Development and characterisation of mgTHP-1, a novel in vitro model for neural macrophages with microglial characteristics

ABSTRACT Neuroinflammation is primarily characterised by activation of the brain’s resident macrophages – the microglia. However, other central nervous system (CNS) cells also contribute to this response, including the astrocytes and endothelial cells. In addition, there is infiltration into the CNS of peripherally derived immune cells. Together these cells mediate inflammation by the production of cytokines, chemokines, reactive oxygen species, and secondary messengers, and enacting of the appropriate response to those signals. However, deciphering the specific contributions of each cell type has been challenging. Studying CNS cell biology is often challenging, as the isolation of primary cells is not always feasible, and differentiation towards microglia-like cells is complex. Here, we demonstrate a novel method whereby THP-1 monocytic cells are differentiated into neural macrophage cells with microglia-like cell characteristics. The cells, designated mgTHP-1, show typical morphological and gene expression patterns of resident CNS macrophages and functionally respond to inflammatory stimuli by producing inflammatory cytokines. Furthermore, with the addition of Vicenin-2 (an anti-inflammatory flavonoid) such responses can be reversed. This novel cell model will allow further investigations, and hence insights, into the neuroinflammatory mechanisms associated with CNS diseases.


Introduction
Neuroinflammation is a hallmark of the pathophysiology of many CNS diseases and plays a pivotal role in disease progression; however, in many cases the cellular and molecular mechanisms underpinning this process remain elusive.Much focus has been on the role of the CNS resident macrophage cells, the microglia [1][2][3][4].It is well established that these cells react to stimuli in the brain to produce the desired immunogenic response -including the production of cytokines and chemokines to prevent damage to the local environment [4,5].Conversely, sustained and chronic overactivation of microglia can also produce a variety of proinflammatory cytokines and chemokines, reactive oxygen species (ROS), proteases, nitric oxide (NO) synthase, and prostaglandins, which increase brain damage [6].Therefore, targeting these cells in order to better understand their properties, and potentially control their responses, has been proposed as a therapeutic target; for example, depleting microglia by targeting microglial proliferation via the inhibition of CSF1R [7] has been shown to have beneficial effects in models of Alzheimer's disease (AD) [8][9][10].However, not all inflammation is 'harmful': although it is generally accepted that enhanced inflammation is detrimental to any tissue, anti-inflammatory signalling has been shown to have detrimental effects on plaque clearing in AD [11].Therefore, the diverse responses of inflammatory cells in different inflammatory settings have a direct effect on their functions, and hence the functions of the CNS, and as such studying such responses in vitro may elucidate roles that inflammation plays in neurodegenerative diseases.
Studying microglia is undertaken using either primary cells (human or murine) or differentiated cells from various sources and using various methods.Obtaining and using primary human cells is difficult; although using cells from either diseased tissue obtained during a biopsy, or cells from postmortem brains give valuable information, it is not certain whether these cells react completely as they do in vivo.On the other hand, given the differences between non-human and human microglia [12], using murine primary cells (or cell models) also comes with its own challenges.As such, much focus in the recent years has been in the development of in-vitro human models for these cells.Most of these models involve iPSCs (induced pluripotent stem cells) as the starting material, using different differentiation protocols which vary in complexity, yield, and validation/characterisation methodology, and result in cells that resemble microglia in many respects.iPSC models, however, are not accessible, in that their derivation, culture, and differentiation may often be laborious, expensive and timeconsuming.
In order to address this issue, we have developed an in-vitro cell model for peripheral blood-derived neural macrophages with microglia-like characteristics (henceforth referred to as mgTHP cells).This model is based on the use of THP-1 cells, a well-established human cell line, that has monocytic characteristics and is often used as a model for monocytes [13,14], and which can be differentiated into macrophage cells readily in culture [15].Herein, we provide evidence of mgTHP-1 cells' ability to differentiate into microglialike cells that express markers associated with microglia cells, such as TMEM119 and Iba1.Furthermore, these cells respond to inflammatory stimuli such as lipopolysaccharide (LPS) and interferon-gamma (IFN-γ) in a typical macrophage-like fashion, by producing inflammatory cytokines and by expressing either M1 or M2 characteristics.Previously, we have shown Vicenin −2 to have potent anti-inflammatory properties when used on PMA differentiated THP-1 cells.In this study, we show that mgTHP-1 cells were also responsive to the anti-inflammatory effects of the previously reported flavonoid Vicenin-2 [16][17][18].

Cell differentiation
For differentiation towards microglia-like cells, THP-1 cells were incubated with serum free supplemented RPMI, enriched with 50× Neural induction supplement (NIS) (Gibco) (to final dilution of 1× and final concentration of 2%), and seeded at 3 × 10 5 cells/ml.Medium was changed every 3-4 days and cultures were maintained for up to 14 days post initiation of differentiation.By day 3 of differentiation cells were becoming adherent with the proportion of adherent cells increasing over time in culture.

Inflammatory stimulation with IFNγ or LPS, and V2 (pre-)treatment
For IFNγ, the concentration used was 500 U/ml for 24 h, and for LPS 100 ng/ml for 24 h, as these conditions have been used in a variety of microglia/mixed cell culture studies [19][20][21].Pre-treatment with V-2 (Sigma, product code 03980585-10 MG), reconstituted in dimethyl sulfoxide (DMSO) to a final concentration of 50 ng/ml (≈80 nM, as per [16]) was undertaken for 1 h before treatment with the above inflammatory stimuli, with V2 not being removed from the incubation medium during the incubation with IFNγ or LPS.To exclude the possibility that V2 on its own can lead to the secretion of the pro-inflammatory cytokines investigated, a 24 h V2 only control was used in some experiments.

RNA extraction, conversion, and qPCR
RNA was isolated using TRIzol® reagent (Life Technologies, UK) according to manufacturer's guidelines.RNA was converted to cDNA using highcapacity cDNA reverse transcription kit (Applied Biosystems, Warrington, UK).Gene expression was analysed using Applied Biosystems Fast 7500 Real-Time PCR System (Applied Biosystems, UK).The comparative CT method was used to calculate relative gene expression using pre-designed probes obtained from ThermoFisher (namely PU.1, Iba1, CD45, P2y12R, CX3CR1, and TBP).The Ct results were normalised against TBP as a housekeeping gene.THP-1 cells were used as experimental controls.

Cytokine expression
Enzyme-linked immunosorbent assays (ELISAs) were used to assay the levels of cytokines in cell-free supernatants from different culture conditions (LPS ± V-2, IFNγ ± V-2).All cytokine kits for the cytokines assayed (IL-6, IL-1β, IL-18, and TNFα) were purchased from Abcam, and manufacturer guidelines were followed in all cases.
To examine the influence of V-2 on cells on its own, the levels of cytokines, growth factors, and chemokines in cell-free supernatants were measured using cytokine arrays (Abcam) at 1, 3, 6 and 24 h of incubation.A table of all 42 of the cytokines analysed is presented (Figure 5(a)), which can be used as a guide for interpreting the heat maps generated (Protein Array Analyser tool by Gilles Carpentier).Having corrected for background noise from each array by comparison with signals from negative controls/dots, the signal from each sample was then normalised (in pairs) against the positive controls, and then compared between conditions (+V2 or -V2).The following cytokines were selected as being of most interest, due to their changes between conditions, or because they have been found to play important roles to neuroinflammation and were then analysed further (i.e.densitometric analysis using Image J) (National Institute of Health, https://imagej.nih.gov/ij/):MCP-1, TNFα, ENA-78, GM-CSF, RANTES, TGF-β1, IL-4, IL-10, IL-12, IL-13, IL-6, and IL-1β (Figure 5(b)).

Statistical analysis
All data are presented as mean ± standard deviation (SD).A one-way or two-way analysis of Variance (ANOVA) was conducted for within group comparisons.GraphPad9 was used for statistical analyses.Statistically significant differences between treatments are denoted as (*) for p < 0.05, (**) for p < 0.01 or (***) for p < 0.001.Error bars represent standard deviation of three replicates.

Results
As shown in Figure 1(a), THP-1 cells in vitro are free floating round cells that expand in number exponentially after 3-5 days in culture.In the experiments outlined here, the aim was to differentiate these cells into microglia-like cells.Using neural induction supplement (NIS; TRIzol®) which is utilised for the differentiation of stem cells (SCs) into neural stem cells (NSCs; TRIzol®) we were able to differentiate THP-1 cells into neural macrophagelike cells that exhibited microglia characteristics.By Day 3 in vitro, sphere-like clumps of semi-adherent cells appeared in the cultures (Figure 1(b)).The cells within these clumps were relatively small in size based on cell diameter measurements (approx.11 μm).At Day 7 (Figure 1(c)) and again at D14 (Figure 1(d)), the adherent cells exhibited characteristic microglia morphological characteristics, with amoeboid and ramified morphologies visible.The cells increased in size accordingly to approx.19.2 μm at D7 and as large as approx.82 μm at D14.By D14, there were very few undifferentiated cells present in the culture.The characteristics of the differentiating cells are summarised in Table 1.By comparison, cells differentiated with PMA into dTHP-1 cells were observed to be less heterogenic with elongated adherent cells, however many still remain immature THP-1-like (round cells) (Supplementary figure S2a).

Expression of microglia markers using qPCR and immunohistochemistry
Microglia cells, which are CNS-resident cells, express many of the same markers as infiltrating immune cells [22].Therefore, distinguishing between these two cell types can be challenging.A microglia-specific marker has been reported -TMEM119 [23], and accordingly we aimed to characterise its expression throughout the mgTHP-1 differentiation process.We also investigated expression of other markers, such as CX3CR1 (a chemokine receptor previously reported to be characteristic of microglia and a subset of monocytes [24]), P2Y12R (a chemoreceptor for ADP which is downregulated during microglial activation, but is expressed highly under normal conditions [25]), PU.1 (a member of the Ets transcription family which is involved in the development of microglial cells [26]), Cd11b (an integrin also known as complement receptor 3α which is known as a microglia marker [23]), CD45 (a tyrosine phosphatase also known as PTPRC, which has been reported to be expressed in microglia [27]), and Iba1 (a Ca 2+ -binding protein involved in actin bundling, cytoskeletal remodelling and phagocytosis, which is expressed in all microglia, but also myeloid cells, and is seen to increase with microglial/macrophage activation [28]).All of the above are regarded as established and widely used microglia markers -without, however, having been shown definitively to be exclusive to this cell type (as analysed by [29]).
Firstly, we used qRT-PCR to assess the expression at the mRNA level of Iba1, CX3CRI, PU.1, P2Y12R and CD45 in experiments in which THP-1 cells were differentiated with NIS for periods of D3, D7 and D14 in vitro.Expression levels were expressed as relative fold change (versus TBP as a housekeeper gene) in the different conditions compared to non-differentiated THP-1 cells.THP-1 cells differentiated with PMA were also assayed (dTHP-1 cells) with similar findings being observed to those seen for undifferentiated THP-1 cells [data not shown].Interestingly, unlike for TBP, significant changes in GAPDH expression were observed over time in vitro, with significantly increased GAPDH expression levels seen at D3 (see Supplementary Fig S1B).This was accompanied by significantly increased intracellular ATP levels at D3 (see Supplementary Fig S1C ), which suggests that cells undergo metabolic as well as morphological changes during the mgTHP-1 differentiation process.As shown in Figure 2(a), we observed a significant increase in Iba1 expression with a peak at day 3 (p < 0.0001) that declined at day 7 (albeit still significantly greater expression compared to control THP-1 cells (p < 0.0001)) and by day 14 expression had returned to control levels.Similarly, the expression of PU.1, which is involved in the development of microglia, was elevated at the beginning of the differentiation process (day 3, p < 0.001 vs control, and p < 0.05 vs days 7 and 14), and then its expression reduced (Figure 2(b)).Interestingly, CD45, which has been reported to be expressed in lower levels in microglia compared to other myeloid cells [30], was expressed in significantly lower levels at day 14 (ie. in fully differentiated mgTHP-1 cells) than in control monocytic THP-1 cells (p < 0.05; Figure 2(c)).In contrast, markers which are reported to be strongly expressed under normal conditions in microglia cells [31], namely CX3CR1 (Figure 2(d)) and P2YR12 (Figure 2(e)), showed higher expression in day 14 cells (ie.mature mgTHP-1 cells) compared to control (p < 0.0001).
Secondly, cultures were immunofluorescently stained on D3, D7 and D14 and viewed using fluorescent microscopy (Figure 3).Over time, the cells showed similar morphological maturation to that previously shown in Figure 1.The expression of TMEM119, Iba1 and CD11b was determined.TMEM119, a transmembrane protein of unknown function, is a developmentally regulated and highly specific cell-surface marker of microglia, which is not expressed by macrophages or other immune or neural cell types [27].Interestingly, as shown in Figure 3, we Table 1.Cells' morphological characteristics during different stages of the differentiation process.For cell shapes, the different morphologies are mentioned in order of them being observed (e.g.In day 7, more clumps were observed within the culture, and only a small number of microglia-like and single non-adherent cells were observed).Cell diameter, signifies the maximum and minimum distance between the cytoplasmatic extensions of the cells.Data are summarised from n = 5 images.found the differentiation of THP-1 to mature mgTHP-1 cells resulted in TMEM119 positive cells in our cultures at all time points assayed.Similarly, at all time points assayed, positive expression of the markers CD11b and Iba1 was also observed, with co-staining of Iba1 and TMEM119 being evident at each time-point.

Functional assays using ELISA and cytokine arrays
As compared with peripheral macrophages, microglia demonstrate some similarities, but also some unique features, with regard to phenotype polarization; this facilitates their distinct innate immunological functions.Microglia can be stimulated by LPS or IFN-γ to an M1 phenotype for expression of proinflammatory cytokines, such as TNFα, IL-6, IL-1β and IL-18 or by IL-4/IL-13 to an M2 phenotype for resolution of inflammation and tissue repair [3].We characterised the M1 phenotype of mgTHP-1 cells at D14 in culture by determining production of the above cytokines following treatment with either LPS or IFN-γ.Based on a preliminary time-course analysis (data not shown) it was decided to compare expression in cultures exposed to the stimulus for 24 h.When compared to vehicle control (DMSO) treated cells, it was observed that LPS and IFN-γ treatment did not elicit any changes in cytokine expression, as shown in Supplementary Figure S3.LPS or IFN-γ induced significant increases in TNFα secretion compared to untreated control mgTHP-1 cells (P < 0.0001 in both cases; Figure 4(a)).Significant increases in IL-6 expression were also seen in response to LPS (p < 0.001) and -to a significantly more marked extentto IFN-γ (p < 0.0001; Figure 4(b)).Levels of IL-1β secretion from mgTHP-1 cells were low (19.83 ± 0.492 pg/ml [control] − 31.41 ± 1.802 pg/ml [LPStreated]), but nevertheless a statistically significant increase in response to LPS (p < 0.0001), and a smaller but significant increase in response to IFNγ (p = 0.026) were seen (Figure 4(c)).Finally, mgTHP-1 cells treated with LPS or IFN-γ exhibited significantly increased expression of IL-18 (p < 0.0001 in both cases), which was significantly more pronounced in response to LPS than to IFN-γ (Figure 4(d)).
Overall, these data show that mature mgTHP-1 cells (differentiated for 14 days in vitro with NIS) are immunogenic in response to inflammatory stimuli, and thus can exhibit an M1-like phenotype capable of producing significant amounts of proinflammatory cytokines.
Vicenin-2 (V-2), isolated from a traditionally used medicinal plant Artemisia capillaris, is a 6,8-di-C-glucoside of apigenin which has been previously reported to possess a wide variety of pharmacological activities including antioxidant, anti-inflammatory, anti-cancer, and hepatoprotective [32].We have previously shown its potent anti-inflammatory effect on dTHP-1 macrophage-like cells [16], and therefore, we assayed V-2's effectiveness in modulating mgTHP-1 responses to LPS and IFN-γ.As described above, expression of the cytokines TNFα, IL-6, IL-1β and IL-18 was assayed via ELISA, and responses at the 24 h time point assessed.As shown in Figure 4, our data indicates that V-2 exerted significant antiinflammatory suppressive effects on LPS-induced expression of TNFa, IL-1β and IL-18 (but that this effect did not achieve statistical significance for IL-6), while for IFN-γ V-2 had a significant antiinflammatory suppressive effect on all cytokines assayed (p < 0.0001 in all cases; Figure 4).(a) TNFα secretion was negligible at baseline, but was strongly induced after treatment with either LPS or IFNγ, with pre-treatment with V2 significantly lowering the response in both cases.(b) IL-6 secretion from the cells significantly increased (compared to control) upon treatment with either pro-inflammatory stimulus, with IFNγ causing significantly higher secretion of the cytokine compared to LPS.Pre-treatment with V2 appeared to decrease (albeit not significantly for LPS) the secretion of IL-6 in both cases.(c) IL-1β secretion was low (<40 pg/ml) in all cases, but was significantly induced when cells were exposed to LPS, or -albeit to a lower degree -after exposure to IFNγ.Again, V2 pre-treatment significantly blunted the response in both cases.(d) IL-18 secretion was significantly increased following exposure to either IFNγ or LPS, with LPS inducing a significantly stronger response; V2 pre-treatment significantly blunted the response in both cases.n > 3 in all cases; * represents P < 0.05; **** represents P < 0.0001.
Given the M1-M2 dichotomy of macrophage polarisation, and as V-2 appeared to have anti-inflammatory effect and was able to blunt responses to proinflammatory stimuli, investigation of the effects of V-2 alone on D14 mgTHP-1 cells was used to confirm its ability to polarise mgTHP-1 cells towards an M2 phenotype.Cytokine arrays were used in which a variety of (neuro)inflammation-related cytokine antibodies are adhered to a solid matrix, in order to identify potential changes in cells' secretomes.As discussed in the Methods section, the following cytokines were selected for quantitative analysis using densitometry: MCP-1, TNFα, ENA-78, GM-CSF, RANTES, TGF-β1, IL-4, IL-10, IL-12, IL-13, IL-6, and IL-1β (see Figure 5(b)).In accordance with the ELISA data presented above, at 24 h of V-2 treatment on D14 cultures, there was decreased intensity of signal for MCP-1, and an increase in the intensity of signal for RANTES, TGF-β1, IL-10, IL-13 and IL-6, all of which are regarded as M2-related cytokines (for review see [33]).Although there is some variation in the dataset, this data broadly corroborates the ELISA data, in that it provides evidence that V-2 can both suppress M1 polarisation as induced by M1-related stimuli (see Figure 4, and MCP-1 graph in Figure 5(b)), and synergise M2 polarisation as induced by anti-inflammatory cytokines such as IL-10 and IL-13 (see Figure 5(b)).

Discussion
This paper describes the development and characterisation of a novel in vitro model for neural macrophages with microglia-like characteristics which we define as mgTHP-1 cells.As shown in Figures 1 and  3, the cells exhibit the typical morphological and functional characteristics of microglia.The differentiation process from THP-1 cells to microglia-like neural macrophage cells includes a stage at which the cells are amoeboid and clump together (D3, D7), but by D14 the cells adopt a mature ramified morphology.This is in accordance with the reported developmental morphology of microglia cells where a typical amoeboid cellular morphology is observed in early foetal development, progressing to a ramified mature microglial morphology after birth [34].Similar phenotypic features have also been reported in other studies of microglia [35,36].Importantly, these phenotypes are not fixed, with microglia in their scavenging state shown to be ramified, whereas when activated, microglia become morphologically amoeboid.Moreover, ramification is associated with the downregulation of markers associated with activation of microglia, such as Iba1, which we (and others [37][38][39][40]) have shown to be downregulated in activated microglia (eg. in day 14, when compared to day 3 -see Figure 2).mgTHP-1 cells appear more heterogous in nature with fewer cells presenting with the THP-1 like morphology, a characteristic found when differentiating THP-1 cells with PMA into dTHP-1 macrophage-like cells (see Supplementary Fig S2A).We also show that during NIS-induced differentiation, mgTHP-1 cells over time show increased expression of SOX and nestin, markers of neural identity (see Supplementary Fig S2b).However, it should be noted that in cell culture conditions, there are limitations with many factors ) were used according to the manufacturers' recommendation; results appear as heatmaps (see left-hand side of [ii]), from which the respective levels of cytokines of interest in tissue-culture supernatants from Day 14 mgTHP-1 cells ± treatment with V-2 (80 nM; 24 h) can be discerned.For selected cytokines (see text for details), heatmap dot intensity was densitometrically quantified using image J software (National Institute of Health, https:// imagej.Nih.gov/ij/), and illustrated using bar graphs in each case (see left-hand side of [ii]).Each dot represents N = 1 (from n = 3 biological replicates pooled in each case); each bar represents a pair of dots.
influencing morphology, including the matrix in which they are grown and the range of compounds to which they are exposed.For example, ramification may be less obvious in 2D cultured cells, as the cells are not tightly packed within the tissue [41,42].It should also be noted that maintaining the differentiation of the cultures over time is challenging, as shown by the time-dependent decreases in expression of certain markers in our study; accordingly, an undertaking for the future is to identify parameters for sustained differentiation and cell yield within the cultures over time.
It has been proposed that glucose metabolism exerts transcriptional control over microglial activation, and that the homeostatic phenotype of microglia in culture preferentially utilize oxidative metabolism [43].Glucose enters the cell through different transporters (GLUTs) but under inflammatory conditions, GLUT1 expression is upregulated to increase glucose uptake and promote glycolysis [43].GAPDH expression increased significantly by D3 (Supplementary Figure S1B), which may be indicative of activationrelated changes in glycolysis corresponding to the first significant changes in cell morphology seen at D3.Given that ATP influences microglia differentiation [44] we confirmed that the ATP content of the cells at D3 was significantly higher than in either the initial control THP-1 cells or D7 and D14 mgTHP-1cells (Supplementary Figure S1C).The significant increase in DNA 5-mC methylation we report at D3, as compared to both D7 or D14 (see Supplementary Fig S1A ), corroborates previous data showing that glycolysis accompanies the differentiation of monocytes into active macrophages [45] and microglial inflammatory activation [46].While methylation is often associated with gene silencing, it can also be associated with upregulation and modulation of gene expression, particularly that involved in cell differentiation [47,48].Combined, these data are indicative of changes in cellular metabolism (possibly involving epigenetically triggered increases in glycolytic flux), which were observed in the differentiation of these cells to a microglial-like phenotype.
Microglial marker expression was further verified using gene expression analysis.From the data presented in Figures 2 and 3, it is evident that these mgTHP-1cells exhibit microglia-like characteristics with the markers Iba1, PU.1 and CD45 exhibiting similar trends in terms of expression peaking at D3 in all cases.PRY12R and CX3CR1 were also shown to have increased expression; albeit at the later D14 timepoint.The increase in the expression of genes associated with neural phenotype in these cells coincided with the microglial-like characteristics of our cultures in response to their differentiation in the presence of NIS.NIS has to the best of our knowledge not been previously reported in the maintenance or transformation of monocytic cells such as THP-1 cells.Whereas NIS is typically used for the differentiation of pluripotent cells into neural precursor cells, here we used it to differentiate THP-1 cells to a neural lineage.Thus, whilst our protocol is unique in using a readily available and validated myeloid cell line closely related to microglia and the use of NIS, the data is in accordance with other published in vitro models of microglia-like differentiation [49][50][51][52][53][54][55][56].
THP-1 cells were investigated in this study in part due to their well-characterized ability to demonstrate phenotypic changes associated with M1 and M2 polarization, which is also known to be highly relevant to the function of microglia.M1 macrophage activation is associated with the release of pro-inflammatory cytokines, including IL-1β, IL-18, IL-6, and TNFα [57].Increases in IL-18 and IL-1β secretion are also induced via the activation of the NLRP-3 inflammasome [58], which is highly expressed in M1 phenotypes [59], and has been found to have a regulatory role with regard to M2 polarisation [60].LPS and IFNγ are both known to polarise macrophages into an M1 state [61]; thus, the data presented in Figure 4 confirm the hypothesis that mgTHP-1cells retain macrophage-like abilities for M1 polarisation, as LPS or IFNγ significantly upregulated the secretion of the above pro-inflammatory cytokines.TNFα, IL-1β and IL-6 were secreted in negligible amounts by untreated mgTHP-1 cells, which is in agreement with findings where microglia cells have a low baseline secretion of TNFα and IL-6 [62,63], while the increase in IL-1β secretion (albeit in relatively low amounts) shown in Figure 4 is similar to that previously reported [64].Finally, IL-18 has been shown to be continuously secreted in microglia as well as monocyte-derived macrophages, with LPS enhancing the secretion in microglia cell cultures [65], which again is corroborated by the data presented here in mgTHP-1 cells (see Figure 4).
Conversely, V-2 is a dietary flavonoid that research in our group and others has previously shown to shift macrophage-like cells towards an M2 phenotype [16,18,66].Overall, we observed a significant blunting of proinflammatory cytokine secretion from cells pretreated with V-2 and then exposed to LPS.Furthermore, a cytokine array analysis of D14 cells in the presence or absence of V-2 at different time points identified additional cytokines (namely RANTES, IL-10, IL-13) that are produced by the mg-THP-1 cells and whose expression is increased by the addition of V-2 to the culture.These cytokines are typically associated with an M2 phenotype (for example, IL-10 is a potent regulator of inflammatory cytokine secretion [15]), and increased IL-10 expression in response to V-2, as shown in Figure 5, indicates that mgTHP-1 cells can be polarised towards an M2 phenotype by stimuli such as V-2.Although M2 macrophages are increasingly being sub-classified into subtypes including M2a, M2b, M2c, 2d and M(Hb), Mox, and M4 that play complex roles in inflammatory disease [67,68], we did not investigate our mgTHP-1 model for its ability to generate M2 sub-types (but that is an area for future investigation).There is a growing body of research exploring a vast range of anti-inflammatory compounds (in particular flavonoids) that could be utilised in these experiments to test for functional effects within our model.However, we chose to focus on V-2 based on previous experience of working with this compound and having an understanding of its mechanism of action.Future and ongoing work will look to expand our data on potential compounds (for a recent review see [69]).
The present findings regarding V-2 are in agreement with those reported by Hassan et al. [16], who proposed that such changes are due to a decrease in NLRP3 expression, and an increase in TET-2 expression.TET2 catalyzes the oxidation of 5-methylcytosine to 5-hydroxymethylcytosine, leading to DNA demethylation and through this action restraining inflammation by enhancing IL-10 expression [16,[70][71][72][73]. Interestingly, flavonoids have been shown to rescue activated microglia from cell death under certain conditions of activation [74] which could have important implications in vivo [75,76].Chemokines such as MCP-1 which not only attract monocytes and macrophages but can also activate them, have been found to be secreted by neurons and astrocytes under inflammatory conditions [77,78].Thus, V-2 regulation of such chemokines in mgTHP-1 could have implications for its anti-inflammatory action via modulating the number of cells attracted in response to signals from microglia.
Overall, some similarities to monocytic THP-1 or macrophage-like dTHP-1 cells exist following differentiation of mgTHP-1 cultures in NIS.In particular, when comparing mgTHP-1 cells to dTHP-1 cells; notable similarities include the macrophage-like characteristics of cytokine secretion and expression of certain markers.However, mgTHP-1 cells show unique characteristics of neural macrophages, in particular expression of microglial-specific markers (see Figures 2 and 3) and markers of neural identity (see Supplementary Fig S2B), which to our mind make these cells a suitable cell model for future studies specifically related to neuroinflammation.
In conclusion, we have shown that, using NIS, we can differentiate THP-1 cells into mgTHP-1 cells that exhibit microglial-like characteristics.Furthermore, these mgTHP-1 cells respond to inflammatory stimuli by secreting proinflammatory cytokines such as TNFα, IL-6, IL-1β and IL-18, a phenomenon resembling polarisation to an M1 phenotype, and one that could be blocked (or at least blunted) by pre-treatment with V-2.Moreover, as flavonoids have been shown to not only change the polarisation of macrophages towards an M2 phenotype, but also rescue activated microglia, it can be concluded that the present study provides evidence for V-2 inducing an antiinflammatory M-2 phenotype in mgTHP-1 cells.Therefore, the present study supports the potential for mgTHP-1 cells to constitute a valid model for studying neuroinflammatory conditions and neuroinflammation in general.

Figure 1 .
Figure 1.Morphologies observed during differentiation of THP-1 cells into mgTHP-1 cells.(a) THP-1 cells formed a suspension culture, comprising spherical cells that did not form clumps.(b) after 3 days in the NIS differentiation medium, clumps started forming, which appeared to adhere to tissue-culture plasticware, with extensions visible at sites where they adhered (see arrows in day 3 photograph).(c) between days 3 and 14, and especially around day 7, clumps started to disappear, and microglia-like cells exhibiting amoeboid and/or ramified morphologies became visible.(d) by day 14, most of the cells exhibited microglia-like 'mgTHP-1' morphologies (see arrows in day 14 photograph); however, the total number of cells declined.Light microscopy images were obtained using a Nikon U-200 camera (Nikon UK Ltd, Kingston-upon-Thames, UK) attached to an inverted leica microscope (Leica Microsystems UK Ltd, Milton Keynes, UK).Scale bar represents 20 μm.Images are representative of at least n = 3 in each case.DM=NIS/RPMI differentiation medium (see Methods section for details of differentiation protocol).

Figure 2 .
Figure 2. Gene expression (as determined using qPCR) of five microglia-related genes [(a) Iba1; (b) PU.1;(c) CD45; (d) CX3CR1; (e) P2Y12R] during differentiation of THP-1 cells into mgTHP-1 cells.(a) Iba1 expression significantly increased at day 3; although it declined thereafter, it continued to be significantly elevated versus control in day 7 (but not day 14).(b) PU.1 expressionsignificantly increased to maximal levels at day 3, before undergoing a significant decline by days 7 and 14 to levels approximating those seen in undifferentiated THP-1 cells.(c) CD45 expression significantly increased to maximal levels at day 3, before undergoing a significant decline by day 7 to levels approximating those seen in undifferentiated THP-1 cells, and by day 14 to levels below those seen in undifferentiated THP-1 cells.(d) levels of CX3CR1 did not undergo any significant changes at day 3 or day 7, but were significantly increased by day 14 (e).P2Y12R expression underwent significant step-wise increases during the differentiation process, with the highest expression levels being seen at day 14.Undifferentiated THP-1 cells were used as controls, and results were normalised against the expression of the housekeeper gene TBP.n > 3 in all cases; * represents P < 0.05; ** represents P < 0.01; *** represents P < 0.001; **** represents P < 0.0001.

Figure 3 .
Figure 3. Immunofluorescence for THP-1/NIS differentiated cells at all three selected differentiation days for markers of interest (Cd11b, Iba1, TMEM119).For day 3, various multinucleated structures (red arrow Cd11b) and multicellular clumps (yellow arrows) as well as single cells (Iba1 stain red arrow) were apparent.At day 7 expression of the respective markers was seen in cell clusters.Day 14 different microglia-like morphologies were more apparent in the culture.Multinucleated cells were present (Iba1 stain), and both ramified and amoeboid cells were observed as well as multinucleated cells.Double labelled Iba1 and TMEM119 cells show co-labelling of cells at each time point.Green: protein of interest (TMEM119, Iba1, and Cd11b respectively), images representative of > n = 3. Cultures were counterstained with Hoechst a nuclear stain (blue).

Figure 4 .
Figure 4. Secretion (as measured by ELISA) of (a) TNFα; (b) IL-6; (c) IL-1β; (d) IL-18 from day 14 mgTHP-1 cells, following treatment with the inflammatory stimuli LPS (100 ng/ml; 24 h) or IFNγ (500 U/ml; 24 h), with and without pre-treatment with V2 (80 nM; 1 h).(a)TNFα secretion was negligible at baseline, but was strongly induced after treatment with either LPS or IFNγ, with pre-treatment with V2 significantly lowering the response in both cases.(b) IL-6 secretion from the cells significantly increased (compared to control) upon treatment with either pro-inflammatory stimulus, with IFNγ causing significantly higher secretion of the cytokine compared to LPS.Pre-treatment with V2 appeared to decrease (albeit not significantly for LPS) the secretion of IL-6 in both cases.(c) IL-1β secretion was low (<40 pg/ml) in all cases, but was significantly induced when cells were exposed to LPS, or -albeit to a lower degree -after exposure to IFNγ.Again, V2 pre-treatment significantly blunted the response in both cases.(d) IL-18 secretion was significantly increased following exposure to either IFNγ or LPS, with LPS inducing a significantly stronger response; V2 pre-treatment significantly blunted the response in both cases.n > 3 in all cases; * represents P < 0.05; **** represents P < 0.0001.

Figure 5 .
Figure5.Cytokine array data illustrating responses of mgTHP-1 cells to treatment with Vicenin-2 (V-2).cytokine arrays (Abcam PLC, Cambridge, UK) comprising a total of 42 cytokines (see[i]) were used according to the manufacturers' recommendation; results appear as heatmaps (see left-hand side of [ii]), from which the respective levels of cytokines of interest in tissue-culture supernatants from Day 14 mgTHP-1 cells ± treatment with V-2 (80 nM; 24 h) can be discerned.For selected cytokines (see text for details), heatmap dot intensity was densitometrically quantified using image J software (National Institute of Health, https:// imagej.Nih.gov/ij/), and illustrated using bar graphs in each case (see left-hand side of [ii]).Each dot represents N = 1 (from n = 3 biological replicates pooled in each case); each bar represents a pair of dots.