mice ( Figure 6A), we analyzed ROS production by Ly6C hi monocytes in both models ( Figure 6B). Consistent with the transcriptomic results, and consistent with a low MFI, the percentage of ROS + Ly6C hi monocytes decreased during E. histolytica infection ( Figure 6C) but increased during L. monocytogenes infection ( Figure 6D).
To further characterize ROS + Ly6C hi monocytes, we examined the expression of surface marker CD38, which exhibits various functions during cell activation [39]. Interestingly, although the number of ROS-producing Ly6C hi monocytes decreased during infection with E. histolytica, the percentage of ROS + CD38 + out of Ly6C hi monocytes increased rapidly, and remained elevated, during infection ( Figure 6E).
As expected, the proportion of these cells also increased during infection with L. monocytogenes, but with a delay compared with parasitic infection, and the final proportion was higher ( Figure 6E). Next, we used knockout mice lacking CD38 [40,41] to investigate whether CD38 + Ly6C hi monocytes contributes to abscess formation during E. histolytica liver infection. Cd38 -/mice had significantly smaller abscesses on d3 p.i. ( Figure 6F) and a significantly lower level of proinflammatory monocytes (similar to naïve mice) ( Figure 6G).
In summary, ROS-production as well as the expression of CD38 characterizes the true proinflammatory phenotype of Ly6C hi monocytes in both infection models. To further characterize ROS + Ly6C hi monocytes, we examined the expression of surface marker CD38, which exhibits various functions during cell activation [39]. Interestingly, although the number of ROS-producing Ly6C hi monocytes decreased during infection with E. histolytica, the percentage of ROS + CD38 + out of Ly6C hi monocytes increased rapidly, and remained elevated, during infection ( Figure 6E).
As expected, the proportion of these cells also increased during infection with L. monocytogenes, but with a delay compared with parasitic infection, and the final proportion was higher ( Figure 6E). Next, we used knockout mice lacking CD38 [40,41] to investigate whether CD38 + Ly6C hi monocytes contributes to abscess formation during E. histolytica

Discussion
Monocytes are critical for the defense against microbial infections, but also for promoting resolution of inflammation. However, an improper balance of these tasks can lead to the collateral damage of host tissues and delay of tissue regeneration [42].
The rationale for the present study arose from the striking differences in the function of classical proinflammatory Ly6C hi monocytes revealed by Ccr2 -/-. Whereas the lack in the egress from the bone marrow and hence the recruitment of Ly6C hi monocyte prevented liver destruction after E. histolytica infection [16], their absence in L. monocytogenes infection exacerbated disease progression [23], pointing to functional differences in this monocyte subset which was originally regarded as homogeneous. Recent studies based on single-cell sequencing actually suggest an even greater diversity. Under conditions of homeostasis three more proinflammatory monocyte subsets: nMO, dcMO [12,13,43,44], and Cxcl10 + and Saa3 + monocytes were identified [14]. Interestingly, all analyzed genes involved in development of nMO were strongly upregulated in proinflammatory monocytes during L. monocytogenes infection, but they were unaffected in monocytes from the amebic model. Enhanced development of Ly6C hi monocytes into nMO has been demonstrated previously, but only under LPS stimulation [45] and the present study is the first to demonstrate its presence in vivo by bacterial infection. Likewise, genes related to MHC-II-mediated antigen presentation and dcMO development were upregulated in listeria infection. However, some factors thought to be important for dcMO development (i.e., Flt3, Pou2f2, Pid1, and Hpgd) were strongly downregulated while their expression by monocytes from the amebiasis model was comparably higher. In addition, genes characteristic for the subset of Cxcl10 + and Saa3 + monocytes that arise under sterile inflammatory conditions (e.g., autoimmune encephalitis) [14] were only upregulated in the listeria model. Taken together, the data suggest that proinflammatory monocytes from L. monocytogenes-infected animals display a distinct proinflammatory phenotype, characterized by the upregulation of genes associated with nMO, dcMO and Cxcl10 + and Saa3 + cells.
Additional relevant transcriptional differences between Ly6C hi monocytes from both models became apparent by examining the expression of selected transcription factors and genes involved in polarization of monocytes. Ly6C hi monocytes from the parasite model exhibited a more anti-inflammatory phenotype and the expression of genes involved in conversion of Ly6C hi monocytes to Ly6C lo monocytes. Altered genes include Nr4a1, a major transcription factor responsible for transition of Ly6C hi to Ly6C lo and survival of Ly6C lo cells [30,46], while Csf1 and Ly6c2, promoting survival and activation of Ly6C hi monocytes, were downregulated [35,47]. Their further polarization towards anti-inflammatory M2 macrophages [9] is supported by the upregulation of MafB, Nr4a1, or Fos [29,30,33]. As already evident from the cluster analysis, Ly6C hi monocytes from the parasite model were overall quite similar to those from naïve animals. However, some genes were differentially regulated, e.g., Arg1/Arg 2, further suggesting an ongoing polarization into an anti-inflammatory M2 phenotype. In contrast, monocytes from the L. monocytogenes infection model were characterized by a classical proinflammatory, interferon-driven transcription factor-like profile (i.e., Mndal, Ifi204 and Irf2) [25,28], pointing towards an M1 phenotype [9].
A suitable antibody panel to distinguish bona fide inflammatory Ly6C hi monocytes from non-inflammatory Ly6C hi monocytes and to study their dynamics of Ly6C hi monocytes in both infection models, was developed by selecting several surface markers that had emerged from transcriptome analysis. These included activation markers such as CD38 [39][40][41], CD74, a receptor for proinflammatory macrophage migration inhibitory factor involved in cell proliferation and antigen presentation [48,49] as well as CD14, coreceptor for several Toll-like receptors involved in proinflammatory processes [50]. While CD38 and CD74 were more highly expressed in Ly6C hi monocytes from the L. monocytogenes infection model, CD14 was the only surface marker with higher expression in Ly6C hi monocytes from the parasite model. Although not differentially expressed on the mRNA level, we included CD86 as an additional proinflammatory M1 marker [41].
The expression of CD14 and CD74 on Ly6C hi monocytes from the parasite model was very similar to those from naïve mice, whereas CD14 in the L. monocytogenes model initially decreased and only increased toward the end of the disease course. CD38 was more highly expressed on monocytes from the parasitic model during the early phase of infection, thus describing inflammatory Ly6C hi before transitioning to Ly6C lo cells. As expected, CD38, CD74, and to a lesser extent CD86 remain highly expressed in Ly6C hi monocytes from the L. monocytogenes model over time, suggesting that these markers are useful for further distinguishing proinflammatory Ly6C hi monocytes. Next, we included the transcription factor IRF2 within the panel. IRF2 in combination with CD14 best distinguishes the Ly6C hi population in the ameba and listeria model from the population in naïve mice at least on day 3 post infection, the peak of liver pathology in these models ( Figure 5B). Subsequent UMAP analysis based on the designated surface markers confirmed clear delineation of Ly6C hi monocyte populations between naïve, E. histolyticaand L. monocytogenes-infected mice, as well as the highest diversity of proinflammatory monocyte subpopulations from the L. monocytogenes model. Finally, we used ROS production as a hallmark of proinflammatory monocyte activation [38]. In contrast to the E. histolytica model, where it remained stable, we observed a continuous increase in ROS-producing Ly6C hi monocytes expressing CD38 in the listeria model. Up-regulation of CD38 on monocytes during infection with listeria has been described previously [39]. Interestingly, genetic deletion of CD38 resulted in increased accumulation of inflammatory monocytes in the liver but not in the spleen and was associated with higher susceptibility to listeria infection, as observed during genetic deletion of Ccr2 [23,39]. Assuming that CD38 + ROS + monocytes are responsible for immunopathological mechanisms during the early phase of hepatic amebiasis, genetic deletion of CD38 should lead to smaller abscesses. Indeed, we were able to detect this phenotype, and it was associated with a marked reduction in the proportion of recruited Ly6C hi monocytes.
In summary, analysis of Ly6C hi monocyte populations from two different infection models shows that proinflammatory Ly6C hi monocytes differ depending on the infectious agent. Based on the present results, we propose that the addition of IRF2, CD14, and CD38 or CD86 to the classical markers (CD11b, Ly6C, Ly6G) can help distinguish true proinflammatory Ly6C hi monocytes from non-inflammatory Ly6C hi monocytes.