Development and characterization of a novel, megakaryocyte NF‐κB reporter cell line for investigating inflammatory responses

Essentials An easily detectable readout in megakaryocyte cell lines will enhance inflammatory research in these cells. Here, we report the development and characterization of a novel megakaryocyte NF‐κB‐reporter cell line (Meg‐01R). Multiple inflammatory molecules modulate NF‐κB activity in Meg‐01R cells. Meg‐01R cells respond to small molecule inhibitors such as IMD0354 and C87 that are known to inhibit NF‐κB activity upon stimulation with TNFα.


| INTRODUC TI ON
Platelets (small circulating blood cells) are known for their ability to regulate hemostatic, innate immune, and inflammatory responses. 1,2 A large number of studies have focused on elucidating the roles of platelets in the regulation of innate immune/inflammatory responses and the molecular mechanisms responsible for these activities. 1,[3][4][5] Furthermore, platelets are recognized as having prominent roles in the adaptive immune system. 1,2 However, only a relatively small number of studies have focussed on determining the impact on megakaryocytes (MKs), the precursors of platelets, on immune/inflammatory responses. 6,7 This is an under researched area, although the reactivity and proteome of platelets produced during inflammation are modified to accustom the pathological conditions. [8][9][10] MKs differentiate from hematopoietic stem cells (the progenitor cells from which innate immune cells such as neutrophils and monocytes are also derived), produce large quantities of proteins, and possess an extended invaginated membrane system for packaging into numerous platelets. 7 MKs in the bone marrow produce platelets by extending proplatelets into sinusoids, where the shear flow of blood leads to budding of platelets into the bloodstream. 7,11 Platelets are also produced in the mouse lungs, where MKs may interact with pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs); however, whether this function translates into humans remains unclear. 12 The environment surrounding MKs may alter the functions of platelets, for example, producing platelets that are more pro-aggregatory or aggressive toward pathogens, although the molecular mechanisms behind these actions are still poorly understood. [8][9][10] Studies have shown that in mice, lipopolysaccharide (LPS) from Gram-negative bacteria causes an increase in circulating platelet count. 8 Mice lacking functional Toll-like receptor (TLR)-4 (the receptor for LPS) have a significant reduction in the number of circulating platelets. 8 Furthermore, some studies suggest that LPS treatment increases platelet production from MKs, as well as MK ploidy, 13,14 although this may be due to the release of several factors from macrophages. 15 Hence, determining the impact of inflammatory responses on MKs and subsequent platelet production will aid in better understanding of the significance of MKs and platelets in the regulation of thromboinflammation in various pathophysiological scenarios.
Meg-01 is a megakaryoblastic cell line frequently used as a surrogate for elucidating signalling pathways and functions in MKs because of the difficulties associated with acquiring large numbers of primary MKs. [16][17][18][19] TLR2 is closely related to TLR4 and it signals through the same MyD88-dependent pathway. A previous study has reported that the treatment of Meg-01 cells and murine MKs with a TLR2 agonist, Pam3CSK4, resulted in increased MK ploidy, activated nuclear factor-κB (NF-κB), and altered protein expression 16 ; these results suggest that the MyD88-dependent pathway is likely to be active in MKs. Because of the lack of a reliable system to determine the role of MKs in the regulation of inflammatory responses, in this study we developed and charac-

| MATERIAL S AND ME THODS
The raw data that support the findings of this study are available upon reasonable request.

| Cell culture
Meg-01 cells were grown in RPMI-1640 media supplemented with 10% (v/v) foetal calf serum (FCS) and 2 mM l-glutamine (Sigma-Aldrich). Cells were kept at 37°C in a humidified 5% CO 2 incubator. Every 2 to 3 days, the cells were removed by scraping, counted, and resuspended at a concentration of 2.5 × 10 5 cells/mL in a vented suspension flask (Sarstedt). The media for Meg-01R cells was supplemented with 1 μg/mL puromycin for selection (Apollo Scientific).
All experiments were conducted in the absence of antibiotics and antimycotics.
HEK-293 cells were grown in high glucose DMEM containing 10% (v/v) FCS and 2 mM L-glutamine in a humidified 5% CO 2 incubator at 37°C.

K E Y W O R D S
Formyl peptide receptor, inflammation, megakaryocyte, NF-κB, Toll-like receptor 4

Essentials
• An easily detectable readout in megakaryocyte cell lines will enhance inflammatory research in these cells.
• Here, we report the development and characterization of a novel megakaryocyte NF-κB-reporter cell line (Meg-01R).
• Meg-01R cells respond to small molecule inhibitors such as IMD0354 and C87 that are known to inhibit NF-κB activity upon stimulation with TNFα.
Doubling time was calculated using Meg-01 and Meg-01R cells cultured in parallel under normal growth conditions in normal growth media using the formula:

| RT-PCR
Total RNA was extracted from Meg-01 and Meg-01R cells and underwent RT-PCR to determine the presence of transcripts for MyD88. Further details are included in Appendix S1.

| XTT assay
Cell viability assays XTT assays were conducted according to manufacturer's instructions (Sigma-Aldrich). Further details are described in Appendix S1.

| Luciferase assay
The luciferase assay performed in this study was based on previous experiments. 23 Briefly, 3.75 × 10 5 Meg-01R cells were seeded into a 24-well suspension plate in starvation media (RPMI-1640 with 2 mM L-glutamine) and left at 37°C in a 5% CO 2 incubator for 4 hours.
Following induction, cells were incubated for 24 hours. Meg-01R cells were removed and any remaining adherent cells were rinsed off with media and collected. Meg-01R cells were washed before resuspension in cell lysis buffer (Promega). Each sample was agitated for 2 hours and then centrifuged at 5000g for 5 minutes at room temperature. NF-κB-mediated luciferase activity was measured by the addition of luciferin (Promega) and the level of luminescence at all wavelengths was recorded using a SpectraMax iD3 multimode microplate reader (Molecular Devices).
Cells were fixed in 4% (w/v) paraformaldehyde for 10 minutes before blocking with FcR blocking reagent in PBS (nonpermeabilized) or 0.02% polybutylene terephthalate (permeabilized). 1/100 dilution of mouse anti-TLR4 antibody (Abcam) was used to detect TLR4 after incubation with cells for 1 hour at room temperature. A 1/300 dilution of Alexa Fluor 555-conjugated goat anti-mouse IgG antibody (ThermoFisher Scientific) was used to visualize binding of the anti-TLR4 antibody to its target. Finally, the nuclei of Meg-01 cells were stained with 1/2000 4′,6-diamidino-2-phenylindole before resuspension in Mowiol 4-88 (Sigma-Aldrich) and mounting on a microscope slide. Slides were then imaged using a Nikon A1R confocal microscope and images were processed using Nikon NIS-Elements software and ImageJ.

| Statistical analysis
Logarithmic dose-response curves for puromycin toxicity were generated using a four-parameter curve with variable slope. For cell doubling time, an unpaired Student's t test was used to compare the mean values. For grouped data, analysis was performed using multiple t tests with the type I error rate controlled for using the Holm-Sidak method. For experiments containing multiple comparisons, data were analyzed using a one-way analysis of variance (ANOVA) with Bonferroni's post-hoc test. All statistical analysis was conducted using GraphPad Prism 8 (GraphPad). Data are represented as mean ± standard deviation.

| Meg-01 cells express various signalling proteins involved in TLR4 pathways
Platelets are reported to contain all the signalling proteins that are involved in MyD88-dependent and MyD88-independent signalling pathways. 25 Figure 1A). Although all of these proteins were detectable in Meg-01 cells, surprisingly we could not detect MyD88 via this method.
However, MyD88 mRNA was detectable via RT-PCR using specific primers for MyD88 ( Figure 1B). This discrepancy may be due to the expression level of this protein being below the detection threshold of the immunoblotting technique or antibody sensitivity. Notably, a previous study 16 has shown that Meg-01 cells are able to respond Pam3CSK4; therefore, MyD88-dependent signalling is likely to be active in these cells.  The optimal concentration of puromycin required to eradicate ~80% of nontransfected Meg-01 cells was determined using an XTT assay ( Figure 2A). The results demonstrate that puromycin was capable of inducing cell death both over 24-and 48-hour time scales with a half maximal effective concentration of 760 and 560 ng/mL, respectively. Based on these results, a concentration of 1 μg/mL puromycin was used to select transduced cells, which contain a puromycin-resistance gene, in subsequent experiments.

F I G U R E 1 Presence of proteins involved in
To elucidate differences or similarities between the transduced Meg-01R and standard Meg-01 cells, the doubling time of these cells was compared. The growth rate of Meg-01 (50.7 ± 4.35 hours) and Meg-01R (53.2 ± 6.28 hours) cells cultured parallelly was not significantly different ( Figure 2B). Moreover, their viability was determined using an XTT assay and it did not show any significant differences after 6 hours except a slight reduction in the viability of Meg-01R cells observed when they were seeded at 2 × 10 5 cells/mL ( Figure 2C). This slight reduction may and Meg-01R at similar levels ( Figure 2E). Moreover, integrin β 3 was detectable in both permeabilized and intact Meg-01 and Meg-01R cells at similar levels ( Figure 2F). 19 To corroborate these data, GPIbα and integrin β 3 expression levels were analyzed using immunoblots, and they were found to not be significantly different in Meg-01 and Meg-01R cell lysates ( Figure 2G). Together, these data suggest that the transduction in Meg-01R cells did not affect the proliferation, growth, and major characteristics of these cells compared with Meg-01.

| Inflammatory molecules stimulate NF-κB activity in Meg-01R cells
To determine whether Meg-01R cells respond to various inflammatory molecules, luciferase assays to detect NF-κB-dependent bioluminescence were performed as a measure for NF-κB activ- Interestingly, although uLPS EC was incapable of inducing NF-κB activity from Meg-01R cells, the non-ultrapure preparation tested was able to stimulate significant NF-κB activity. These results demonstrate that Meg-01R cells produce functional luciferase as a marker for NF-κB activity and successfully respond to various inflammatory molecules.

| Small molecule inhibitors reduce TNF-αinduced NF-κB activity in Meg-01R cells
IMD0354, C87, and 1,8-cineole were reported as small molecule inhibitors that are able to affect TNF-α-induced NF-κB activity in other cell types. IMD0354 acts as an inhibitor of IKKβ to prevent the phosphorylation and subsequent degradation of IκB. 23,36 C87 acts as an antagonist to TNF-α by directly binding to TNF-α and thus has been proposed to disrupt the TNF-α-TNFRI/II complex. 37,38 The 1,8-cineole is a plant-derived compound that has been shown to inhibit nuclear translocation of p65 and degradation of IκBα. 24,39 Here, these compounds were tested to determine their impact on TNF-α-mediated NF-κB activity in Meg-01R cells.
In the absence of TNF-α, IMD0354 ( Figure  can also be determined using Meg-01R cells. F I G U R E 5 Impact of TNF-α pathway inhibitors on NF-κB activity in Meg-01R cells. Different concentrations of an IKKβ inhibitor, IMD0354 (A), a TNF-α antagonist C87 (B), and 1,8-cineole (C) were tested in both the presence and absence of 2.5 ng/mL TNF-α over a 24-hour period. Data represent mean ± standard deviation (n = 3) and statistical significance (*P < .05, **P < .01, and ****P < .0001) was calculated using a one-way ANOVA with Bonferroni post hoc test

| TLR4 is predominantly present inside Meg-01/ Meg-01R cells
To further scrutinize the lack of Meg-01R response to the ultrapure preparation of LPS, the level of TLR4 expression on the surface and inside of Meg-01 and Meg-01R cells was examined.
The results obtained using confocal microscopy demonstrate that TLR4 is largely detectable inside Meg-01 cells as only the permeabilized cells show strong binding to anti-TLR4 antibodies ( Figure 6A and 6B). The z-stack image in Figure 6C further demonstrates that TLR4 is expressed ubiquitously within the cytoplasm.
Additionally, the absence of TLR4 on the surface of Meg-01R cells was corroborated using a flow cytometry-based assay. Here, the binding of an anti-TLR4 antibody to the surface of Meg-01R cells was not

| CD14 and LL37 do not enable LPS to stimulate NF-κB activity in Meg-01R cells
To determine if TLR4-induced NF-κB activity could be promoted Meg-01R cells were co-incubated with uLPS EC or ultrapure Salmonella minnesota LPS (uLPS SM ) for 24 hours with/without CD14 or LL37. A physiologically relevant concentration of 2μg/mL was chosen for CD14. 45 The results show that CD14 did not significantly alter NF-κB activity in Meg-01R cells on its own, and it was not capable of promoting uLPS EC or uLPS SM to stimulate NF-κB activity ( Figure 7A). Furthermore, LPS chemotypes did not induce any activity on their own. When LL37 and uLPS EC or uLPS SM were added simultaneously ( Figure 7B) or following 15 minutes preincubation with LL37 ( Figure 7C), there was no change in NF-κB activity compared with LL37 alone ( Figure 7C). These results demonstrate that CD14 and LL37 do not modulate LPS-induced NF-κB activity.

| D ISCUSS I ON
The role of MKs during inflammation is not fully understood.
Thrombopoietin is known to stimulate the growth and development of MKs and it is removed from the plasma by platelets, which means that, during thrombocytopenia (which can be induced by inflammatory diseases such as sepsis 5 ), high levels of thrombopoietin are detected in the plasma. 7,46 Very few studies have been F I G U R E 7 Effect of LPS-binding molecules CD14 and LL37 to induce LPS-mediated NF-κB activity in Meg-01R cells. NF-κB activity was measured in lysates from Meg-01R cells that were stimulated with uLPS EC or uLPS SM in the presence of 2 μg/mL CD14 (A; n = 3), or 10 μM LL37 following no co-incubation before addition (B; n = 6) or 15 minutes co-incubation (C; n = 5). Data represent mean ± standard deviation and statistical significance was calculated using a one-way ANOVA with Bonferroni post hoc test. The P values (**P < .01 and ****P < .0001) are as calculated by GraphPad Prism  conducted to examine the effect of PAMPs using MK cell lines but they demonstrate that NF-κB can be activated. 16,28 Mice lacking TLR4 have significantly lower platelet counts than their wild-type counterparts but intravenous injection of a sublethal dose of LPS (0.2 mg/kg) leads to a significant increase in platelet count regardless of whether TLR4 was present. 8 Moreover, TLR2 and TLR4 have been shown to induce the production of interleukin-6 (IL-6), via NF-κB, in CD34 + cells, which leads to increased MK maturation and platelet production. 13   IκB prevents binding of p65 (a member of the NF-κB family of transcription factors) to its specific DNA promoter sequence but is degraded following phosphorylation by IKKβ and ubiquitinated before its degradation. 5,52,53 p65 is a transcriptional activator that enables transcription of inflammatory cytokines in response to PAMPs and DAMPs, such as ligands of TLRs and TNF-α. 5,53,54 The MyD88dependent pathway directly leads to NF-κB activation brought about by IκBα degradation; however, cross-talk between the two pathways mean that the MyD88-independent pathway can also induce NF-κB activation. 5,28 Meg-01 cells and MKs are reported to express TLR1, 2, 3, 4, and 6 and Meg-01 cells are frequently used as a surrogate for elucidating signalling pathways in MKs. [16][17][18][19]28,[55][56][57] Because these receptors all can induce NF-κB activation, we endeavored to develop a reporter cell line to determine the role of MKs in regulating inflammation.
Therefore, we transfected HEK-293 cells with a previously published and commercially available reporter plasmid with a puromycin-resistant gene (pGreenFire-NFκB-Puro) 23 to produce lentiviral particles that can stably transfect other cell types with this reporter con- Pam3CSK4 is a synthetic ligand for the TLR1/2 heteroreceptor and it has previously been shown to induce phosphorylation of NF-κB's p65 subunit and degradation of IκB over the period of 1 hour. 16 Phosphorylation of p65 was also shown to occur in murine MKs following 30 minutes of treatment with Pam3CSK4. 16 In accordance with this previous study, Pam3CSK4 was able to induce luciferase production downstream of NF-κB albeit not as strongly as the lowest concentration of TNF-α tested.
Poly(I:C) is a synthetic ligand for TLR3 and, although TLR3 couples to the MyD88-independent pathway, signalling downstream of TLR3 can lead to NF-κB activation. 28 Poly(I:C) and Poly(A:U) (another synthetic TLR3 ligand) have been shown to induce IκB degradation and phosphorylation of p65 in human CD34 + cells derived from umbilical blood. 28 In this study, Poly(I:C) was incapable of stimulating the NF-κB activity at any of the concentrations tested. This discrepancy may be due to changes in the expression levels of specific receptors and characteristics observed during MK maturation.
The preparation of LPS EC can alter the characteristics of the responses induced by this ligand therefore an ultrapure and a non-ultrapure version were both tested. This has been hypothesized to be due to the presence of bacterial contaminants in non-ultrapure preparations that can activate other pro-inflammatory receptors such as TLR2. 20,51 The ultrapure version of LPS EC was unable to activate NF-κB; however, the non-ultrapure preparation was capable of inducing activity at a range of concentrations. This finding lends more weight to the hypothesis that many of the effects observed when using non-ultrapure LPS EC may be induced by contaminants in the preparations.
Moreover, FPR2/ALX ligands were tested on Meg-01R cells to determine if they could affect NF-κB activity in this cell type. The agonist, LL37, 31 was able to significantly inhibit NF-κB activity on its own over a period of 24 hours in contrast to the annexin A1-mimetic, Ac2-26. 33 Amyloid-β 1-42 is a neuropeptide involved in the progression of Alzheimer's disease. Indeed, it has been related to inflammation in the brain and is capable of inducing NF-κB activity in a glioblastoma cell line. 23,34 However, amyloid-β 1-42 was incapable of activating NF-κB at any of the tested concentrations in this study.
This may be due to the concentrations tested being subthreshold.
Interestingly, the synthetic lipoxin A4-analogue, BML-111 was incapable of modulating NF-κB activity in the absence or presence of TNF-α. 35 However, this may be due to signalling via a different downstream pathway.
A potential application of this novel reporter cell line is for the identification of compounds that affect inflammatory responses in MKs during disease states in an NF-κB-dependent manner. 15 To evaluate this, Meg-01R cells were treated with an IKKβ inhibitor, although the mechanism behind this is currently unknown. 1,8-cineole had no effect in either stimulated or unstimulated cells.
Meg-01 cells have previously been described to express TLR4 when the cells were fixed with methanol, a fixative known to be capable of permeabilizing cells. 56 To confirm the distribution of TLR4 in Meg-01 cells, they were fixed with 4% PFA (for immunocytochemistry) or 0.2% formyl saline (for flow cytometry) and then cells were left intact or permeabilized with PBS containing Triton-X100.
Moreover, two different anti-TLR4 antibodies were used to confirm its presence depending on the detection method. With both methods, TLR4 was not detectable on the surface of Meg-01 cells; however, it was detectable intracellularly. Previous studies have suggested that MKs express TLR4 on their surface but this increases during MK maturation. 57 Although TLR3 is predominantly detectable intracellularly, poly(I:C) is capable of being internalized leading to its activation. 28 In contrast, LPS is not capable of crossing the cell membrane on its own. 43 LPS is internalized after binding to CD14-TLR4 to enable it to enter endosomes and induce MyD88-independent signalling. 40 shown to be independent of TLR4 and dependent of caveolae and functional epidermal growth factor receptor. 43 Three treatments -2 μg/mL CD14, simultaneous addition of 10 μM LL37 and LPS, addition of 10 μM LL37 and LPS after 15 minutes of co-incubation -were tested to determine if these molecules could deliver LPS to its target receptor. Two μg/mL CD14 had no effect on NF-κB activity in Meg-01R cells and was unable to promote LPS to induce luciferase synthesis. Furthermore, in both conditions, LL37 significantly inhibited NF-κB activity; however, this was not modulated by the simultaneous addition of LPS, nor the co-incubation of LL37 and LPS. This may suggest that TLR4 in Meg-01R cells is not found in the endosome and therefore it may not stimulate MyD88-independent signalling nor interact with internalized LL37/LPS complexes.
In conclusion, here we developed a MK reporter cell line and demonstrated that MKs are responsive to a range of PAMPs and DAMPs, and their pathway inhibitors. Hence, this reporter cell line can be applied to screen a broad range of inflammatory molecules that act via NF-κB to determine their functions in MKs.

CO N FLI C T O F I NTE R E S T
The authors declare no competing interests.