Ikaros mediates the DNA methylation-independent silencing of MCJ/DNAJC15 gene expression in macrophages

MCJ (DNAJC15) is a mitochondrial protein that regulates the mitochondrial metabolic status of macrophages and their response to inflammatory stimuli. CpG island methylation in cancer cells constitutes the only mechanism identified for the regulation of MCJ gene expression. However, whether DNA methylation or transcriptional regulation mechanisms are involved in the physiological control of this gene expression in non-tumor cells remains unknown. We now demonstrate a mechanism of regulation of MCJ expression that is independent of DNA methylation. IFNγ, a protective cytokine against cardiac inflammation during Lyme borreliosis, represses MCJ transcription in macrophages. The transcriptional regulator, Ikaros, binds to the MCJ promoter in a Casein kinase II-dependent manner, and mediates the repression of MCJ expression. These results identify the MCJ gene as a transcriptional target of IFNγ and provide evidence of the dynamic adaptation of normal tissues to changes in the environment as a way to adapt metabolically to new conditions.

metabolism that acts as a break to attenuate mitochondrial metabolism during adaptation to metabolic stress conditions. MCJ was initially identified as a gene expressed in some but not all ovarian cancer cell lines and primary ovarian cancer tumors 6 . MCJ is expressed in breast and uterine cancer cells that are sensitive to different chemotherapeutic drugs, but not in those that are multidrug resistant 2,6 . In normal human and murine tissues, MCJ is highly expressed in heart, liver and kidney and within the immune system, in CD8 + T cells and macrophages 2,7 . DNA methylation constitutes the only mechanism associated with the regulation of MCJ expression. In ovarian cancer cells, the presence of high levels of CpG island methylation within the first exon of the MCJ gene is associated with loss of expression and correlates with a diminished response to chemotherapy and poor survival 1,6,[8][9][10] . However, the mechanisms that regulate MCJ expression in normal tissues and cells are not known.
We have shown that MCJ modulates macrophage responses to a variety of proinflammatory insults 7 . Short-term induction of inflammation by infection with Staphylococcus aureus or injection with LPS prevented TNF production in vivo and the development of acute fulminant hepatitis in mice in the absence of MCJ 7 . MCJ is therefore, a potential therapeutic target under conditions of persistent inflammation. Here, we report that IFNγ regulates the expression of MCJ in macrophages through a mechanism that involves the transcriptional regulator, Ikaros. These data demonstrate a novel mechanism of MCJ gene expression regulation that is independent of DNA methylation.

Results and Discussion
Loss of MCJ expression in heart-infiltrating macrophages during infection with B. burgdorferi. During short-term in vivo inflammatory conditions, MCJ regulates the response of macrophages to Staphylococcus aureus as well as LPS treatment in mice sensitized with galactosamine 7 . In order to determine the role of MCJ on the local macrophage response during an infectious process that requires a more complex and long lasting interaction between the pathogen and the host, we infected MCJ KO and WT mice with Borrelia burgdorferi. After 3 weeks of infection, macrophage infiltration was not significantly different in infected MCJ KO mice and WT animals (Fig. 1A). In addition, the amount of TNF expressed in the cardiac tissue upon infection was not altered in the absence or presence of MCJ (Fig. 1B). We also assessed the level of expression of MCJ in heart-infiltrating macrophages at the peak of infection with the spirochete. Surprisingly, in contrast to bone marrow-derived macrophages (BMMs), real time RT-PCR failed to detect appreciable levels of MCJ mRNA in macrophages infiltrating the hearts (Fig. 1C). The downregulation of MCJ expression during infection was selective of macrophages since total heart MCJ expression levels were readily detected in the infected mice (Fig. 1C). The histological analysis of infected joint and heart tissue showed that the degree of cardiac inflammation was not affected by the lack of the MCJ gene (Fig. S1A,B). Furthermore, the levels of spirochetal DNA were similar in The base of the hearts of 3 week infected and uninfected (UI) mice were used to extract RNA and assess macrophage infiltration and TNF expression levels by qRT-PCR using primers specific for F4/80 (A) or TNF (B). NS; Not significant. (C) Macrophages were purified from the hearts of 3-week infected B6 mice and used to extract RNA. qRT-PCR was then performed to detect MCJ mRNA levels, compared to bone marrow-derived macrophages (BMM). As a control, MCJ mRNA levels were also determined in whole heart tissue of 3-week infected mice. The data shown correspond to 5 mice per group and are presented as the mean ± SE. *; Student´s t test, p < 0.05.
WT and MCJ KO mice (Fig. S1C). These results suggested that upon infection with B. burgdorferi, MCJ expression is repressed specifically in macrophages infiltrating the heart. MCJ expression in macrophages is selectively downregulated by IFNγ. In order to determine whether the interaction of macrophages with bacterial products results in reduced levels of MCJ, we stimulated RAW cells and BMMs with live B. burgdorferi and assessed the levels of MCJ. Stimulation with the spirochete did not affect MCJ protein ( Fig. 2A) or mRNA (Fig. 2B) levels. LPS stimulation also failed to alter the levels of MCJ in macrophages ( Fig. 2A). These data indicate that the regulation of the expression of MCJ occurs independently of pattern-recognition receptor (PRR) stimulation, including TLR4, TLR1/2 and other PRRs stimulated by the interaction of live B. burgdorferi with macrophages [11][12][13][14][15] . Since IFNγ is a major contributor to macrophage function during cardiac infection with B. burgdorferi 16,17 , we stimulated macrophages with IFNγ . Treatment with IFNγ resulted in lower levels of MCJ protein in both RAW cells and BMMs (Fig. 2C,D). Because MCJ is localized in mitochondria, we examined the effect of IFNγ on mitochondrial mass; however, no difference was observed as determined by levels of the mitochondrial protein, VDAC1 (Fig. 2E). The effect of IFNγ was selective of macrophages, because it did not affect MCJ levels in the murine tumor cell line, Hepa 1-6 or primary CD8 + T cells (Fig. 2C). IL-6 has been shown to downregulate MCJ levels in breast cancer cell lines 2 . Similarly, we found that IL-6 induced the downregulation of MCJ in Hepa liver cancer cells (Fig. 2F). However, IL-6 failed to downregulate MCJ expression in RAW cells or BMMs (Fig. 2F). These results show that MCJ expression in macrophages is selectively silenced by IFNγ .

IFNγ inhibits MCJ gene transcription independently of DNA methylation.
To determine if the downregulation of MCJ protein levels by IFNγ in macrophages was due to an effect on MCJ gene expression, we assessed MCJ mRNA levels in macrophages stimulated with IFNγ . The treatment with IFNγ resulted in a significant decrease in MCJ mRNA levels in RAW cells and BMMs (Fig. 3A). No previous studies have characterized the human or mouse MCJ gene promoter region and addressed transcriptional regulation. We identified a 1 kb region upstream of the start initiation site of the murine MCJ gene (Fig.  S2A), that was capable to mediate high levels of transcription in RAW cells in luciferase reporter assays (Fig. 3B). Treatment with IFNγ caused a pronounced decrease in the transcriptional activity of this region of the MCJ promoter (Fig. 3B).
The only described mechanism of regulation of MCJ involves the methylation of CpG rich regions of the gene 9 . Thus, we addressed whether IFNγ could silence MCJ expression through DNA methylation. BMMs were treated with IFNγ in the presence of the methylation inhibitors, decitabine (DEC) and 5-azacitidine (Aza). Both DEC and Aza failed to prevent the downregulation of MCJ expression by IFNγ (Fig. 3C). We also analyzed by bisulfite sequencing the methylation status of CpG residues present in the gene region identified as distinctively methylated between CD8 + T and B cells 18 and that correlates with the level of expression of MCJ in these cells 5,7 . Six CpG residues were identified in this region (Fig. 3D). Of these, the first two were methylated in 100% of the BMMs samples analyzed (Fig. 3D). Importantly, the stimulation of BMMs with IFNγ did not affect the methylation of these CpG residues (Fig. 3D), indicating that IFNγ effect is independent of DNA methylation mechanisms. We further analyzed the Black circles indicate 100% of the samples contained these residues methylated, while white circles represent 0%. The analysis corresponds to BMMs isolated from 6 mice. (E) CHIP analysis of BMM DNA immunoprecipitated with antibodies against the H3 marks corresponding to trimethylation of Lys 4 (H3K4m3) and 27 (H3K27m3) or H3 pan-acetylation (Pan Ac-H3). The binding leves are relative to total H3. The results correspond to the average ± SE of 3 independent experiments. effect of IFNγ treatment on histone H3 marks associated with the activation and repression of gene expression (ref. 31). The treatment of BMMs with IFNγ did not affect the binding of thrimethylated H3 at Lys 4 and Lys 27 or acetylated H3 to the MCJ promoter (Fig. 3E). These data revealed that transcriptional regulation is an alternative mechanism of MCJ expression modulation that is independent of DNA methylation or alteration on histone marks.

Ikaros is an inducible repressor of MCJ gene transcription.
To identify the specific mechanism by which IFNγ represses MCJ gene transcription, we performed a search for potential transcription factor binding sites within the 1kb region of the mouse MCJ gene promoter using the tool TFSearch 19 . Two consensus binding sites for Ikaros (− 350 to − 361 and − 706 to − 717) were identified (Fig. S2A). Ikaros is known to act primarily as a repressor of gene expression 20 . To demonstrate whether Ikaros binds to these putative binding sites in the MCJ promoter and address whether binding was regulated by IFNγ , we performed chromatin immunoprecipitation (ChIP) assays in BMMs. Binding of Ikaros to both sites was almost undetectable in untreated BMMs (Fig. 4A). However, Ikaros binding to Site 1 (the most proximal to the transcription start site; Fig. S2A) was highly induced in cells treated with IFNγ (Fig. 4A). Ikaros binding to Site 2 (Fig. S2A), however, was not induced by IFNγ (Fig. 4A).
To further analyze the contribution of both putative Ikaros binding sites to the regulation of MCJ gene expression, we generated deletion mutants of both binding sites, as well as a double deletion mutant lacking both binding sites. The deletion of Site 1 resulted in significantly increased transcriptional activation in reporter assays (Fig. 4B) suggesting that this site is bound under basal conditions to a negative gene expression regulator. However, the deletion of Site 2 did not affect the expression activity of the promoter (Fig. 4B). Furthermore, the deletion of both sites resulted in transcriptional activity that was equivalent to that observed with the deletion of Site 1 (Fig. 4B). Overall, these data suggest that Ikaros binds to Site 1 upon induction with IFNγ and displaces a weaker negative regulator of gene expression.
To demonstrate the role of Ikaros in the IFNγ -dependent repression of MCJ gene expression, we generated stable lentiviral transductants containing a short hairpin sequence (shRNA) specific for the IKFZ1 gene encoding Ikaros. Transduction with shIKFZ1 in RAW cells caused a prominent reduction of Ikaros levels (Fig. 4C). Importantly, while IFNγ downregulated MCJ levels in control cells, it did not affect MCJ levels in shIKFZ1-transduced cells (Fig. 4D). These results demonstrate that silencing of MCJ expression by IFNγ is mediated by Ikaros and reveal this repressor as a key factor in the alternative mechanism regulating MCJ expression.
We then assessed whether IFNγ upregulates Ikaros expression. The stimulation with IFNγ did not affect Ikaros levels in macrophages (Fig. 4E), suggesting that the increased binding to the MCJ promoter region was due to post-translational modifications induced by IFNγ . Ikaros activity is regulated by phosphorylation mediated by Casein Kinase 2 (CK2) 21 . It has also been reported that IFNγ regulates the expression of a subset of genes through the activation of CK2 21,22 . To investigate whether IFNγ promotes Ikaros binding through CK2, cells were treated with IFNγ in the presence of a CK2 specific inhibitor, 4,5,6,7-tetrabromobenzotriazole (TBB) 23 . CHIP analysis demonstrated that the pretreatment with TBB abrogated IFNγ -induced binding of Ikaros to the MCJ promoter (Fig. 4F).
We then investigated whether silencing of MCJ expression by IFNγ was mediated by CK2. Pretreatment of macrophages with the CK2 inhibitor prevented downregulation of MCJ expression by IFNγ (Fig. 4G). In addition, inhibition of CK2 also prevented IFNγ from suppressing MCJ promoter transcriptional activity (Fig. 4H). Together, these results show that IFNγ represses MCJ gene transcription in macrophages by promoting CK2-dependent DNA binding of Ikaros to the proximal region of the MCJ promoter.
Our studies identify a novel mechanism of regulation of MCJ gene expression that is independent of the well-established DNA methylation pathway described in several tumors. MCJ/DnaJC15 is emerging as an important regulator of mitochondrial activity and cellular function in vitro and in vivo 5,7 . Therefore, the control of MCJ transcription constitutes a mechanism to regulate cellular responses to environmental changes. As opposed to DNA methylation, which is considered a long-term mechanism to silence gene expression 24 , the transcriptional control of MCJ gene expression by Ikaros may allow normal tissues to adapt dynamically to a changing environment. Here, we demonstrate that Ikaros represses MCJ expression in response to IFNγ in macrophages. Similar mechanisms could be used to alter MCJ levels in other cells or tissues in response to changes in the environment as a way to adapt metabolically to new conditions.
Our results identify MCJ gene expression as a transcriptional target of the cytokine IFNγ , contributing to the regulation of their inflammatory output 7 . Our data also reinforces the role of IFNγ as a cytokine that exerts a protective effect during infection with B. burgdorferi 16,17,25 . Overall, we hypothesize that the combined effect of IFNγ , including the regulation of MCJ expression, results in a more efficient elimination of the bacteria from the infected tissue without a concomitant increase in the inflammatory damage.

Methods
Mice. MCJ-deficient mice in a C57Bl/6 (B6) background 5 and wild type B6 mice were bred at UMass Amherst and CIC bioGUNE. The Institutional Animal Care and Use Committees at UMass Amherst and CIC bioGUNE approved all procedures involving animals.

Infections.
Groups of WT and KO mice were infected by subcutaneous injection with 10 5 Borrelia burgdorferi 297 in the midline of the back. The mice were sacrificed after 3 weeks of infection and analyzed for inflammatory symptoms in joints and hearts stained with hematoxilin and eosin. Signs of arthritis and carditis were determined blindly as described 26 . The number of spirochetes in heart tissue was determined by real-time PCR, using primers specific for the recA gene (Table S1) standardized to μg of total DNA with primers corresponding to Glyceraldehyde 3-Phosphate Dehydrogenase, GAPDH, (Table S1) 17 .

Cells.
Infiltrating cardiac macrophages were isolated from 3-week infected B6 mice. Hearts were perfused with cold Hank´s balanced salt solution (HBSS, Lonza, Anaheim, CA) and cut into small pieces, followed by digestion with 1 mg/mL of collagenase/dispase (Roche) and homogenization in a Dounce homogenizer. The digest was passaged through a 16" gauge syringe to obtain single cell suspensions. The cellular suspension was layered on top of a 3 mL layer of Ficoll (GE Healthcare, Piscataway, NJ) and centrifuged at 400 × g for 40 min without brakes. Monocytes were then purified from the interphasic cellular fraction using a one-step discontinuous Percoll gradient (46%) under isosmotic conditions 27 . Monocytes were used for RNA extraction.
Bone marrow-derived macrophages were generated as described 17 using 30 ng/mL of M-CSF (Miltenyi Biotec, Bergisch Gladbach, GE). Macrophages were allowed to differentiate in 100 mm × 15 mm petri dishes (Fisher Scientific, Pittsburgh, PA) for 8 days. Non-adherent cells were then eliminated and adherent macrophages were scraped, counted and resuspended in serum-free RPMI medium 2 h prior to use.
CD8 + T cells were purified by positive selection from the spleens of B6 mice using biotinylated anti-CD8 (BD Biosciences, San Diego, CA), anti-biotin microbeads and the MACS system (Miltenyi Biotec, Auburn, CA).
Lentiviral particles containing shRNA targeting Ikaros (Ikzf1 gene, Sigma Chemical Co, St. Louis, MO) were produced as described 28 . Supernatants containing the virus were used to infect RAW 264.7 cells, followed by incubation with puromycin at 2 μ g/mL to generate stable lines. Cells containing the empty vector, pLK0.1, were used as a control.
In vitro stimulation. Cells were incubated with 100 ng/mL of murine IFNγ or human IL-6 for the Real-time RT-PCR. RNA from isolated cells or cardiac tissue was extracted by the thioisocyanate method (Amresco, Solon, OH), treated with DNase I (Qiagen), and reverse transcribed using the SuperScript VILO cDNA synthesis kit (Life Technologies). Real-time PCR was then performed using SYBR Green PCR Master Mix (Life Technologies) on a BioRad CFX96 Real-Time System (Bio-Rad, Hercules, CA). Fold induction of the genes was calculated relative to actin, using the 2 −ΔΔCt method. The primers used are listed in Table S1.
Western blot. Five to 20 μ g of protein were run on SDS-PAGE, transferred to nitrocellulose membranes and tested with antibodies specific for MCJ 5 , VDAC1 (D-16) and Ikaros (M-20, Santa Cruz Biotechnology, Dallas, TX). Equal loading was determined using antibodies against GAPDH (6C5) or actin (I-19) from Santa Cruz Biotechnology. Epifluorescence (Apotome) microscopy. Cells were grown in 8-well chamber slides (Nunc Thermo Scientific, Waltham, MA). Upon incubation with 100 ng/mL of IFNγ (eBioScience, San Diego, CA) for 3 days, the cells were processed as described 29 using anti-MCJ Abs, followed by an anti-rabbit IgG conjugated to Alexa Fluor 594.
Cloning of the proximal 1 kb MCJ promoter and luciferase assays. The proximal 1 kb promoter of the murine MCJ gene was cloned into pGL3 using the primers in Table S1. Deletion mutants corresponding to the putative Ikaros binding sites of the MCJ promoter (Fig. S2) were generated using the QuickChange Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA) and the primers listed in Table S1. 1.9 μ g of these constructs plus 0.1 μ g of pSVL40 plasmid were cotransfected into RAW cells using the X-TremeGene HP DNA tranfection reagent (Roche). After 6h, the cells were treated with IFNγ in the presence or absence of the specific inhibitor, TBB. After 20 hr incubation, the cells were lysed in lysis buffer (Promega, Madison, WI) and Firefly and Renilla luciferase activities were determined by the Dual Luciferase reporter system (Promega).
Bisulfite sequencing. DNA was extracted from BMMs treated with IFNγ and controls, denatured and subjected to bisulphite conversion as described by Clark and colleagues 30 . The resultant product was PCR amplified using the primers in Table S1, corresponding to the region in the MCJ gene described by Meissner and colleagues 18 . Chromatin immunoprecipitation. Fifteen million BMMs were stimulated with 100 ng/mL of IFNγ in the presence or absence of TBB for 16h. CHIP assays were performed using the SimpleChip Enzymatic Chromatin IP kit-Magnetic beads (Cell Signaling, Beverly, MA) following the manufacturer´s instructions using anti-Ikaros, anti-H3K4m3, anti-H3K27m3, anti-pan acetylated H3 antibodies and anti-H3 (Cell Signaling) or normal rabbit IgG as negative control. The immunoprecipitated DNA was subjected to q-PCR using primers encompassing the two putative Ikaros binding sites (Table S1). The results are presented as fold induction over rabbit IgG immunoprecipitates or total H3 relative to input (percent input method), following the formula: