MyD88 protein destabilization mitigates NF-κB-dependent protection against macrophage apoptosis

Abstract

Various signaling pathways are essential for both the innate immune response and the maintenance of cell homeostasis, requiring coordinated interactions among them. In this study, a mutation in the caspase-1 recognition site within MyD88 abolished inflammasome-dependent negative regulation, causing phenotypic changes in mice with some similarities to human NEMO-deficiencies. The MyD88^D162E mutation reduced MyD88 protein levels and colon inflammation in DSS-induced colitis mice but did not affect cytokine expression in bone marrow-derived macrophages (BMDMs). However, compared to MyD88^wt counterparts, MyD88^D162E BMDMs had increased oxidative stress and dysfunctional mitochondria, along with reduced prosurvival Bcl-xL and BTK expression, rendering cells more prone to apoptosis, exacerbated by ibrutinib treatment. NF-κB activation by lipopolysaccharide mitigated this sensitive phenotype. These findings underscore the importance of MyD88^wt signaling for NF-κB activation, protecting against macrophage premature apoptosis at resting state. Targeting MyD88 quantity rather than just its signaling could be a promising strategy for MyD88-driven lymphoma treatment.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12964-024-01930-1.

Background

MyD88 (Myeloid differentiation primary response 88) protein is a master regulator of Toll-like receptor (TLR) (except for TLR3) and interleukin-1 receptor (IL-1R) signaling (reviewed in [1, 2]. It consists of the death domain (DD), intermediate (INT), and TIR (Toll/Interleukin-1 receptor) domains. MyD88 has a propensity to self-assemble to MyD88 clusters observed already at resting state [3]. Upon ligand binding to the ectodomain of TLRs or IL-1R MyD88 interacts with receptors through TIR domains [4]. Oligomeric MyD88 recruits kinase IL-1R-associated kinase 4 (IRAK4) and IRAK1/2, which initiates formation of a supramolecular organizing center (SMOC) called myddosome [5, 6] and activating nuclear factor kappa B (NF-κB) or activator protein-1 (AP-1) TF family thus inducing transcription of inflammatory genes. MyD88 signaling not only increases cytokine expression following receptor activation, but also acts as a global immune response regulator by modulating the expression of innate immune receptors such as NLRP3 [7], mitophagy-promoting proteins [8] and anti-apoptotic proteins [9].

Excessive TLR or IL-1R activation may have severe consequences, so regulating MyD88 signaling is pivotal for cell homeostasis and occurs on several levels. Inhibition can be performed directly on MyD88 or other molecules involved in myddosome formation and downstream signaling. The mechanisms can be transcriptional [10, 11], epigenetic [12] or by direct protein inhibition [13]. Lipopolysaccharide (LPS) stimulation increases NF-κB dependent expression of tumor necrosis factor alpha-induced protein 3 (TNFAIP3/A20) or miR-146a, which both negatively regulate NF-κB signaling [11, 14]. LPS also induces alternative splicing, where the MyD88_S isoform lacks INT domain and prevents recruitment of IRAK4 and IRAK4-mediated IRAK1 phosphorylation [13, 15].

Our recent study revealed that human and murine MyD88 possess a caspase-1 recognition site, presenting additional negative regulation through inflammasome-activated caspase-1 cleavage [16]. Herein we generated MyD88^D162E mice with a mutated caspase-1 site to investigate its physiological impact. These mice exhibited previously unseen phenotypic changes in MyD88 mutated or knock out mice. Around one-quarter of MyD88^D162E mice died within eight weeks of birth, though survivors showed no disabilities and were protected from DSS-induced colitis and LPS-induced inflammation. MyD88^D162E protein showed reduced stability compared to wt protein, leading to lower cytokine secretion in the gut, serum, and bone marrow-derived dendritic cells (BMDCs). Interestingly, mutation did not affect cytokine secretion in bone marrow-derived macrophages (BMDMs) and peritoneal macrophages (PEMs). At resting state, MyD88^D162E BMDMs displayed decreased glycolysis rates, increased mitochondrial oxidative stress, reduced expression of prosurvival (non)phosphorylated Brutonˋs tyrosine kinase (BTK) and anti-apoptotic B-cell lymphoma-extra large protein (Bcl-xL), resulting in heightened apoptosis, particularly under apoptosis-inducing drugs for MyD88-driven lymphomas. LPS stimulation alleviated this susceptibility by reducing mtROS, recovering mitochondrial function, and mitigating apoptosis. These findings highlight how MyD88 quantity and thus changed MyD88 signaling profoundly impacts basal NF-κB activity, cell homeostasis, and systemic immune response. The results also show that targeting MyD88 quantity has more drastic consequences for macrophage survival than for immune signaling, suggesting potential implications for treating MyD88-related pathologies.

Methods

Generation of MyD88^D162E (Myd88^em1Bfro) mice

All animal experiments were performed according to the directives of the EU 2010/63 and were approved by the Administration of the Republic of Slovenia for Food Safety, Veterinary Sector and Plant Protection of the Ministry of Agriculture, Forestry and Foods, Republic of Slovenia (Permit Number U34401-6/2017/4).

gRNA targeting the third exon of the mouse Myd88 gene was designed using the CRISPR MIT design tool and Benchling. gRNA target (CCTGTCCTCAGGACAAACGC CGG) with the best score was used for this study. Cas9 and gRNA with targeting sequence CCTGTCCTCAGGACAAACGC were ordered from Synthego (USA). PAGE purified ssODN (IDT) was used as DNA template, coding nucleotide substitution (GAT into GAG for aspartate into glutamate change at the position 162; silent PAM mutation was also introduced) with 90 bp length of homology arms. Endonuclease-mediated mutant mice (Myd88^em1Bfro) were generated using standard microinjection procedures. Briefly, the FVB/NHsd (Envigo, Italy) line was used as embryo donors. On day 1, females were superovulated by PMSG (7.5 U) followed by injection of freshly thawed aliquot of hCG solution (7.5 U) and mated overnight in single pairs with the FVB/NHsd males. On day 4, embryos were collected, and treated with hyaluronidase to remove cumulus cells in the M2 medium. A microinjection mixture of sgRNA, ODN, and Cas9 was prepared according to Synthego protocol and microinjected into the pronuclei of one-cell embryos using Narishige (Japan) micromanipulators and incubated overnight in M16 medium. On day 5, two-cell embryos were transferred in groups of 20 into the oviducts of plug-positive Hsd:ICR foster mothers. D162E single point mutation in mice was confirmed in F1 offspring by Sanger sequencing (Macrogen) of the targeted DNA amplicon, using forward primer TGGGAATAATGGCAGTCCTC and reverse primer GCAATGGACCAGACACAGGT. Standard restriction fragment length polymorphism (RFLP) analysis using HaeIII (NEB) enzyme confirmed the single nucleotide substition. The absence of the off-target site (determined by Benchling) for gene Gm17092 was checked with forward primer GCCCCATCGACCAGGAAGGA and reverse primer GCCAAGGTTGGTCGGGCGCT.

Phenotype characterization was carried out from the birth of the animals until 12 weeks of age of the F3 generation of mice. We tried to include the equal numbers of both sexes of the animals. The survival rate for 56 mice of each genotype was observed for 25 weeks from birth.

Dextran Sodium Sulfate (DSS)-induced acute murine colitis experiment

The role of Myd88^D162E mutant was determined in acute DSS model. B6-MyD88^D162E and corresponding wild type C57BL6/J OlaHsd mice were used. All animal experiments were performed according to the directives of the EU 2010/63 and were approved by the Administration for Food Safety, Veterinary Sector and Plant Protection of the Ministry of Agriculture, Forestry and Food, Republic of Slovenia (Permit Number U34401-21/2022/8). Eight to ten weeks male and female old laboratory animals were housed in IVC cages GM500 (Techniplast), fed standard chow (Mucedola) and tap water was provided ad libitum. The cages were enriched using Nestlets nesting material and mouse houses. Mice were maintained in 12–12 h dark–light cycle at approximately 40–60% relative humidity with 22 °C of ambient temperature. All animals, used in the study, were healthy; accompanied with health certificate from the animal vendor. Health/microbiological status was confirmed by FELASA recommended Mouse Vivum immunocompetent panel (QM Diagnostics). To induce DSS acute colitis model [17, 18] mice were given either tap water or water supplemented with 2% DSS (TdB Consultancy, DB001-31) for 8 days ad libitum, whereas on day 2 and 6 fresh 2% DSS was provided. Number of animals (n = 8), allocated to the experimental groups was determined by GPower 3.1 software. Mice were daily weighed, their stool consistency and presence of gross or occult bleeding was checked. At the day 8, animals were humanely euthanized and then colon length was measured from the body of the caecum to the distal part of the rectum, part of the colon tissue was taken for further analysis. Based on stool consistency and bleeding disease activity index (DAI) for each animal was given, where each score was determined as follows: stool blood, 0–2 (0, no blood; 1, hemoocult bleeding; 2, gross bleeding); and stool consistency, 0–2 (0, normal stool; 1, soft stool; 2, diarrhea). Blood was taken at the end of the experiment to obtain mice sera for cytokine determination. Serum was collected by centrifuging mice blood in Multivette-600 tubes (Sarstedt) at 2000 rpm for 20 min at 4 °C. Blinding of the animal study was conducted as researchers were not aware of the treatment given to each experimental group of the animals.

LPS-induced inflammation

B6-MyD88^D162E and corresponding wild type C57BL6/J OlaHsd mice were used. Mice were injected intraperitoneally (i.p.) with LPS (0,1 mg/kg of body weight) E.coli O55:B5 (Sigma-Aldrich). Four hours after the LPS injection, blood was taken and sera prepared. Mouse IL-6 and mouse IL-10 cytokines were determined using ELISA (Invitrogen) according to the manufacturers recommendation.

Preparation of BMDMs and BMDCs

For BMDMs and BMDCs generation, bone marrows of femur, tibia and humerus were collected from mice. For differentiation to macrophages (BMDMs), bone marrow cells were plated at a density of 1 × 10⁶ cells/ml in RPMI 1640 media containing 20% FBS, 1% penicillin–streptomycin (all Invitrogen), and macrophage colony-stimulating factor (M-CSF) (eBioscience; 40 ng/ml). On day 3, culture media was changed for 40 ng/ml M-CSF for additional 2 days. Attached cells were tripsinized and used for further experiments. For differentiation to dendritic cells (BMDCs), bone marrow cells were plated at a density of 2 × 10⁶ cells/ml in RPMI 1640 media containing 10% FBS, 1% penicillin–streptomycin (all Invitrogen) and incubated with granulocyte–macrophage colony-stimulating factor (GM-CSF) (RD systems; 20 ng/ml). On day 3 culture media was changed for 20 ng/ml GM-CSF. On day 5 half volume of fresh media was added. On day 8 floating and loosely attached cells were collected and used for further experiments.

Isolation of peritoneal macrophages

To isolate PEMs, mice were given 1 ml of 3% thioglycolate broth (Sigma) intraperitoneally 4 days before collecting them with peritoneal lavage with 6 ml of ice cold PBS with 27G needle. The cells were grown in RPMI 1640 media, supplemented with 10% of FBS (Invitrogen).

qPCR

BMDMs were seeded at 0.5 × 10⁶ cells/24-well, PEMs at 1 × 10⁶/24-well and left untreated or stimulated with 10 ng/ml LPS (from Salmonella abortus equi HL83; a gift from K. Brandenburg, Forschungszentrum Borstel) or 10 ng/ml Pam2CSK4 (InvivoGen) for 4 h in RPMI 1640 media + 10% FBS. Total RNA was isolated using High pure RNA isolation kit (Roche). Colon tissue was homogenized in TRIZOL (Roche) and RNA was isolated. cDNA was prepared using high capacity cDNA reverse transcription kit (Applied Biosystems) and qPCRs for Gapdh as internal control, Il6, Il10, Il1b, MyD88, A20/Tnfaip3, p62, Nox4, Hmox1, and Nqo1 were performed using SYBR green I master kit (Roche) on LightCycler 480 (Roche). The data are presented as fold increase of mRNA in treated cells relative to untreated cells using ΔCt values.

Dual luciferase test

HEK293 MyD88 KO cells, prepared in our lab [19], were growing in DMEM containing 10% FBS (both Invitrogen) and 6 × 10⁴ cells/96-well were seeded. To determine the role of caspase-1 in MyD88 signaling regulation cells were transfected with pcDNA3.1 mouse MyD88^wt (a gift from R. Medzhitov (Addgene plasmid #13,092; http://n2t.net/addgene:13092; RRID:Addgene_13092)) or pcDNA3.1 mouse MyD88^D162E plasmids (both 0.05 ng/well) w/o increasing concentrations of pCMV mouse Caspase-1 (0.05, 0.1 ng/well) as well as pELAM1-luciferase (NF-κB promotor; 30 ng/well)) (a gift from C. Kirschning, Germany) and phRL-TK Renilla luciferase (3 ng/well) (Promega) for normalization using lipofectamine 2000 transfection reagent (Invitrogen). 6 h after transfection, media was changed and cells were stimulated with 10 ng/ml of recombinant human IL-1β (Invivogen) for 16 h and afterwards lysed. To determine the role of MyD88^D162E in downstream signaling HEK293 cells or HEK293 MyD88-KO cells were transfected with pcDNA3.1 mouse MyD88^wt or pcDNA3.1 mouse MyD88^D162E plasmids (both 5 ng/well) as well as luciferase under NF-κB (pGL4.32 NF-kBluc2P; Promega) (3 ng/well), AP-1 (pGL3 3xAP-1luc; Addgene plasmid #40,342) [20] (5 ng/well) promoters, or promoters for CCL20 (pGL3 human CCL20luc) (10 ng/well) and IFNβ (pGL3 IFNβluc; a gift from J. Hiscott, Canada) cytokines, and Renilla luciferase (3 ng/well) for normalization. Cells were lysed 24 h after transfection and analyzed for luciferase activity using dual luciferase assay on Orion luminometer (Berthold).

ELISA

BMDMs were seeded in 96-well plates at 2 × 10⁵/well, BMDCs at 2 × 10⁵/well, PEMs at 3 × 10⁵/well and left untreated or stimulated with 10 ng/ml LPS or 10 ng/ml Pam2CSK4 for 16 h. For NLRP3 activation, BMDMs were seeded at 2 × 10⁵/well and primed with LPS (10 ng/ml) for 2 or 4 h. Media was changed and further stimulated with 3% DSS for 24 h or 5 μM nigericin for 1 h.. For tolerance assay, BMDMs were seeded at 2 × 10⁵/well and primed with LPS (100 ng/ml) for 24 h. Media was changed and cells rested for 2 h and then restimulated with LPS (10 ng/ml) for 6 h. Supernatants were collected and analyzed for IL-6, IL-10, TNFα, IL-1β or IL-18 (Invitrogen) according to manufacturers’ instructions. Mice sera and colon tissue from DSS and LPS experiments were also tested for IL-6 and IL-10 cytokine levels.

Western blot

HEK293T cells (seeded at 7.0 × 10⁵/6-well) were transiently transfected with MyD88^wt or MyD88^D162E w/o mCaspase-1 plasmids or w/o pcDNA3 hCaspase-3 (Addgene #11,813) [21]. After 24 or 48 h cells were lysed and WB was performed.

BMDMs (seeded at 1.3 × 10⁶/24-well) were stimulated for 6 h with 100 ng/ml LPS for NLRP3, ASC, Casp-1, p62, and Bcl-xL detection.

For inhibition of protein translation and proteasomal degradation of MyD88 cells were treated with either cycloheximide (100 μM; Sigma) or lactacystin (50 μM; AdooQ Bioscience) for 5 h. Cells were lysed in RIPA buffer containing protease inhibitors (1:100; Sigma); for pBTK and BTK detection cells were lysed in RIPA buffer containing protease inhibitors (1:100; Sigma) and phosphatase inhibitors (1:50; Millipore).

20–40 μg of proteins were used for detection; SDS-PAGE and western blot (WB) were performed. Anti-NLRP3 Ab (Adipogen, 1:2000), anti-mCasp-1 p20 (Adipogen, 1:2000), anti-mCasp-1 p10/p12 (Abcam, 1:1000) anti-ASC (Adipogen, 1:1000), anti-p62 (Progen, 1:1000), anti-mMyD88 (Sigma, 1:000), anti-pBTK (Cell Signaling, 1:1000), anti-BTK (Cell Signaling, 1:1000;), anti-Casp-3 (Cell Signaling, 1:1000), and anti-Bcl-xL (Santa Cruz, 1:1000) were used. Secondary Ab with HRP against mouse (Jackson Immuno Research, 1:3000), rabbit (Abcam, 1:4000), and guinea pig (Progen, 1:25,000) were used. Anti-β-actin Ab (Cell Signaling, 1:5000) and anti-Gapdh (ProteinTech, 1:20,000) were used as a loading control.

BMDMs were seeded in 24-well plates at 2 × 10⁶/well. After 1 h nuclei were isolated using nuclear/cytosol fractionation kit (Abcam). 40 μg of nuclear proteins were used for detection; SDS-PAGE and WB were performed. Anti-p65 (NF-κB) Ab (Santa Cruz, 1:500) and anti-PCNA (Abcam, 1:1000) as loading control were used. As secondary Ab anti mouse-HRP Ab (Jackson Immuno Research, 1:3000) were used.

Transmission electron microscopy

BMDMs were fixed with 2.5% (v/v) glutaraldehyde in 0.1 M cacodylate buffer (pH 7.2–7.4) for 1 h at room temperature. The fixation was followed by rinsing in 0.33 M sucrose in 0.1 M cacodylate buffer three times for 30 min at room temperature. The samples were then post-fixed in 1% (w/v) osmium tetroxide with 0.8% potassium ferrocyanide in 0.1 M cacodylate buffer for 30 min at room temperature, rinsed in distilled water, incubated in 2% uranyl acetate for 30 min at room temperature and then incubated in distilled water for 5 min. Next, the samples were dehydrated in a graded ethanol series (50, 70%) for 3 min each, 90% ethanol for 5 min and 100% ethanol for 5 min, 2 times. Dehydrated samples were embedded in Epon (Serva Electrophoresis, Heidelberg, Germany) by infiltration; samples were first immersed in a mixture of 100% ethanol and Epon (1:1 ratio) for 1 h at room temperature, followed by immersion in Epon for 30 min, repeated 3-times at room temperature. The polymerization of Epon was performed over the next 5 days with gradual temperature increase (35 °C, 45 °C, 60 °C, 70 °C and 80 °C) every 24 h. Finally, the 60 nm ultrathin sections were contrasted with uranyl acetate and lead citrate and examined at the operation voltage 80 kV with a Philips CM100 electron microscope (Philips, Eindhoven, The Netherlands) equipped with the CCD camera (AMT, Danvers, MA, United States).

Flow cytometry

Apoptosis was measured using annexin V apoptosis kit (Invitrogen). BMDMs were seeded at 3.0 × 10⁵/48-well. The next day the cells were collected and washed with PBS. Cells were resuspendend in 100 μl of 1 × binding buffer and incubated for 15 min at room temperature after the addition of 5 μl of eFluor450 conjugated annexin V. Cells were washed, resuspendend in 200 μl of 1 × binding buffer and 5 μl of 7-AAD viability staining solution was added prior measuring.

For mitochondria staining, BMDMs or BMDCc were seeded at 3.0 × 10⁵/48-well. The next day, cells were collected and resuspended in 50 μl prewarmed PBS + 2% FBS. Cells were incubated with 100 nM MitoTracker Green and 30 nM Mitotracker Deep Red (both Invitrogen) at 37° C for 15 min. After incubation 100 μl of prewarmed PBS + 2% FBS was added and after centrifugation resuspended in 200 μl of PBS + 2% FBS prior measuring.

For intracellular ROS detection, BMDMs or BMDCs were seeded at 3.0 × 10⁵/48-well. The next day cells were collected, resuspended in 50 μl of PBS + 2% FBS and further incubated with 5 μM CM-H2DCFDA (Invitrogen) at room temperature for 15 min prior measurement. For mitochondrial ROS detection, cells were incubated with MitoSox Red (at 3.85 μg/ml; Invitrogen) in 200 μl of PBS + 2% FBS at 37° C for 20 min. After incubation 300 μl of PBS + 2% FBS was added and after centrifugation resuspended in 200 μl of PBS + 2% FBS prior measuring.

Cells were analyzed on a spectral flow cytometer with 405-nm 100 mW violet, 488-nm 50 mW blue and 650-nm 80 mW red lasers (Aurora, Cytek). Data were analyzed using FlowJo software (Tree Star).

Analysis of oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) with Seahorse XF analyzer

BMDMs were seeded in Seahorse XF24 cell culture microplates (Agilent) at a cell density of 2.5 × 10⁵ cells/well in RPMI 1640 with 10% FBS. 18 h after seeding, cell culture medium was replaced with assay medium (Seahorse XF DMEM supplemented with 10 mM glucose, 2 mM glutamine and 1 mM pyruvate (all from Agilent)). Cells were incubated for 1 h at 37 °C in humidified air and then transferred into Seahorse XFe24 Analyzer (Agilent). In Seahorse, cells were incubated for additional 2 h and then mito stress test was performed with 2 μM oligomycin (OM), 4 μM carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP) and 0.5 μM rotenone + 0.5 μM antimycin A (R + AA) (all from Merck). OCR and ECAR were normalized to protein content, which was determined with BCA protein assay (Thermo Fisher Scientific) after lysis of cells with 0.1% (w/v) SDS in water.

Apoptosis measurment

BMDMs were seeded in black microplates (Agilent) at cell density of 2 × 10⁵ cells/well in RPMI 1640 with 10% FBS. One hour later, the floating cells were removed, and cells were treated with ibrutinib (4 μM) or left untreated. At the same time, Incucyte caspase-3/-7 green dye (1:1000, Sartorius) was added. Apoptosis was measured around 20 h with a 45-min cycle on CellInsight CX7 LZR High-Content Screening (HCS) Platform (Thermo Fisher) at 20 × magnification. Sixteen fields were scanned in each well every hour over night. Data was analyzed using the open-source software CellProfiler 4.2.4 [22]. Cells were thresholded and segmented and number of caspase-3/-7 positive cells was counted.

Use of large language models

During the preparation of this work, the authors used ChatGPT in order to shorten and improve the language of the abstract and the discussion. After using this service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

Statistical analysis

GraphPad Prism 8 was used for preparing the graphs and performing statistical analysis. Data from independent experiments are represented as mean ± SEM. A paired t-test was used for flow cytometry data. LN (natural log) transformation was applied for qPCR and ELISA data used in statistical tests. Studentʾs unpaired t-test was used. For the analysis of mice experimental data, two-way ANOVA (SIDAKʾs multiple comparison test) or unpaired t-test were used. One way ANOVA with Bonferroniʾs multiple comparisons test was used for the analysis of ECAR and OCR data. *p < 0.05; **p < 0.01; n.s. – not significant.

Results and discussion

Phenotypic characterization of MyD88^D162E mice

As shown previously, human and mouse MyD88 contain the caspase-1 recognition site to negatively regulate MyD88-signaling after NLRP3 inflammasome activation in different cell lines [16]. Mutation of an aspartic acid residue to a glutamic acid (D148E in humans and D162E in mice) prevented MyD88 cleavage by caspase-1 and did not reduce NF-κB signaling in the presence of an increased amount of caspase-1 [16] (Fig. S1A and B). Asp162 residue in mouse MyD88 is at the beginning of the TIR domain and is conserved between human and mouse MyD88 and Mal protein as well. Lin et al. [23] identified this homologous D87A in human Mal protein as a mutation with decreased NF-κB activation capability, but with unchanged TIR MyD88 and TIR TLR4 interaction capability. We compared mouse MyD88^wt- and MyD88^D162E-dependent activation of several response elements and promoter regions without caspase-1 coexpression. In HEK293wt and HEK293 MyD88-KO cell lines no substantial decrease in activation of NF-κB and AP-1 response elements as well as promoters for CCL20 and IFNβ were observed when mouse MyD88^D162E was expressed (Fig. S1C).

To evaluate the physiological role of caspase-1 inhibition on MyD88 signaling, we generated single point mutation Myd88^em1Bfro mice (further referred as MyD88^D162E mice) as described under Method section and Fig. S2. Phenotypic characterization of newly established MyD88^D162E mice was conducted in the third generation of offspring. Mutant mice are significantly lighter compared to their wild-type littermates until the 3rd week of life, until weaning, whereas this distinction in body mass was abolished after sexual maturity as no significant difference was seen nor for male neither for female mice (Fig. S2A). By observing 56 newborn pups of MyD88^wt and MyD88^D162E mice from birth till 25th week of age, we discovered that approximately 25% of mutant mice died until sexual maturity, where 12.5% of mice death was observed in the first week of life (Fig. S2B), excluding other causes of death of the animals. Some mutant pups exhibit head malformation (Fig. S2C) which resulted in animal dying in few weeks. In some cases, we observed alopecia in young mutant mice (Fig. S2D). We are aware that alopecia due to barbering can manifest as a form of stereotypic behavior in C57BL/6 mice [24], but this was not observed in MyD88^wt mice. At the age of 10 weeks, the weight of major internal organs was similar, but the weight of the spleen was increased in MyD88^D162E mice compared to MyD88^wt mice (Fig. S2E). A complete blood count and extensive blood metabolic panel was conducted, where no differences were seen between mutant and wild-type mice (results not shown).

MyD88^D162E mice are protected from DSS-induced acute colitis and LPS-induced inflammation

Due to impaired negative regulation of MyD88 signaling upon NLRP3 inflammasome and subsequent caspase-1 activation and some physiological changes observed in mutant mice, increased disease pathology was expected. Mice were subjected to well established DSS-induced acute colitis model as it involves NLRP3 inflammasome and Casp-1 activation [18, 25] (Fig. S3A). MyD88^D162E mice developed less severe acute colitis than MyD88^wt mice. Disease activity index (DAI), which is based on scoring of blood presence and stool consistency, was lower in mutant mice (Fig. 1A, left). Additionally, the percent of body weight decrease and colon length change were less prominent in MyD88^D162E mice than MyD88^wt mice (Fig. 1A, middle and right, and Fig. S3B). Better disease outcome was confirmed by the detection of lower levels of IL-6 cytokine in colon washout and mice serum. Also, lower mRNA expression of cytokines (Il6 and Il1b) in colon tissue was determined (Fig. 1B-E). Despite similar Myd88 mRNA expression, lower protein levels of MyD88^D162E were detected in the colon tissue of control and DSS-treated mice than in MyD88^wt mice (Fig. 1F and G), contributing to the observed lower cytokine expression and secretion in the colon tissue and serum as well. To confirm this low inflammatory phenotype in MyD88^D162E mice, LPS was injected i.p. for 4 h. IL-6 as well as IL-10 cytokine levels in serum were lower than in MyD88^wt mice (Fig. 1H).

Fig. 1.

MyD88^D162E mutant mice are protected from DSS-induced acute colitis. Schematic presentation of the DSS-induced colitis experiment. To induce acute colitis, mice were drinking tap water, supplemented with 2% DSS for 8 consecutive days ad libitum. A Disease activity index was defined by observing the stool consistency and the presence of the gross or occult blood. The weight loss was calculated from the weight difference on day1 and day 8 and the colon length was measured from the body of the caecum to the distal part of the rectum. B The part of the colon was washed and IL-6 cytokine was measured using ELISA assay. C Blood was collected and IL-6 was detected in serum using ELISA. D-F Il6, Il1b, and MyD88 mRNA expression was detected using qPCR in colon tissue homogenate. G MyD88 protein detection from obtained colon tissue of treated mice, using Western blot. β-actin was used as loading control. H Schematic presentation of the experiment. LPS was i.p. injected. After 4 h IL-6 and IL-10 cytokines were analyzed in the serum. Combined data from one (B, C. H) or two (A, D-F) indep. exp. are shown as mean ± SEM. LN transformation was applied for the data used in statistical tests (D-F). Studentʾs unpaired t-test was used. p values of < 0.05 (*), < 0.01 (**) are indicated

MyD88^D162E BMDMs produce similar cytokine levels despite lower MyD88 protein levels

To further characterize the protective immune response phenotype, BMDMs were used. After the differentiation of bone marrow cells to macrophages, around one-third less of MyD88^D162E BMDMs were obtained from the mutant mice compared to wt mice. Similarly, as observed in the colon tissue, the amount of MyD88^D162E protein, but not mRNA, was lower than in the MyD88^wt counterparts (Fig. 2A and B). Both proteins degraded similarly during 5 h incubation time with cycloheximide, but it seems that the degradation route was not only proteasomal as lactacystin did not fully prevent it. Although the amount of MyD88^D162E was lower, it did not influence the expression of cytokines. mRNA levels of Il10 and Il6 were similar between mutant and wt BMDMs after 4 h of LPS and Pam2CSK4 stimulation (Fig. 2C), presumably due to all-or-nothing cell response [26]. In addition, cytokine levels of IL-6 and IL-10 were comparable after 16 h stimulation (Fig. 2D). Moreover, LPS tolerance was induced with similar efficiency in both cell types (Fig. 2E). Similar results on cytokine expression levels were obtained from the PEMs (Fig. S4A). As dendritic cells also use MyD88-dependent activation of TFs IRF5 and IRF7 [27, 28], we stimulated BMDCs with LPS and imiquimod, a TLR7 agonist. MyD88^D162E BMDCs produced less IL-6 cytokines than wt BMDCs (Fig. S4B), probably contributing to protective properties in DSS disease pathology.

Fig. 2.

MyD88 protein levels are lower in MyD88^D162E BMDMs, but cells express similar cytokine levels. A MyD88^D162E and MyD88^wt BMDMs were seeded and left untreated for 4 h for MyD88 mRNA detection using qPCR. B BMDMs were seeded and incubated with either cycloheximide (100 μM) or lactacystin (50 μM) for 5 h to inhibit protein translation and proteasomal degradation. Cells were lysed and MyD88 protein was detected using WB. β-actin was used as loading control. BMDMs were stimulated with either LPS or Pam2CSK4 (both 10 ng/ml) or left untreated for 4 h for Il6 and Il10 mRNA detection by qPCR (C) and 16 h for IL-6 and IL-10 detection using ELISA (D). E Cell tolerance was induced by stimulating BMDMs with LPS (100 ng/ml) for 24 h. Media was changed and after 2 h of resting, cells were restimulated with LPS (10 ng/ml) for 6 h. TNFα was detected using ELISA. A representative experiment of two indep. exp. is shown (B). Combined means from four (C) and three (A, D, E) indep. exp. are shown as mean ± SEM. LN transformation was applied for the data used in statistical tests (A). Studentʾs unpaired t-test was used. n.s.- not significant

MyD88^D162E BMDMs experience mitochondrial oxidative stress and have lower glycolysis rates with no profound consequences on protective mechanisms against oxidative stress

Metabolic changes mediated by mitochondrial reprogramming with oxidative stress as a contributing factor, can modulate the immune response [29]. We therefore measured the cytosolic and mitochondrial ROS (mtROS) levels in the cells. We observed higher MyD88^D162E BMDM cell numbers experiencing mitochondrial oxidative stress but not cytosolic ROS (Fig. 3A and B). MitoTracker Deep Red stains polarized mitochondria with the negative charge, whereas MitoTracker Green binds to mitochondrial proteins resulting in charge-independent stain. Dual labelling revealed significant quantitative difference in MyD88^D162E BMDMs (Fig. 3C) and with increased mtROS displaying certain level of mitochondrial dysfunctionality. Next, we performed TEM analysis to investigate the structural changes in mitochondria. In MyD88^D162E BMDMs we observed some disorganized mitochondrial structural network with inhomogeneous cristae distribution (Fig. 3D). In BMDCs no changes in mtROS or dysfunctional mitochondria levels were observed (Fig. S4B).

Fig. 3.

MyD88^D162E BMDMs experience oxidative stress and have lower glycolysis rates. MyD88^D162E and MyD88^wt BMDMs were seeded. The next day, the cells were stained with CM-H2DCFDA for intracellular ROS detection (A), MitoSox Red for mtROS (B), MitoTracker Green and MitoTracker Deep Red for dysfunctional mitochondria detection (C) and analyzed using flow cytometry. D TEM analysis showing mitochondria of wt BMDMs with well-developed cristae, while MyD88^D162E BMDMs have mitochondria with less cristae (arrows) or without cristae (stars). Bars: 500 nm. E BMDMs were seeded and the next day mito-stress test was performed on Seahorse analyzer with oligomycin (2 μM), FCCP (4 μM) and rotenone (0.5 μM) + antimycin A (0.5 μM). OCR and ECAR were normalized to protein content. F MyD88^D162E and MyD88^wt BMDMs were seeded and primed with LPS (10 ng/ml) or left untreated. Medium was changed and cells were further stimulated with 5 μM nigericin (after 4 h of LPS) for 1 h or 3% DSS (after 2 h of LPS) for 24 h. Secretion of IL-1β and IL-18 was measured using ELISA. BMDMs were stimulated with LPS and Pam2CSK4 (both 10 ng/ml) or left untreated for 4 h and next Il1b mRNA was detected using qPCR. G BMDMs were stimulated with LPS (100 ng/ml) or left untreated for 6 h and expression of NLRP3, proCasp-1, ASC and p62 protein was detected using WB. Gapdh and β-actin were used as loading controls. H, I BMDMs were stimulated with LPS and Pam2CSK4 (both 10 ng/ml) or left untreated for 4 h and A20/Tnfaip3 and p62 mRNA were detected using qPCR. J p62 mRNA expression was detected using qPCR in colon tissue from mice DSS-colitis experiment. Colon tissue was homogenized and lysed, next the p62 protein was detected using WB. β-actin was used as loading control again. Combined means from three (C, F, H) and four (A, B, E, I) indep. exp. are shown as mean ± SEM. A representative experiment of two independant exp. is shown (G). Combined data from two (J) indep. exp. are shown as mean ± SEM. p values of < 0.05 (*) are indicated; n.s.- not significant. Paired t-test was used for cytometry data. LN transformation was applied for the data used in statistical tests (F, H). Studentʾs unpaired t-test was used. p values of < 0.05 (*), are indicated. n.s.- not significant

Cellular stress affects cell metabolism. Basal oxygen consumption rate (OCR), which reflects the rate of mitochondrial respiration, and basal extracellular acidification rate (ECAR) that reflects the glycolysis rate were comparable between wt and mutant cells (Fig. 3E). The addition of mitochondria uncoupling agent, FCCP, or after adding mitochondria respiration blockers rotenone and antimycin A, measured glycolysis was lower in mutant cells, showing changes in cell metabolism. Interestingly, no substantial changes in oxygen consumption were observed despite the mitochondrial malfunctioning.

To get an insight into the mechanism of ROS production and their influence on mutant cells, the expression of several proteins was analyzed. NADPH oxidase 4 (NOX4) is a constitutively active H₂O₂ generating enzyme. We observed increased Nox4 mRNA expression in MyD88^D162E BMDMs under steady state conditions (Fig. S4C), stating potential contribution to the observed increased cellular ROS. NOX4 upregulation can also be protective when linked to the Nrf2 antioxidant signaling pathway [30, 31]. However, no substantial changes were observed in expression of Nrf2-dependent Nqo1 or Hmox1 mRNA in BMDMs (Fig. S4C) or colon tissue (Fig. S4D), confirming the involvement of NOX4 in ROS production, but do not indicate the contribution of Nrf2 signaling pathway in the protection of MyD88^D162E mice.

MyD88^D162E BMDMs have preserved NLRP3 inflammasome activity

Increased mtROS and glycolysis can influence the immune response, such as inflammasome activation [32, 33]. We primed BMDMs with LPS and subsequently activated NLRP3 inflammasome with a common, very potent inducer nigericin or 3% DSS. IL-1β cytokine release was similar after 1 h stimulation with nigericin, but significantly lower in mutant BMDMs after 24 h stimulation using DSS, although mRNA expression of Il1b was comparable (Fig. 3F). On the other hand, NLRP3-dependent IL-18 release, which expression is MyD88-independent, was not significantly affected by the MyD88 D162E mutation after DSS stimulation (Fig. 3F). Besides IL-1β expression, NLRP3 expression is also MyD88-dependent [7]. After LPS stimulation, NLRP3 expression was increased similarly in both cell types, whereas no differences in ASC and caspase-1 expression was observed (Fig. 3G) excluding direct NLRP3 inflammasome involvement. Pro-IL-1β processing is also a regulated process and its inhibition can be achieved by restricting ubiquitination of pro-IL-1β by (de)ubiquinating enzyme A20/Tnfaip3 [34, 35]. Expression of A20/Tnfaip3 was indeed increased in MyD88^D162E mutant cells after stimulation compared to MyD88^wt cells (Fig. 3H), showing a potential mechanism of mitigated IL-1β secretion after prolonged DSS stimulation.

Myddosomes and damaged mitochondria are removed from the cell by the autophagic clearance. TLR, as well as NF-κB signaling, induce autophagy/mitophagy [8, 26, 36]. NF-κB mediates LPS-induced expression of an adapter protein p62/SQSTM1required for the removal of damaged mitochondria, maintenance of mitochondrial respiration [8] and myddosome clearance [26]. No changes were observed between the wt and mutant BMDMs on the mRNA and protein level, even after LPS stimulation (Fig. 3G and I). Similar observations were made also on colon tissue from the DSS colitis model (Fig. 3J).

MyD88^D162E BMDMs are committed to higher apoptosis rates with mitigation of preterm apoptosis by LPS stimulation

Observing a quantitative difference after BMDMs differentiation from bone marrow led us to investigate the level of apoptosis as a potential underlying cause. Sixteen hours after seeding the cells under steady state conditions, similar levels of apoptotic cells (ApoCells) were detected on the flow cytometer in MyD88^D162E cells, whereas higher apoptotic bodies (ApoBD) were determined (Fig. 4A). Anti-apoptotic Bcl-xL prevents Bcl-2 associated X-protein (Bax)-dependent formation of pores in the mitochondrial outer membrane and cytochrome c release [37]. Lymphoma cells expressing constitutively active MyD88^L265P express more Bcl-xL and are better protected from apoptosis than wt cells [9]. Therefore, we compared Bcl-xL expression which was lower in mutant BMDMs (Fig. 4B), indicating the cause of mitochondrial dysfunction and increased apoptosis. Moreover, also the amounts of BTK and phosphorylated BTK, which also increase MyD88-driven lymphoma cell survival, were similarly decreased (Fig. 4C).

Fig. 4.

MyD88^D162E BMDMs are committed to higher apoptosis rates, but LPS stimulation mitigates preterm apoptosis. A MyD88^D162E and MyD88^wt BMDMs were seeded. After 16 h the cells were stained with annexin V-eFluor450 to determine the apoptosis by flow cytometry. B, C BMDMs were stimulated with LPS (100 ng/ml) or left untreated for 6 h and expression of Bcl-xL and BTK/pBTK was detected using WB. β-actin was used as loading control. BMDMs were seeded and after one hour treated with ibrutinib (4 μM) or left untreated (D) or stimulated with LPS (100 ng/ml) or left untreated (F). At the same time, caspase-3/-7 green dye was added. Caspase-3/-7 activity was measured for 20 h using a 45-min cycle. E BMDMs were seeded and after one hour, nuclear fractions were isolated and p65 (NF-κB) was detected using WB. PCNA was used as loading control. BMDMs were seeded and stimulated with LPS (100 ng/ml) or left untreated. After 16 h the cells were stained with annexin V-eFluor450 for apoptosis (G) or MitoSox Red for mtROS, MitoTracker Green and MitoTracker Deep Red for dysfunctional mitochondria detection, and CM-H2DCFDA for cytosolic ROS detection analyzed via flow cytometry (H). Combined means from three (A, G, H) indep. exp. are shown as mean ± SEM. Representative data from two (B, C, E) or three (D, F) indep. exp. are shown. A paired t-test was used for cytometry data. p values of < 0.05 (*) or < 0.01 (**) are indicated

To further evaluate apoptosis susceptibility, BMDMs were seeded and left untreated or treated with ibrutinib, an inhibitor of BTK, also used for treatment of MyD88-dependent lymphomas. The activity of caspase-3/-7, which ultimately trigger apoptosis, was measured for 20 h and was more extensive at the beginning of measurement in MyD88^D162E BMDMs compared to wt cells and was further promoted by ibrutinib treatment. Interestingly, approximately after 18 h the apoptosis reached similar levels in both cell types, in accordance with flow cytometry results, however the kinetic of apoptosis in mutant cells was faster (Fig. 4D).

Increased Casp-3 activity in mutant BMDMs raises question whether Casp-3 could cleave mMyD88 as already shown for hMyD88 [38]. HEK293T cells transfected with either mMyD88^wt or mMyD88^D162E and Casp-3 were incubated with raptinal (a Casp-3 activator) (Fig. S5A) resulting in mMyD88^wt but not mMyD88^D162E cleavage (Fig. S5B), identifying the same cleavage site as for Casp-1, but excluding higher Casp-3 activity as the cause of lower amounts of MyD88^D162E.

Portion of immortalized BMDMs forms active myddosomes and signaling pathways at steady state [26]. Also in monocyte-differentiated macrophages consitutively activated NF-κB has been detected and is essential for preserving macrophage viability and mitochondrial homeostasis [39, 40]. Suppressing the constitutive activation of NF-κB already induces release of cytochrome c from the mitochondria and a stimuli-independent apoptosis [40, 41]. We detected less NF-κB in nuclei of MyD88^D162E BMDMs at resting state (Fig. 4E), so we decided to test whether LPS stimulated MyD88^D162E BMDMs could escape preterm apoptosis induction. More than 20 h-LPS stimulation increased apoptosis rates in both BMDMs (Fig. 4F), however shorter stimulation time using LPS resulted in lower apoptotic bodies detection (Fig. 4G). Furthermore, LPS stimulation also significantly reduced mtROS and mitigated mitochondrial dysfunction, but not changing cytosolic ROS levels (Fig. 4H). The effect was more profound in MyD88^D162E BMDMs than the wt BMDMs, confirming that MyD88-dependent NF-κB activation at resting state is important for prevention of preterm apoptosis, probably through increased Bcl-xL expression (Fig. 4B).

Altogether, our findings underscore MyD88's pivotal role in immune response, mitochondrial function, and cell death regulation. In vitro, the D162E mutation prevented NLRP3 inflammasome-dependent (i.e. caspase-1-dependent) downregulation of MyD88 signaling, but our results show this role is not relevant for the in vivo effects we observed not even the protection in DSS-induced colitis. Reduced MyD88^D162E protein levels in colon tissue and immune cells suggest protein instability. Low inflammation markers detected in serum and colon tissue in DSS colitis-administered MyD88^D162E mice suggest additional factors beyond TLR/MyD88 signaling. Reduced numbers of MyD88^D162E immune cells prone to apoptosis contribute to gut protection, evidenced by milder colitis in mutant mice.

Several single nucleotide polymorphisms (SNPs) and MyD88 mutations that lead to a decreased amount of functional MyD88 due to loss of full-length protein expression and myddosome formation have been identified [42]. For instance, children with germline loss-of-function MyD88 due to mutations in MyD88 or lacking MyD88 expression, typically survive into adulthood only with strict, ongoing antibiotic therapy as they are highly susceptible to various infections [43]. Human MyD88 deficiencies share similar immunological phenotype and clinical features to IRAK4 deficiencies (reviewed in [44]). Furthermore, NF-κB deficiencies due to the mutations in NF-κB essential modulator (NEMO) have been described. They may cause incontinentia pigmenti (IP), an uncommon X-linked dominant genodermatosis, primarily affecting tissues of ectodermal origin such as the skin, teeth, eyes, and central nervous system [45]. Interestingly, one of the persistent markers of IP is alopecia, which was observed in MyD88^D162E mice as well. NEMO-KO cells also have increased apoptosis [46, 47], which has not been observed in MyD88- or IRAK4-KO cells.

One of the MyD88 functions is increased cell survival and protection against apoptosis. For example, gain-of-function mutation Leu265Pro in MYD88 gene, which is present in more than 90% of patients with Waldenström’s macroglobulinemia patients, a type of lymphoplasmacytoid cell malignoma, increases expression of Bcl-xL, enhances cell survival and protects the cells upon bortezomib induced apoptosis [9]. Additionally, CpG-stimulation of activated CD4⁺ T cells also enhances their survival mediated by NF-κB-dependent Bcl-xL expression in a MyD88-dependent manner [48]. A fine balance of binding affinities and expression levels of pro- and anti-apoptotic proteins regulates sensitivity or resistance to apoptosis [49]. In MyD88^D162E BMDMs, we observed decreased Bcl-xL expression, which can disrupt the balance that must be well orchestrated at all time. Bcl-xL performs its anti-apoptotic activity by preventing mitochondrial outer membrane pore formation induced by oligomerization of pro-apoptotic Bax and subsequent release of cytochrome c [50]. Bcl-xL also regulates mitochondrial dynamics (fusion and fission) and increases mitochondrial biomass [51]. The observed Bcl-xL decrease in MyD88^D162E BMDMs suggests a cause for increased mtROS-positive cells and mitochondrial dysfunction. Alongside with decreased expression of another prosurvival protein BTK and its phosphorylated form it rendered cells more prone to apoptosis at resting state and especially with the BTK inhibitor ibrutinib. LPS stimulation alleviated this sensitivity, indicating the importance of MyD88-dependent signaling and NF-κB activation in cell protection against preterm apoptosis. This apoptosis sensitivity is specific to immune cells, as colon length and DIA index remained comparable between control and D162E mice. However, in NEMO-KO mice leading to complete NF-κB deficiency Nenci et al. detected increased apoptosis of colon epithelial cells as well [47].

To conclude, lower MyD88^D162E levels disrupt basal NF-κB activity, thus increasing macrophage susceptibility to apoptosis, likely via Bcl-xL and BTK expression regulation. While TLR-dependent signaling and cytokine secretion are less affected by MyD88 protein levels, our findings support exploiting MyD88 stability disruption-mediated apoptosis as a therapeutic approach for MyD88-driven malignancies, potentially through proteolysis-targeting chimeras strategy (PROTACs) [52] combined with existing treatments.

Authors’ contributions

DL and MMK together with RJ designed the study, and both performed/contributed to most of the experiments and analyzed the data. SH prepared MyD88^D162E mice. DL performed mice experiments, MMK performed in vitro experiments. KD in SP performed Seahorse experiments, FI apoptosis experiments on CX7 and RR TEM experiments. MMK and DL wrote the manuscript. All authors discussed the results and commented on the manuscript.

Funding

This study was financially supported by the Slovenian Research and Innovation Agency (research core no. P4-0176 and projects J7-4640, J4-4563). R.R. was supported by research core no. P3-0108; S.P. and K.D. by research core no. P3-0043 and project J7-3153 and Centre for the Technologies of Gene and Cell Therapy, which has received funding from the European Union’s Horizon research and innovation Teaming program Grant agreement ID: 101059842.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.