Differential Cytokine-Induced Responses of Polarized Microglia

The role of select pro- and anti-inflammatory mediators in driving microglial cell polarization into classically (M1), or alternatively, (M2) activated states, as well as the subsequent differential responses of these induced phenotypes, was examined. Expression of PD-L1, MHC-II, MHC-I, arginase 1 (Arg-1), and inducible nitric oxide synthase (iNOS) was assessed using multi-color flow cytometry. We observed that both pro- and anti-inflammatory mediators induced PD-L1 expression on non-polarized microglia. Moreover, IFN-γ stimulated significant MHC class I and II expression on these cells. Interestingly, we observed that only IL-4 treatment induced Arg-1 expression, indicating M2 polarization. These M2 cells were refractory to subsequent depolarization and maintained their alternatively activated state. Furthermore, PD-L1 expression was significantly induced on these M2-polarized microglia after treatment with pro-inflammatory mediators, but not anti-inflammatory cytokines. In addition, we observed that only LPS induced iNOS expression in microglial cells, indicating M1 polarization. Furthermore, IFN-γ significantly increased the percentage of M1-polarized microglia expressing iNOS. Surprisingly, when these M1-polarized microglia were treated with either IL-6 or other anti-inflammatory cytokines, they returned to their non-polarized state, as demonstrated by significantly reduced expression of iNOS. Taken together, these results demonstrate differential responses of microglial cells to mediators present in dissimilar microenvironments.


Introduction
It is well-established that upon infection or insult within the brain there is a secretion of a wide array of cytokines and chemokines by both immune and non-immune cells [1]. These low molecular weight polypeptides invigorate the cells and draw them towards sites of neuroinflammation. They can act either at the site they emanated (autocrine effect) or on neighboring cells (paracrine fashion), or even on remote cells (endocrine effect/systemic effect) [2]. It is also common that individual cytokines have varying effects on different cell types (pleiotropy). Cytokines are redundant in their function and are often produced in a cascade (i.e., one triggers its target cells to make additional cytokines, which can work synergistically or antagonistically). They modulate cell-to-cell interactions that are directly involved in the advancement of disease pathogenesis [3,4]. Various cytokines are secreted within the brain, which can be either pro-inflammatory or anti-inflammatory [2]. Pro-inflammatory cytokines are secreted by activated cells and are involved in upregulating inflammatory responses. There is evidence suggesting that certain pro-inflammatory cytokines are involved in neuroinflammation-induced sensitization of the brain, such as interleukin (IL)-1β, IL-6, and tumor necrosis factor (TNF)-α [5,6]. IL-1β is primarily produced by macrophages as well as non-immune cells, such as fibroblasts and endothelial cells, during cell injury or inflammation [2,7]. IL-6 is known to be involved in microglial activation and plays a central role in the development of neuropathic pain following a

Ethical Statement
This study was carried out according to guidelines for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the University of Minnesota's Institutional Animal Care and Use Committee (Protocol Number: 2001-37741A). All animals were routinely cared for according to Research Animal Resources (RAR) guidelines. Animals were euthanized after isoflurane inhalation, whenever required and efforts were made to ameliorate suffering.

Experimental Animals
C57BL/6 (7-10 d old) mice were used for the isolation of microglial cells. Pathogenfree C57BL/6 (JAX stock#000664) mice were purchased from The Jackson Laboratory (Bar Harbor, Maine). Animals were lodged in individually ventilated cages and were provided with food and water ad libitum at RAR, University of Minnesota.

Primary Murine Microglial Cell Cultures
Murine microglial cells from 7-10 d old mice were isolated using the Adult Brain Dissociation Kit followed by CD11b microbeads (Miltenyi Biotec, Gaithersburg, MD, USA). Briefly, brains were dissected, weighed, and enzymatically digested using the dissociation kit for 30 min at 37 • C (>100 mg tissue). Further tissue processing was carried out at 4 • C. The cell suspension was passed through a 70 µm cell strainer and the supernatant was centrifuged at 300× g for 10 min. Cells were resuspended in cold D-PBS and tissue debris was removed by using Debris Removal Solution and overlayed with an appropriate amount of cold D-PBS. Cells were centrifuged at 3000× g for 10 min and the top two phases (containing debris) were discarded. Cells were washed and further processed for removing red blood cells (RBCs). For this, the cell pellet was resuspended in cold 1X RBC Removal Solution and refrigerated for 10 min. Cells were then washed and resuspended in an appropriate volume of cold magnetic buffer (PBS + 0.5% BSA). To isolate the microglia, cells were incubated with 10 µL of CD11b microbeads per 10 7 cells/90 µL for 15 min in the refrigerator. CD11b + cells were separated in a magnetic field using MS columns. The amounts of antibodies and the buffers were calculated based on the number of cells obtained after RBCs removal, using the manufacturer's guidelines. Purified microglial cells stained >95% positive for the microglial phenotype (CD45 int CD11b + ).

In-Vitro Stimulation of M1 and M2-Activated Microglia
Microglial cells were treated with LPS or IL-4 to polarize them to M1-or M2-activated states for 48 h. The polarized cells were then treated with select pro-and anti-inflammatory cytokines for 24 h, as discussed previously. Cells were then trypsinized and collected for flow cytometry.

Statistical Analyses
GraphPad Prism software was used to determine statistical significance (version 9.2; Graphpad Software, La Jolla, CA, USA). For comparing groups, Dunnett's multiple comparisons test and a two-tailed unpaired T-test were employed. A p-value < 0.05 was regarded as significant.

Expression of MHC Molecules on Non-Polarized Microglial Cells
We further went on to investigate the role of different cytokines in modulating m croglial activation by analyzing MHC molecules. Both MHC class I and class II are upr ulated on antigen-presenting cells to present the peptides to the immune cells for th effective clearance. For this experiment, microglial cells were treated with different cy kines as described previously, and 24 h later, cells were stained for both MHC class I a class II molecules. We observed that only IFN-γ induced significant MHC-II express on microglia after 24 h of treatment ( Figure 2). Since we did not observe any MHC clas expression at 24 h or 48 h post-IFN-γ treatment (Supplementary Figure S1), we analyz its expression at 72 h post-treatment. We observed that IFN-γ, TNF-α, and IL-1β all duced significant MHC-I expression on microglial cells after 72 h of treatment compar to the untreated control ( Figure 3).

Expression of MHC Molecules on Non-Polarized Microglial Cells
We further went on to investigate the role of different cytokines in modulating microglial activation by analyzing MHC molecules. Both MHC class I and class II are upregulated on antigen-presenting cells to present the peptides to the immune cells for their effective clearance. For this experiment, microglial cells were treated with different cytokines as described previously, and 24 h later, cells were stained for both MHC class I and class II molecules. We observed that only IFN-γ induced significant MHC-II expression on microglia after 24 h of treatment ( Figure 2). Since we did not observe any MHC class I expression at 24 h or 48 h post-IFN-γ treatment (Supplementary Figure S1), we analyzed its expression at 72 h post-treatment. We observed that IFN-γ, TNF-α, and IL-1β all induced significant MHC-I expression on microglial cells after 72 h of treatment compared to the untreated control ( Figure 3).     Experiments were carried out three times with duplicate wells. * p < 0.05; ** p < 0.01; *** p < 0.001.

Polarization of Microglial Cells to Alternatively Activated State (M2)
We have previously demonstrated in vitro microglial cell polarization following IL-4 treatment using real-time RT-PCR for arginase-1 (Arg-1) mRNA [31]. In this study, we treated microglia with several cytokines and analyzed Arg-1 expression using both flow cytometry and RT-PCR at 24 h. We observed that, of the eight different treatments, only IL-4 induced Arg-1 expression indicating M2 (alternatively activated) polarization ( Figure 4A,B; Supplementary Figure S2). These M2-polarized cells (treated with IL-4 for 48 h) were then treated with both pro-and anti-inflammatory cytokines for 24 h, after which they were analyzed for the expression of Arg-1. Interestingly, these M2-polarized microglial cells maintained their Arg-1 expression, and thus their alternatively activated state, even after being exposed to various pro-and anti-inflammatory mediators ( Figure 4C).

Polarization of Microglial Cells to Alternatively Activated State (M2)
We have previously demonstrated in vitro microglial cell polarization following IL-4 treatment using real-time RT-PCR for arginase-1 (Arg-1) mRNA [31]. In this study, we treated microglia with several cytokines and analyzed Arg-1 expression using both flow cytometry and RT-PCR at 24 h. We observed that, of the eight different treatments, only IL-4 induced Arg-1 expression indicating M2 (alternatively activated) polarization ( Figure  4A,B; Supplementary Figure S2). These M2-polarized cells (treated with IL-4 for 48 h) were then treated with both pro-and anti-inflammatory cytokines for 24 h, after which they were analyzed for the expression of Arg-1. Interestingly, these M2-polarized microglial cells maintained their Arg-1 expression, and thus their alternatively activated state, even after being exposed to various pro-and anti-inflammatory mediators ( Figure 4C).

Cytokine-Induced PD-L1 Expression on M2-Polarized Microglia
We went on to analyze PD-L1 expression on the M2-polarized microglia. Microglia were polarized with IL-4 for 48 h and then treated with various cytokines for an additional 24 h. Cells were first gated for Arg-1 expression, as shown in Figure 4A, and then the Arg-1 positive microglial cells were analyzed for PD-L1 expression by flow cytometry. We observed that M2-polarized microglial cells significantly induced the expression of PD-L1 after treatment with pro-inflammatory mediators, with the exception of IL-6. However,

Cytokine-Induced PD-L1 Expression on M2-Polarized Microglia
We went on to analyze PD-L1 expression on the M2-polarized microglia. Microglia were polarized with IL-4 for 48 h and then treated with various cytokines for an additional 24 h. Cells were first gated for Arg-1 expression, as shown in Figure 4A, and then the Arg-1 positive microglial cells were analyzed for PD-L1 expression by flow cytometry. We observed that M2-polarized microglial cells significantly induced the expression of PD-L1 after treatment with pro-inflammatory mediators, with the exception of IL-6. However, we did not observe any PD-L1 induction on these cells following treatment with anti-inflammatory cytokines ( Figure 5A,B). Interestingly, when we compared the expression of PD-L1 on non-polarized (M0) as well as M2-polarized microglial cells, we observed significantly less expression of PD-L1 on M2-polarized cells when compared to M0-microglia after treatment with both pro-and anti-inflammatory cytokines with the exception of TNF-α, Il-6, and TGF-β ( Figure 5C). we did not observe any PD-L1 induction on these cells following treatment with anti-inflammatory cytokines ( Figure 5A,B). Interestingly, when we compared the expression of PD-L1 on non-polarized (M0) as well as M2-polarized microglial cells, we observed significantly less expression of PD-L1 on M2-polarized cells when compared to M0-microglia after treatment with both pro-and anti-inflammatory cytokines with the exception of TNF-α, Il-6, and TGF-β ( Figure 5C). UT-untreated control. Experiments were carried out three times with duplicate wells. * p < 0.05; ** p < 0.01; *** p < 0.001.

Polarization of Microglial Cells to Classically Activated State (M1)
We also wanted to investigate cytokine-induced polarization of microglia to the M1 state. For this, we analyzed iNOS expression after treatment with both pro-and anti-inflammatory cytokines using both flow cytometry and RT-PCR. Interestingly, we observed that only LPS induced iNOS expression after 24 h of treatment, indicating their M1 (classically activated) state ( Figure 6A,B; Supplementary Figure S2). Further, when these cells were polarized to M1 using LPS and then treated with either pro-or anti-inflammatory cytokines, we observed that IFN-γ significantly potentiated the % of microglial cells expressing iNOS ( Figure 6C,D). Interestingly, when these M1-polarized microglia were treated with either IL-6 or the anti-inflammatory cytokines (IL-4, IL-10, and TGF-β), they returned to their non-polarized state, as demonstrated by their significantly reduced expression of iNOS ( Figure 6C,D).

Polarization of Microglial Cells to Classically Activated State (M1)
We also wanted to investigate cytokine-induced polarization of microglia to the M1 state. For this, we analyzed iNOS expression after treatment with both pro-and anti-inflammatory cytokines using both flow cytometry and RT-PCR. Interestingly, we observed that only LPS induced iNOS expression after 24 h of treatment, indicating their M1 (classically activated) state ( Figure 6A,B; Supplementary Figure S2). Further, when these cells were polarized to M1 using LPS and then treated with either pro-or anti-inflammatory cytokines, we observed that IFN-γ significantly potentiated the % of microglial cells expressing iNOS ( Figure 6C,D). Interestingly, when these M1-polarized microglia were treated with either IL-6 or the anti-inflammatory cytokines (IL-4, IL-10, and TGF-β), they returned to their non-polarized state, as demonstrated by their significantly reduced expression of iNOS ( Figure 6C,D).

Discussion
Resident brain cells and their microenvironment actively modulate immune responses [16,32]. Microglia display many characteristics similar to peripheral macrophages and they are considered sentinels that can orchestrate potent inflammation [33]. They respond to infection and injury by changing morphology and migrating. Their activation depends on disease context and factors such as aging, environment, or cell-to-cell interaction [33]. In this study, we aimed to elucidate microglial activation and polarization in response to select cytokines. Cytokines are small signaling polypeptides that play a piv- UT-untreated control. Experiments were carried out three times with duplicate wells. * p < 0.05; ** p < 0.01; *** p < 0.001.

Discussion
Resident brain cells and their microenvironment actively modulate immune responses [16,32]. Microglia display many characteristics similar to peripheral macrophages and they are considered sentinels that can orchestrate potent inflammation [33]. They respond to infection and injury by changing morphology and migrating. Their activation depends on disease context and factors such as aging, environment, or cell-to-cell interaction [33]. In this study, we aimed to elucidate microglial activation and polarization in response to select cytokines. Cytokines are small signaling polypeptides that play a pivotal role in disease [3]. They display multifaceted interactions involved in disease pathogenesis. Recent advances in cytokine biology have led to the development of new anti-viral therapies based on mimicking cytokine networks [3].
In this study, we cultured and treated primary murine microglial cells with several prototypical pro-and anti-inflammatory cytokines. First, we analyzed the expression of PD-L1 on resting non-polarized microglial cells using flow cytometry. We observed that both pro-and anti-inflammatory cytokines enhanced PD-L1 expression, however, the levels were different. These results are consistent with our previous report where IFN-γ, TNF-α, and IL-1β induced PD-L1 expression on primary cultured microglia (from d1 pups) as shown by real-time RT-PCR [28]. Others have also positively correlated IFN-γ level with PD-L1 expression in murine glioma cells, bone marrow-derived macrophages, and primary cultured microglia [34]. We did not observe a significant percentage of microglia expressing PD-L1 after treatment with IL-6 or TGF-β. Reports suggest that IL-6 signaling is important for PD-L1 expression after viral infection, but this may not be the case in vitro, as shown in our study [35]. Similarly, TGF-β induces PD-L1 expression in murine fibroblasts, but this has not been shown in microglial cells [36]. It has been previously shown that IL-4 upregulates mRNA for PD-L1 in human monocytes [37]. Investigations into the role of IL-10 have been mainly restricted to PD-1 expression on tumor-associated dendritic cells (DCs) and bone marrow-derived DC [38]. In this study, we show that IL-4, as well as IL-10, induces the expression of PD-L1 on microglia (Figure 1).
We have previously demonstrated the expression of MHC-II and MHC-I on primary microglial cells isolated from d1 pups following IFN-γ treatment [28]. In this study, we expanded our investigation to assess the effects of additional cytokines (as well as IFN-γ) using primary microglia isolated from d7-10 old pups (using the Miltenyi adult dissociation kit and CD11b selection). We examined IFN-γ, TNF-α, IL-1β, IL-6, IL-4, IL-10, and TGF-β; and established that IFN-γ most robustly induced the expression of both MHC-II and MHC-I (Figures 2 and 3). However, MHC-II was expressed within 24 h of treatment, while MHC-I expression took 72 h. Besides IFN-γ, TNF-α and IL-1β treatments also upregulated MHC-I on microglial cells after 72 h, but the expression observed was significantly less than that induced by IFN-γ. IFN-γ-induced expression of MHC-II has also been previously demonstrated on mouse macrophage cell lines, bone marrow-derived macrophages, and astrocytes [39][40][41]. It has also been previously shown that IFN-γ upregulates MHC class I on primary murine neurons after 72 h of treatment [42].
Investigators will encounter a body of published work regarding the terminology of macrophage/microglial polarization, mainly into the M1 and M2 state. This phenotypic characterization was adopted to simplify the interpretation of data [43]. To investigate the role of cytokines in polarizing microglial cells to the M1 or M2 state, we analyzed cells for expression of iNOS and Arg-1, respectively. Similar to our previous report, IL-4 was found to drive microglial polarization to M2 [31]. However, in this study, we have more comprehensively analyzed a panel of pro-and anti-inflammatory cytokines. Besides IL-4, no other cytokine was found to drive polarization to the M2 state ( Figure 4). Interestingly, when these M2-polarized microglial cells were further exposed to pro-and anti-inflammatory cytokines, they retained their M2 state, as indicated by Arg-1 expression. Hence, these M2-polarized cells were found to be quite refractory to depolarization. These M2-polarized microglial cells do significantly express PD-L1 after treatment with proinflammatory stimulants (LPS, IFN-γ, TNF-α, and Il-1β), but not after being exposed to anti-inflammatory cytokines. On comparing PD-L1 expression between M0-and M2polarized microglial cells, we observed a significant decrease in the percentage of M2polarized cells expressing PD-L1 after treatment with both pro-(LPS, IFN-γ and IL-1β) and anti-(IL-4 and IL-10) inflammatory stimulants as compared to M0-microglial cells ( Figure 5). It has been shown in the case of macrophages that PD-L1 is expressed on both M1 and M2 cells, and there is an increased expression when the macrophages are polarized from M2 to M1 [44].
We went on to analyze the polarization of microglia to an M1 state. Surprisingly, only LPS treatment was found to activate microglial cells to M1. This is contradictory to what we have observed previously where IFN-γ polarized microglia to the M1 state [31]. This could be attributed to the fact that microglial cells were isolated in a different way and from different aged mice in these two studies. When these M1-polarized microglial cells were exposed to pro-inflammatory cytokines, they maintained their M1 state. Moreover, IFN-γ significantly potentiated the number of microglial cells expressing iNOS, thereby the number of M1 polarized cells. Most interestingly, when these M1-polarized microglial cells were exposed to anti-inflammatory cytokines, they returned to the M0 state. Hence, the M1 state of microglial cells is very sensitive to depolarization and environmental stimuli (Figures 6 and 7).
We went on to analyze the polarization of microglia to an M1 state. Surprisingly, on LPS treatment was found to activate microglial cells to M1. This is contradictory to wh we have observed previously where IFN-γ polarized microglia to the M1 state [31]. T could be attributed to the fact that microglial cells were isolated in a different way a from different aged mice in these two studies. When these M1-polarized microglial ce were exposed to pro-inflammatory cytokines, they maintained their M1 state. Moreov IFN-γ significantly potentiated the number of microglial cells expressing iNOS, there the number of M1 polarized cells. Most interestingly, when these M1-polarized microg cells were exposed to anti-inflammatory cytokines, they returned to the M0 state. Hen the M1 state of microglial cells is very sensitive to depolarization and environmental sti uli (Figures 6 and 7).

Conclusions
In summary, we demonstrated that LPS treatment polarized microglia to an M1 sta as shown by increased iNOS expression (and no Arg-1 expression), while IL-4 polariz the cells to an M2 state as shown by increased Arg-1 (and no iNOS expression) (Figu 7A). When the M2-polarized microglial cells are exposed to various pro-and anti-infla matory cytokines, these cells stay M2-polarized (as indicated by positive Arg-1 expressi and no iNOS expression). They appear to be quite resistant to depolarization. Howev approximately 15% of M2-polarized microglial cells do exhibit an intermediate state activation after treatment with LPS, as shown by both positive Arg-1 and iNOS staini ( Figure 7B). On the other hand, microglial cells (pretreated with LPS) remain M1-pol

Conclusions
In summary, we demonstrated that LPS treatment polarized microglia to an M1 state, as shown by increased iNOS expression (and no Arg-1 expression), while IL-4 polarized the cells to an M2 state as shown by increased Arg-1 (and no iNOS expression) ( Figure 7A). When the M2-polarized microglial cells are exposed to various pro-and anti-inflammatory cytokines, these cells stay M2-polarized (as indicated by positive Arg-1 expression and no iNOS expression). They appear to be quite resistant to depolarization. However, approximately 15% of M2-polarized microglial cells do exhibit an intermediate state of activation after treatment with LPS, as shown by both positive Arg-1 and iNOS staining ( Figure 7B). On the other hand, microglial cells (pretreated with LPS) remain M1-polarized after additional treatment with IFN-γ, TNF-α, and IL-1β (as shown by positive iNOS expression). IFN-γ significantly increased the percentage of microglial cells expressing iNOS and thereby the % of M1-polarized microglial cells ( Figure 7C). Interestingly, when these M1-polarized microglial cells were treated with either IL-6 or the anti-inflammatory cytokines (IL-10 and TGF-β), they returned to their non-polarized state (negative iNOS and Arg-1 expression). Most importantly, when the M1-polarized microglial cells were treated with IL-4, a percentage of these cells de-polarized towards an M2 state (indicated by positive Arg-1 expression) ( Figure 7C). Although, classification into M1 and M2 polarization states has been helpful for conceptualizing microglial activities in vitro, it is increasingly acknowledged that the M1/M2 paradigm is oversimplified, artificial, and inadequate to describe the many states of microglial cell activation in vivo [43]. There is always a question about the relevance of classifying cell polarization as M1 and M2, as there are likely multiple phenotypes of microglia [45].
Limitation of the study: it is a comprehensive in vitro study investigating the modulation of microglial cells and their polarization in response to both pro-and anti-inflammatory cytokines. Although we examined the effects of the most common select, prototypical cytokines, we were limited in the number of mediators we could study. Responses could also vary because of treatment kinetics. This study utilized microglia isolated from 7-10 d-old pups. It would be interesting to compare these findings with those obtained using cells from older mice; since aged microglia may be phenotypically and functionally different from neonatal cells [46].