Chau N.P. Do1, Jane A. Awuh2, Anne Marstad2, and Trude H. Flo2
The Mycobacterium species include pathogens causing serious diseases in mammals.Mycobacterium avium-intracellular complex (MAC), while it has low virulence to normal people, it is a high potential pathogen causing Tuberculosis-like disease in AIDS patients. Keap1 have studies by many researchers as a potential treatment in cancers. And its multi-binding property might give it an important role as a link between ROS, inflammation and autophagy. We knocked down the Keap1 in primary human monocyte-derived macrophages and investigated its role in immune response mechanism and killing of M.avium in human primary macrophages. Here, we found that not only IKKβ is under Keap1’s effect, but also all IKK complex molecules and TBK1 area affected by Keap1. Our study was done in vitro and limited to the human monocyte derived macrophages only, in summary, it does not necessarily mean that the Keap1 gives disadvantage in controlling the M.avium infection. The role of Keap1 in the adaptive immune system should be studied more to understand the effect of Keap1 in whole system.
The Mycobacterium species include pathogens causing serious diseases in mammals. One of them is tuberculosis (TB), and most TB-infected patients got infection from Mycobacterium tuberculosis. Commonly,AIDS patients who are with TB-like disease or who easily get TB development have infection of Mycobacterium avium-intracellular complex (MAC), including Mycobacterium avium (M.avium) consideredas an opportunistic mycobacterium causing non-tuberculous pulmonary disease in the later stages of AIDS, in immuno-compromised people, genetic-deficiency individuals, healthy children, in some elderly men andwomen with Lady-Windermere syndrome. Cervical lymphadenitis is also found in non-tuberculous M.avium infection in children, who do not develop pulmonary diseases (1). Mycobacteria infected host macrophages, and with the combination of innate and adaptive immune systems, an effective anti-microbial defense will be occurred.
Like other microorganism, M.avium has several moleculars for pattern recognition receptors (PRRs), the host’s protectors against disease. This generation of an enormous numbers of cells and molecules actingtogether in a dynamic network that specifically recognizes and eliminates an apparently limitless variety of foreign invaders. Five different classes of PRR families have been identified: Toll-like receptors, Retinoicacid-inducible gene-I-like receptors, Nucleotide-binding oligomerization domain-like receptors, C-type lectinreceptors, and DNA sensors (2, 3). They have different cellular locations that are able to respond to both extracellular and intracellular microbes. Activation of PRRs is a complex downstream signaling event leading to the secretion of various inflammatory cytokines, chemokines, and antimicrobial proteins. This will result inactivating the first defense line of antigen- presenting cells, production of pro-inflammatory cytokines, type 1interferons (IFN) and chemokines, promoting the phagocytosis to engulf and digest the pathogens, and the expression of the peptide antigen on the cell surface, and influencing as a bridge to activate adaptive immune response by activating T-cells and B-cells (4-6). Under infection, cells are also under stress conditions with various cytokines, chemokines, and inflammasomes from PRR network. This leads to the oxidative stressin cells (7). Oxidative stress will lead to the generation of reactive oxygen species (ROS) and electrophiles, which have a high impact on survival, growth development, and evolution of living organisms (8). To protect against these stress factors and to control the balance of the cellular redox state, eukaryotic cells have developed complex system to detoxify and maintain cellular homeostatics, which consists of four categories of several enzymes, such as thioredoxin, superoxide dismutases, glutathione peroxidases andcatalases (9-11). The nuclear factor (NF)-erythroid 2 (E2)-related factor 2 (Nrf2) is well established as themain mediator of cellular adaption to redox stress. In additional to two previous biological process, autophagy also has immunological role in the infection of microorganisms invading intracellularly (12, 13). Several studies have shown that autophagy is one of the immunity and inflammation downstream outcomes of PRR signaling (14), displaying a prominent property that can effectively eliminate intracellular bacteria, such asmycobacteria.
The identification of Kelch-like ECH-associated protein 1 (Keap1), an inhibitor of Nrf2, was firstannounced by Itoh and colleagues in 1999 (15). Later, many studies about this molecule arose and gave more characteristics for this Keap1 molecule (16). Keap1 exists as a homodimer binding one Nrf2 and it has threemajor domains (Figure 1a and b). The N-terminal BTB (broad complex, tramtrack, bric-a-brac)/POZ (poxvirus, zinc finger) is the protein-protein interaction domain containing the Cys151 residue, which binds to Cullin 3(Cul3)/RING-box protein 1 (Rbx1) E3 ubiquitin ligase complex. This domain also acts as the stress-sensingdomain. The linker region, or intervening region (IVR), is cysteine-rich domain consisting Cys273 and Cys288.This region contributes to the activity of Keap1 as the second stress-sensing domain (17). The C-terminal Kelchdomain (CTR) is linked with a double glycine repeat (DGR) containing six conserved Kelch repeat sequences forming a β-propeller structure. This domain is where the Keap1 binds to Nrf2 through the Neh2 domain on Nrf2. In normal state, the Keap1 serves as an adapter for Cul3/Rbx1 E3 ubiquitin ligase complex and leads to the ubiquitination and degradation of Nrf2 through the 26S proteasome.
Fig. 1: Kelch-like ECH-associated protein 1 (Keap1) protein. A. Represented 3D-structure of Keap1 structure binding Nrf2 (taken and modified from Zhang et al. (18)). B. The arrangment of protein with three major domains. The N terminal BTB (broad complex, tramtrack, bric-a-brac) domain is responsible for Cul3/Rbx1 E2 ubiquitin ligase complex binding. The intervening region (IVR) distributes to the activity of Keap1. These two BTB and IVR consist of important cysteine residues for stress sensing and the C-terminal with 6-repeated-Kelch domain/double glycine repeat (DGR) has function of binding to Neh2 region on Nrf2 (19). C. Mechanism of Keap1 as interplay factor of oxidative stress, autophagy and inflammatory. Keap1 has direct interaction with IKKβ, Bcl2 and Nrf2 and guides them for ubiquitination (Ub). p62 has negative effect on Keap1 through binding. So that, Keap1 indirectly participates in the ROS, inflammation and autophagy (taken and modified from Stepkowski and Kruszewski (20)).
From the time it was discovered, there are many studies showing the role of Keap1 as a key element in cancer treatments (Figure 1c). In the hypothesis paper, Stepkowski showed an interest in the molecular crosstalkbetween the Nrf2-Keap1 signaling pathway with autophagy and apoptosis (20). Most recently, Tian and others also give a mini review on the multi-functional Keap1 on “killing three birds” which are Nrf2, IKKβ, and Bcl-2 and p62 in ROS, inflammatory and autophagy, respectively (21).
Wakabayashi and others proved the significant role when they ablated Keap1 in mouse (22). They showed evidence that Nrf2 constitutively accumulated in the nucleus and stimulated transcription of cytoprotective genes and produced phase II detoxifying enzymes and mice with Keap1-deficiency died postnatally. More examinations showed the hyperkeratosis in esophagus and forestomach, together with the induction of keratin K1, K6 and loricrin. Sykiotis and Bohmann added more to Keap1’s importance with studies on Drosophila (23). It is shown that Drosophila males with Keap1 loss-of-function mutation had longer life-time and higher tolerance to oxidative stress than others. Many studies about cancers also state the importance of Keap1 in Nrf2-interaction. It was reviewed by Hayes that Nrf2 is accumulated and mutated due to dysfunction Keap1 in lung, breast, and bladder cancers causing problems in prognosis (8, 19). The most recent study from Williamson and Johnson showed that the up-regulated activation of Nrf2-ARE pathway,when Keap1 was knockdown by siRNA, increased the protection against oxidative stress in primary astrocytes,and also protected against 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-induced dopaminergic neuron damage/loss in mice (24).
Recently, IKKβ was identified as another substrate for the Keap1-Cul3-E3 ligase complex that influenced the expression of NF-κB (25, 26). They showed that IKKβ contains an E(T/S)GE motif similar tothe one of Nrf2 and Keap1 binds to IKKβ through the same DGR region directly (21, 25). Moreover, IKKβ is one of the IKK complex molecule, and an important molecule in NF-κB signaling pathway. The transcriptionfactor NF-κB is activated through the releasing of IκB induced by phosphorylated IKK complex. It had beenproven in these studies that Keap1-IKKβ had effect on NF- κB expression. Lee et. al. made Keap1 depletionand observed up-regulation of NF-κB-derived tumor angiogenic factors (26). Meanwhile, Kim et. al. made an introduction of Keap1 gene and NF-κB reporter gene into HEK293 cells, they observed that NF-κB activity was inhibited. And knockdown Keap1 gene gave a higher basal signal of TNFα-induced activity compared to control cells (25). Overall, Keap1 and IKKβ interaction has an effect on NF-κB which is one of theinflammatory signaling pathways.
Bcl-2 is B-cell lymphoma 2, which was first discovered as a regulatory protein regulating cell death orapoptosis. Later, it was shown that Bcl-2 also is a multifuctional protein influencing many cellular processes (27). Bcl-2 has been known with antiapoptotic and antiviral effects on preventing virus-induced cell death and restricting viral replication (28). p62, also called Sequestosome 1 (SQSTM1), is a linking polyubitinated protein to the autophagic system (29, 30). p62 is involved in various cellular processes by serving as amolecular hub mediating various protein- interaction (31, 32). Recently, Niture and Jaiswal had shown the interaction between Bcl-2 and Keap1 through cell line models (33). Structure of Keap1 protein has multibinding domain DGR, which binds to Bcl-2 at BH2 domain. And, multidomain adaptor p62 protein binds tovarious signaling proteins, including Keap1-interacting region and TRAF6 binding site, for oligomerization and forming cytosolic speckles (20). An antioxidant signaling of p62 is showed through Keap1-interactingregion (KIR), similar to Nrf2, thus p62 acts as a Keap1 binding competitor with Nrf2 (34). Thus, Keap1 might have indirect interaction with autophagy via Bcl-2 and p62.
Keap1 might have a role as a link between ROS, inflammation and autophagy and these pathways have an important role in controlling M.avium infection. The question here is whether and how Keap1 has an effect in controlling M.avium infection through the interplaying of different innate immune pathways.
Mycobacterial strain: Mycobacterium avium (M.avium) strain 104, kind gift from David R. Sherman, was used in this project (35). This strain stably expressing firefly luciferase and has fulfilled some criteria of pathogenic mycobacteria. It was originally isolated from an AIDS patient with relatively stable genome and easy for transformation (36). The strain was cultured in Middlebrook 7H9 broth supplemented with glycerol, Tween-80 and ADC.
Human monocyte-derived macrophages (MDMs): The MDMs were isolated by density centrifugation by LymphoprepTM (Axis-Shield PoC AS, Norway) from fresh buffy coats provided by The Blood Bank of St.Olavs Hospital, Trondheim, Norway (37). The use of peripheral blood mononuclear cells (PBMCs) from healthy blood donors is approved by the Regional Committees from Medical and Health Research Ethics at NTNU after informed consent. The final PBMCs solution was grown in cell culture medium RPMI 1640 (Sigma Aldrich, USA) containing 30% corresponding heat-inactivated human serum which was provided by the Blood Bank (St. Olavs Hospital) to get 5x106 cells/ml. The cell culture medium RPMI 1640 had been mixed with 3.4% L-glutamine and 5mM Hepes solution (Life Technologies). The macrophages will be derived frommonocytes within 3-5 days when they are incubated at 37oC and 5% CO2. After differentiation, the MDMs would be stimulated and infected in 10% serum L-glutamine-and-Hepes containing RPMI 1640.
siRNA transfection assay: Only 7-day macrophages with healthy cells and 75% of confluence wereproceed for Keap1 transfection. The lipid transfection was used as the method to deliver siRNA to humanMDMs. The cells were treated with 20nM siRNA twice at 0 and 48 hours for 72 hours. The reaction mixtures and transfection procedure were followed the RNAi transfection protocol of IntrogenTM for Lipofectamine®RNAiMAX reagent (#13778-150). While the Opti-MEM®1 (1x) (#11058-021) was from Gibco®, the AllStarNegative control (#1027281) and Keap1 Gene solution siRNAs (SI00451675, SI03246439, SI04155424,SI04267886) were from QIAGEN.
In vitro infection: Before M.avium infection, the siRNA media was removed from Keap1 knocked-down MDMs and 10% serum RPMI was added for 1-hour incubation at 37 oC and 5% CO2. The multiple of transfection (MOI) was calculated as 10 M.avium cells over 1 knocked-down MDM (MOI 10:1). The infection was taken in within 4 hours for assessing inflammatory response at different time courses.
Western Blot assessment of protein levels: The total proteins were collected in cold lysis buffer containing complete, Mini, EDTA-free Protease Inhibitor cocktail (#11836170001) from Roche AppliedScience (Switzerland) and 0.25U/ml benzonase nuclease (Novagen, USA) for bulky DNA cutting. The induced inflammatory proteins were analyzed by Western blotting. The sodium dodecyl sulfate polyacrylamide gelelectrophoresis (SDS-PAGE) was run with 2 steps: 30 minutes at 100V then 90 minutes at 150V. Transferring protein to nitrocellulose membrance was followed the protocol of InvitrogenTM iBlot dry blotting system with20V in 9 minutes. 5% non-fat milk or 5% bovine serum albumin-fraction V (BSA) in Tris Buffered Saline-with Tween-20 (TBS-T) was used as blocking solution and antibody-diluted solution. The ratios for dilution were1:1000 with primary antibody (Cell Signaling Technology®, Santa Cruz Biotechnology Inc., Abcam and ProteintechTM) and 1:2000 with secondary antibody conjugated with horseradish peroxidase (HRP) (Dako). The blots were then developed with SuperSignal West Femto Chemiluminescent Substrate (Thermo Scientific) and visualized by Kodak Image Estimation 2000R.
Statistical analysis: A statistical method was used to analyse the results of different conditions and treatment overall. As least five independent replications were done for each experiment, and three to fourreplications were chosen. The GraphPad Prism version 6 was used to plot graphic data. The difference betweenconditions was investigated by using two-tailed student’s paired samples t-test, with significant difference level at p-value < 0.05.
Results and Discussions
With initial observation from the measurement of both NF-κB and IRF regulated cytokines, it was shownthat Keap1 does not only regulate the NF-κB response, but also regulates the type I IFN response ininflammation resulting from M.avium infection (38). It was shown through the down regulation of the expression of the IL-6, TNF-α, IL-1β, type I IFN, IFN-β and IP-10 in both ELISA and real-time PCRmeasurement. Several studies have shown the importance of the type I IFN response to the bacterial infection,in addition to its antiviral activity, including Escheriachia coli and M.tuberculosis (38-40).We knocked down the endogenous Keap1 in human MDMs, and measured the knockdown efficiency by comparing the difference between Keap1 siRNA conditions and scramble siRNA control samples, based on the fold induction of protein by infection - the ratio of protein induction in infected samples over uninfected samples after normalization to housekeeping protein GAPDH (Figure 2). We measured the phosphorylated and total protein levels of 8 different proteins involved in PRR signaling pathways, both NF-κB and type I IFN pathways, and including p38-MAPK for further investigation of mechanism by which Keap1 regulates inflammatory signaling in primary human MDMs during M.avium infection.
Fig. 2. Level of Keap1 in siRNA knockdown compared to siNTC sample: one representative western- blotused to quantify Keap1 knockdown efficiency (a), and the protein level of induced Keap1 in human with Keap1-knocked-down sample compared to non-targeted siRNA for negative control (siNTC) (b). The samples were collected from cells with no infection and with M.avium infection for 30 minutes (30min), 60 minutes (60min) and 4 hours (4hour) with MOI 10:1. Results were graphed with mean and SEM values.
Keap1 down regulates the NF‐κB signaling pathway through IKK complex during M.avium infection: With two different NF-κB pathways, the components of the complex, IKKα, IKKβ and IKKγ, arelikely to be regulated by different factors and under particular conditions (41). The IKK complex activation is at the center of the NF-κB signaling pathway. Recent studies have shown that Keap1 down-regulates TNF induced IKKβ phosphorylation in human cancer cell lines through binding and guiding degradation of IKKβ (25, 26). In those experiments, Lee and colleague showed the upregulation of IKKγ in Keap1-knockdown sample, but they did not go deep into that protein. The IKKβ was investigated in detail and they discovered the E(T/S)GE binding site on this IKKβ to Keap1, while IKKγ protein sequence does not have that domain. Same with IKKα, even IKKα and IKKβ have similar genetic structure with 50.1% identity and 67.1% similarityin amino acid (Sequence Manipulation Suite: Pairwise Align Protein), it does not contain E(T/S)GE binding site for Keap1. And thus Kim and others did not make the observation on IKKα (25). We observed the negative regulation of Keap1 in all IKK molecules and showed evidence that IKKα is also under the effect ofKeap1 during M. avium infection in MDMs, especially at 4-hour post infection. This is in turn resulted in the down regulation of downstream NF-κB activation as well. As figure 3 shows, the difference is easily seen inIKK complex and NF-κB in fold induction by infection between negative control and Keap1-knockdown samples, even though the SEM is high. Knocking down Keap1 significantly increases more than 1.6 factor int-IKKα and t-IKKγ at 60 minutes and 4 hours post infection, while the increase of t-IKKβ was significant at 60 minutes after infection. Both t- and p-NF-κB were significantly higher in Keap1 knockdown samples at 4 hours of infection compared to negative control.
Fig. 3. Keap1 down regulated the NF-κB signaling pathway through IKK complex during M.aviumin fection: one representative western-blot used to quantify level of different proteins (a), and statistical analysis between negative control and Keap1-knockdown MDMs (b). Here, Human MDMs were transfected with siKeap1 for Keap1-knockdown sample (siKeap1) and with non-targeted siRNA for negative control (siNTC). Phosphorylated (p-) and total(t-) protein levels were collected from cells with no infection and with M.avium infection for 30 minutes, 60 minutes, and 4 hours (4 hour) post infection with MOI 10:1. Results were analysed by two-tailed student’spaired samples t-test with p-value<0.05 is level of significant difference (*).
Genetic experiments have shown that IKKβ is the predominant IκB kinase, relative to IKKα (42). Both Lee et al. and our studies supported the finding of Ramakrishnan et al. that the phosphorylation of IκB isdependent on IKKβ and IKKγ in a classical pathway (41). And somehow Keap1 might have role in IKKγ induction also. Because IKKα does not have the E(T/S)GE binding site for direct interaction with Keap1,we anticipated that the IKKα was indirectly under the effect of Keap1. It was mentioned early that theactivation of IKKα is characterized as the alternative NF-κB pathway and IKKα is activated through TNF cytokines family, but not TNF-α (41). Moreover, we observed that the increase of IKKα induction was smaller and later than the increase of IKKβ and IKKγ in Keap1-knockdown samples. Thus, the increase of total IKKα in Keap1- knockdown could be the result of the effect of Keap1-knockdown in cytokine production. In Tamura et al.’s review, IKKα was shown that it may have role in type I IFN pathway, especially it may activate IRF7 (43).
Together with the examination of Keap1’s effect on NF-κB signaling pathway, we also did thei nvestigation if Keap1 affects the signaling pathway leading to type I IFN induction. The TBK1 molecule is atthe center of the IFN signaling pathway, which commonly phosphorylates IRF3 and IRF7.
Keap1 negative regulates TBK1 during M.avium infection: The most important finding in our studies is the control of Keap1 to TBK1 in M. avium-infected MDMs. There is no or research that shows the relation or interaction between TBK1 and Keap1. And we found that both the total and phosphorylated TBK1 is increased in Keap1-knockdown MDMs. As showing in figure 4, both phospho- andtotal TBK1 are higher in Keap1-knockdown than negative control samples after infection with M.avium.However, the difference is significant only with p-TBK1 at 60-minute post infection. Meanwhile, with IRF3, we cannot detect any clear difference between two types of samples.
Fig. 4. Keap1 down regulated the TBK1 protein in type I IFN signaling pathway during M.avium infection: one representative western-blot used to quantify level of different proteins (A), and statistical analysis betweennegative control and Keap1-knockdown MDMs (B). Here, Human MDMs were transfected with siKeap1 for Keap1-knockdown sample (siKeap1) and with non-targeted siRNA for negative control (siNTC). Phosphorylated (p-) and total(t-) protein levels were collected from cells with no infection and with M.avium infection for 30 minutes, 60 minutes, and 4 hours post infection with MOI 10:1. Results were analysed by two-tailed student’spaired samples t-test with p-value<0.05 is level of significant difference (*).
In theory, TBK1 is the central protein of type I IFN pathways and phosphorylated TBK1 leads to the activation of IRF3, one of the transcription factor for type I IFN. However, we could not detect any clear difference in IRF3 between Keap1-knockdown and negative control samples. While the cytokine experiments showed the difference in induced IFN, we believed that other IRFs may be involved in MDMs during M.avium infection. Moreover, TBK1 also provides many roles in different layers of immunity. Not only involvedin type I IFN pathways, but TBK1 also have shown that it is a kinase that activates NF-κB pathway and it is a mediator that positively regulates autophagy. Under particular conditions, TBK1 is reported to have a functionas an NF-κB effector by phosphorylating related proteins for NF-κB translocation (44, 45). In addition, many studies implicate the function of TBK1 as mediation of autophagy. Bacterial proliferation is restricted by TBK1through phosphorylating the autophagy receptor Optineurin (OPTN), thus it increases the LC3 binding affinityand antibacterial autophagy (46, 47).
Keap1 has no effect on p38 in MAPK pathway: The MDMs infected with M.avium were also examined regarding the effect of Keap1 on the MAPK pathway. One review from Schorey and Cooper stated the role ofMAPK as the leader for macrophage signaling upon mycobacterial infection (48). As seen in figure 5, the fold induction of p-p38 by infection is not different between negative control and Keap1-knockdown samples. Here,we examined the regulation of p38 as a representative protein in the MAPK pathway that is involved in NF-κBactivation and related genes.
Fig. 5. Keap1 had no effect on p38-MAPK protein during M.avium infection: one representative western blot used to quantify level of different proteins (a), and statistical analysis between negative control and Keap1-knockdown MDMs (b). Here, Human MDMs were transfected with siKeap1 for Keap1-knockdown sample (siKeap1) andwith non-targeted siRNA for negative control (siNTC). Phosphorylated (p-) and total (t-) protein levels were collectedfrom cells with no infection and with M.avium infection for 30 minutes (30min), 60 minutes (60min) and 4 hours(4hour) post infection with MOI 10:1. Results were analysed by two-tailed student’s paired samples t-test with pvalue<0.05 is level of significant difference (*).
Together, our observation and these data indicate that Keap1 regulation might have broad consequences in the innate immune response to mycobacterial infection at initial infection and loop feedback. We observed that the Keap1-knockdown MDMs with M.avium infection have upregulated both IKK complexand TBK1. IKKα is reported to be activated though TNF cytokines family (not TNF-α), so that Keap1 has anindirect negative effect to IKKα through the feedback of inflammatory cytokines. With IKKγ and TBK1, there is no study about the effect or the binding site of these proteins with Keap1. Our results support the evidence that Keap1 has control in both IKKγ and TBK1. However, the mechanism of interaction is still unclear. Both total and phospho-IKKγ are affected by Keap1, theorily; the effect could be the results of the transcriptional process, the protein synthesis, the binding leading to degradation of proteins, the prevention of phosphorylationor the degradation of phosphoprotein. It is the same with TBK1, there is little knowledge about TBK1 and Keap1 interaction. It is needed to investigate the mechanism more in detail.
This work was carried out at the Mycobacteria Research Group, Department of Cancer Research and Molecular Medicine in the Faculty of Medicine, Norwegian University of Science and Technology (NTNU), Trondheim, Norway. Financial support was from the Norwegian Research Council.