Critical Role of Methylglyoxal and AGE in Mycobacteria-Induced Macrophage Apoptosis and Activation

Apoptosis and activation of macrophages play an important role in the host response to mycobacterial infection involving TNF-α as a critical autocrine mediator. The underlying mechanisms are still ill-defined. Here, we demonstrate elevated levels of methylglyoxal (MG), a small and reactive molecule that is usually a physiological product of various metabolic pathways, and advanced glycation end products (AGE) during mycobacterial infection of macrophages, leading to apoptosis and activation of macrophages. Moreover, we demonstrate abundant AGE in pulmonary lesions of tuberculosis (TB) patients. Global gene expression profiling of MG-treated macrophages revealed a diverse spectrum of functions induced by MG, including apoptosis and immune response. Our results not only provide first evidence for the involvement of MG and AGE in TB, but also form a basis for novel intervention strategies against infectious diseases in which MG and AGE play critical roles.


INTRODUCTION
Tuberculosis (TB) represents a major threat to humankind, causing approximately 2 million deaths each year. In the majority of individuals infected with Mycobacterium tuberculosis (MTB), the bacilli cause long-term asymptomatic infection termed latent TB during which bacilli persist within granulomas in a dormant stage. Latently infected individuals have a ca. 10% risk of progression to clinical disease during their lifetime [1,2].
Within granulomas, macrophages undergo apoptosis, which participates in host defence against TB [3,4]. Correlation between apoptosis and virulence/persistence of mycobacteria remains controversial. Virulent MTB have been described to induce more profound apoptosis in macrophages and dendritic cells in vitro than attenuated Mycobacterium bovis BCG (BCG) [5]. Yet, evidence has also been presented that MTB attenuates host cell apoptosis in infected macrophages via induction of antiapoptotic mechanisms and/or upregulation of antiapoptotic genes [6][7][8]. Thus, although the exact impact of MTB on apoptosis remains unclear, a role of apoptosis in protection against, and pathogenesis of, TB is likely.
Apoptosis can be induced by various physical and chemical insults. Recent studies revealed that methylglyoxal (MG) induces apoptosis in a cell-type-dependent manner [9][10][11]. The cytotoxicity of MG involves apoptotic events. MG induces apoptosis in Jurkat cells by activating the c-Jun N terminal kinase (JNK) signal transduction pathway, which decreases the mitochondrial membrane potential, followed by caspase-3 activation [11]. In cultured HUVEC cells, MG induces tyrosine phosphorylation and aggregation of a number of cellular proteins [12]. In aortic VSMC from hypertensive rats, elevated levels of MG and advanced glycation end products (AGE) are found, in parallel with evidence that MG increases oxidative stress, activates NF-kB, and enhances ICAM-1 expression [13]. MG can modify proteins rapidly and cause cross-linking, probably changing the properties of these proteins and thus triggering intracellular responses. Abordo and Thornalley demonstrated that MG-modified human serum albumin (MG-HSA) induces increased expression of TNFa in human monocytic THP-1 cells [14]. Together, these studies suggest that MG-modified proteins play an intriguing role in the development of distinct disease states.
Results of global gene expression analysis of MTB isolated from active pulmonary TB indicate that the tissue environment of MTB in vivo, particularly in granulomatous lesions, is rich in MG [15]. Since a significant number of macrophages in these lesions undergo apoptosis [3], we hypothesized that MG levels in macrophages determine apoptotic events during mycobacterial infection. To contest this hypothesis, we measured MG levels in macrophages during mycobacterial infection, characterized MGinduced apoptosis, and ascertained the relationship between elevated levels of MG/AGE and macrophage functions during mycobacterial infection. Our findings point to a close interrelation between MG/AGE and apoptosis as well as activation of macrophages during TB.

Mycobacterial Infection of Macrophages Triggers Increased MG Production
Our DNA array results suggest that the environment of tubercle bacilli at sites of infection is rich in aldehydes, especially MG [15].
To verify the assumption of an MG-rich environment during mycobacterial infection of the lung, we infected MH-S cells ex vivo with mycobacteria and measured relative MG levels in these cells. MH-S cells are an established murine alveolar macrophagederived cell line. Alveolar macrophages play a pivotal role at the initial innate immune response against infection of tubercle bacilli via the respiratory tract [1]. Moreover, recent evidence reveals that alveolar macrophages are able to mount appropriate immune response [16]. We observed that MG levels significantly increased in MH-S cells 1 day after mycobacterial infection as determined by HPLC (Figures S1A-C). In three independent experiments, MG levels were 2.0-3.7 times higher in mycobacteria-infected macrophages as compared to noninfected macrophages. This level of augmentation is comparable to that in in vitro-induced hyperglycemia and in the blood of diabetic patients [17,18].
These findings demonstrate for the first time that MG levels are significantly increased during mycobacterial infection of macrophages. In general, MG is a physiological product of various metabolic pathways, including anaerobic glycolysis, oxidative decomposition of polyunsaturated fatty acids as well as fragmentation of triosephosphate or catabolism of ketone bodies and threonine. MG, 3-deoxyglucosone (3-DG), and glyoxal (GO) are formed during each step of the glycation process by degradation of glucose or Schiff's base in early glycation, or from Amadori products such as fructosamine in the intermediate stages of glycation [19,20]. MG levels in cells are regulated by several factors including the glyoxalase system and other components which influence the glycolysis pathway. The glyoxalase enzyme system in eukaryotic cells, which comprises glyoxalase I and II, is primarily responsible for detoxification of MG [21][22][23]. Increased levels of glucose or deprivation of phosphate in the cytoplasm can induce MG production by interfering with the glycolysis pathway [24]. It is thus possible that mycobacterial infection of macrophages triggers shifts in these metabolic pathways leading to increased MG production. We cannot formally exclude that mycobacteria directly contribute to increased MG levels of infected macrophages. For instance, reactive oxygen radicals produced during mycobacterial infection may oxidize polyunsaturated fatty acids, which represent abundant components of the mycobacterial cell wall, leading to their decomposition and MG formation. Moreover, our microarray data from pulmonary TB patients indicate that MTB experiences nutrient deprivation at sites of pulmonary TB [15]. Such conditions may alter MTB metabolic pathways leading to MG formation. Due to its highly reactive and small molecular nature, MG can easily diffuse out of mycobacterial cells into host cells.

MG Induces Apoptosis in MH-S Cells
In an attempt to unravel the function of MG during mycobacterial infection, we assessed whether MG plays a role in apoptosis of infected MH-S cells. Basically two observations support this notion. First, MG induces apoptosis in different cell lines [11]; second, granuloma formation is accompanied by apoptosis of macrophages [3]. Consistent with this, a high proportion of apoptotic cells was found in bronchoalveolar lavage from patients with reactive or disseminated TB [25]. In vitro models have demonstrated that mycobacteria cause apoptosis and apoptosis of host macrophages has been claimed to restrict growth of intracellular mycobacteria [6].
Accordingly, we first investigated the ability of MG to induce apoptosis in MH-S cells. Apoptosis occurred in this cell line, as assessed by staining of chromosomal DNA with propidium iodide (PI) followed by cellular analysis using the FACScalibur flow cytometer ( Figures 1A and S2). The percentage of apoptotic cells peaked at an MG concentration of 0.8 mM. At an MG concentration of 1.6 mM the proportion of apoptotic cells decreased. A possible explanation for this observation is that MG induces profound necrosis rather than apoptosis in MH-S cells at this high concentration, as observed recently in Jurkat cells [11]. Alternatively, at this high concentration MG triggers growth arrest at the G1 phase in MH-S cells as reported previously for other cell types [10]. We observed increased externalization of phosphatidylserine, as an early sign of apoptosis 4 h after MG treatment ( Figure S3) indicating that apoptotic signals were induced prior to this time point.
JNK plays an important role in apoptotic processes [26][27][28]. The induction of apoptosis by MG has been correlated with the activation of c-Jun N-terminal kinase (JNK) [11]. To examine the apoptotic route of MG-treated MH-S cells, total proteins of MGtreated MH-S cells were isolated at different time points after treatment and JNK, caspase-3, and caspase-9 activation were determined. JNK was activated as early as 1 h after MG treatment and this activation was maintained for 2 h. Caspase-3 and caspase-9 activation occurred 4 h after MG treatment ( Figure 1B).
The activation of JNK can occur upstream or downstream of caspase activation. When JNK activation precedes caspase activation, it is usually required for apoptosis. This is particularly relevant when cells are exposed to insults caused by various chemical agents, irradiation, or oxidation [29,30]. Our results suggest that activation of JNK precedes the onset of apoptosis, and initiates an intracellular signaling cascade in preparation for apoptosis induced by MG in MH-S cells. We conclude that JNK activation is required for MG-induced apoptosis of MH-S cells. Our results agree with previous reports [11], extending the apoptosis-inducing potential of this small molecule to alveolar macrophages.

Increased AGE Levels in Mycobacteria-infected Macrophages
Under physiological conditions, most MG is bound to cellular proteins as adducts formed with Lys, Arg, and Cys residues. Although the reaction with Cys is considered reversible, elevated MG concentrations can irreversibly modify Lys and Arg residues through AGE formation. AGE can enhance formation of oxygen radicals with subsequent activation of NF-kB and release of proinflammatory cytokines [31,32]. AGE can stimulate mitogenactivating protein kinase (MAPK), activator protein-1 (AP-1), and induce expression of transforming growth b1 (TGF-b1) and of extracellular matrix (ECM) proteins [33,34]. It is, therefore, conceivable that abundant MG levels present during mycobacterial infection of macrophages lead to AGE formation of distinct proteins, which then trigger changes in global gene expression. In hyperglycemia, increased levels of MG are accompanied by elevated AGE formation [17]. We observed abundant AGE during mycobacterial infection of macrophages (Figure 2A). More importantly, we detected AGE in TB lesions of patients. Figure 2B shows a histological specimen of lung lesion from a TB patient. AGE staining colocalizes with the accumulation of macrophages suggesting that macrophages are one of the main sources of AGE in active TB. Thus, formation and accumulation of AGE occurs in TB. Until recently, increased AGE formation has been mostly associated with diabetes, renal failure, aging, and cancer [35]. Therefore, our observation extends the roles of AGE to chronic infectious diseases such as TB and hence, provides new insights into regional immunity against, and pathogenesis of, a chronic bacterial infection.

Mycobacteria-induced Apoptosis Requires MG and AGE
Reduced glutathione (GSH) serves as a cofactor of the glyoxalase system and catalyzes the conversion of MG to D-lactate. MG binds nonenzymatically to GSH forming hemithioacetal, which in turn serves as a substrate for glyoxalase I (GLO1). Glyoxalase II (GLO2) hydrolyzes the product from the GLO1-catalyzed reaction, S-D-lactoylglutathione to D-lactate. GSH is also a wellknown antioxidant, which neutralizes oxidative radicals, including oxygen radicals. Under various experimental conditions, induction of apoptosis is preceded by a rise in intracellular oxygen radicals [9]. MG can induce generation of reactive oxygen radicals through multiple pathways [36][37][38] and these radicals contribute to apoptotic events induced by MG [39]. We show that preincubation of macrophages with GSH reduced AGE formation after infection with mycobacteria ( Figure 3A). At the same time we observed a significant reduction in apoptosis ( Figure 3B). Similar results were obtained for MG-treated macrophages, which had been preincubated with GSH ( Figure 3C). This finding suggests that mycobacteria-induced apoptosis of macrophages involves elevated AGE levels.

MG Modulates Expression of Apoptosis-and Immune Response-related Genes
In an effort to unravel the mechanism(s) of apoptosis induced by MG, we performed global gene expression profiling of MG-treated macrophages. Promptly after MG treatments (30 min), a cluster of genes involved in apoptosis were induced (Tables S1-6) and many of these were increasingly expressed at later time points. TNFa was induced at all time points analyzed confirming previous reports that MG and AGE can induce TNF-a production [14]. Increased transcription of TNF-a was accompanied by a high level of TNF-a protein in supernatants of macrophages treated with MG ( Figure 4A). In addition, several genes associated with TNF-a, such as TRAF1, TRAF2, and members of the TNF-a receptor superfamily were also induced (Tables S1-6). The expression of some DNA damage-inducible genes, such as DDIT3 and GADD45G was strongly elevated, indicating that MG directly or/and indirectly causes DNA damage as a characteristic feature of apoptosis (Tables S1-6).
Notably, MG does not only induce apoptosis-related genes, but also a large number of genes related to macrophage activation and induction of immunity. Beside TNF-a, a series of genes belonging to the chemokine and cytokine families (ligands and receptors) were induced (Tables S1-6). Among highly induced chemokines were CXCL2 and CXCL10, which are highly abundant in lung granulomas induced by mycobacteria [40]. The high expression of CXL10 was confirmed by real-time RT-PCR (Table S7) and ELISA ( Figure 4B). Furthermore, real-time RT-PCR results showed that pretreatment of macrophages with GSH reduced the induction of the corresponding genes (Table S7). This directly indicates the involvement of MG in induction of these genes.
Lung granulomas represent the hallmark of pulmonary TB and the prime focus for the continuous crosstalk between MTB and immune cells [41]. Induction of immunity involves elaborate interactions between chemokines and cytokines produced by immune cells as well as local tissue cells [40,41]. TNF is a pivotal player in granuloma formation [42,43]. TNF-deficient mice infected with MTB are highly susceptible to disease, and granuloma formation is severely impaired [42]. Therapeutic TNF neutralization in rheumatoid arthritis patients with latent MTB infection is associated with increased risk of reactivation TB [44]. Thus, considering the large number of genes related to chemokines and cytokines, the expression of which is strongly regulated by MG, we suggest that MG and AGE play pivotal roles in immunity to TB with a focus on the granulomatous lesion. This assumption is supported by the fact that .70% of the genes, which were regulated by MG (this study), are similarly regulated in  Our global gene expression profiling of MG-treated macrophages revealed that MG effects the expression of numerous genes, many related to apoptosis, activation, and immunity. Moreover, MG seems to effect genes associated with regulation of cellular growth. The glyoxalase system and MG have been proposed as major regulators of cell division based on their ubiquitous occurrence and the high reactivity of MG with cellular thiols underlying cell division [45,46]. In sum, our experiments revealed numerous roles of the small molecule MG in diverse cellular events. Understanding the mechanism(s) involved in these events can provide the basis for designing novel intervention strategies against infectious diseases in which MG and AGE are critical.

MATERIALS AND METHODS In vitro Culture of Bacteria and Culture Conditions
BCG Copenhagen organisms were grown in Middlebrook 7H9 media supplemented with 10% (v/v) ADC (Difco Laboratories) and 0.05% (v/v) Tween 80 (Sigma) to mid-log phase at 37uC with shaking at 90 rpm in screw-capped bottles.

Culture of MH-S Cells
MH-S cells (ATCC No. CRL-2019) were cultured in RPMI medium supplemented with 10% FCS, 1 mM Na-pyruvate, and 2 mM L-glutamate, 10 mM HEPES, 0.05 mM 2-mercaptoethanol, 100 U/ml penicillin, and 10 mg/ml streptomycin sulphate. The cells were maintained in a humidified atmosphere of 5% CO 2 at 37uC. During mycobacterial infection, culture medium without antibiotics was used. Infection of MH-S cells with BCG Copenhagen was performed as described previously [6,7].

Flow Cytometric Analysis of DNA Fragmentation
MH-S cells were seeded at 2.5610 5 cells/ml into a 24-well tissue culture plate for 1 day (2.5610 5 cells/well). After treatment with 0.8 mM MG (Sigma) for 24 h, cells were harvested, resuspended in 500 ml of propidium iodide buffer (0.1% Triton X-100, 16PBS) containing 50 mg/ml propidium iodide (Sigma), and incubated for 15 min on ice. DNA fragmentation analysis was carried out using a FACScalibur flow cytometer (Becton Dickinson). Apoptosis was scored by the percentage of cells appearing in the area below the G1/G0 peak representing cells containing DNA,2N.

Annexin V Binding
One million MH-S cells were seeded into a 6-well plate for 1 day and then treated with 0.8 mM MG, and harvested after MG treatment for 4, 6, and 24 h. The MH-S cells were harvested, washed twice with 16PBS, resuspended in annexin V binding buffer, and stained with annexin V according to the manufacturer's instructions (Sigma) before flow cytometric analysis.

Determination of MG Levels in MH-S Cells
Measurement of MG levels was performed as described previously [23] with a few modifications. MH-S cells (2610 7 ) were infected with BCG at MOI of 50:1. One day after infection, MH-S cells were harvested, washed twice with 16PBS, and lysed by sonication in 5 ml 16PBS. Perchloric acid (PCA) (Sigma) and ophenylenediamine (o-PD; Sigma) were added to a final concentration of 0.5 M and 5 mM, respectively. The mixture was incubated at 20uC for 24 h. PCA precipitate was removed by centrifugation. The supernatant was passed through a C 18 solid-phase extraction cartridge (Waters Sep-Pak tC18 plus cartridge, Millipore), which had been prepared by flushing with 6-8 ml of acetonitrile and 6-8 ml of 10 mM KH 2 PO 4 (pH 2.5). The sample was eluted from the cartridge with 4 ml acetonitrile (Sigma). The 2-methylquinoxaline (2-MQ) content was analyzed by HPLC. 2-MQ was eluted at a retention time of 7.1-7.3 min. Increase in MG was determined by calculating the ratio of peak areas between MH-S cells with and without infection. The obtained ratio was normalized with the ratio of the corresponding cell numbers.

Cell Lysis and Immunoblot Analysis
Cells were lysed in lysis buffer containing 0.5% Nonidet P-40 (NP-40), 20 mM Tris-HCl (pH 7.6), 0.15 mM NaCl, 3 mM EDTA, 3 mM EGTA, 1 mM DTT, 2 mM sodium vanadate, and 16proteinase inhibitor cocktail (Sigma). Lysates were cleared by centrifugation at 13,000g for 10 min, denatured by boiling in Laemmli buffer for 5 min, separated on 12.5% SDS-PAGE gels, and blotted onto nitrocellulose membrane. Nonspecific binding was blocked by incubating the membrane with 0.1% Tween-20/ PBS containing 5% nonfat dry milk for 1 h at room temperature. Membranes were incubated in the blocking buffer with the primary antibody: caspase-3 (Cell Signaling Technology), caspase-9 (Cell Signaling Technology), or P-JNK (Sigma) overnight at 4uC, followed by incubation with horseradish peroxidases (HRP)conjugated secondary antibody, and the specific immune complexes were detected using the Western blot enhanced chemiluminescence (ECL)-based reagent (Amersham Biosciences).

Detection of AGE in Mycobacteria-infected Macrophages
Monoclonal antibody (mAb) against AGE (RDI, ADYGLYabm, clone 6D12) was labeled with Cy3 or Cy5 using Cy3/Cy5 Ab Labeling Kit (Amersham Biosciences) according to the manufacturer's protocol. Infected macrophages were harvested, fixed with 4% paraformaldehyde, and permeabilized in 16PBS containing 0.1% Triton X-100. The fixed macrophages were stained with the labeled antibody and subjected to FACS analysis.

Immunohistological Staining
Due to extensive lung disease and prevalence of MDR isolates, TB patients in Russia frequently undergo elective thoracic surgery. Destructed parts of their lungs are removed within the anatomical borders of lung segments to decrease mycobacterial load. Lung tissue was obtained at the Department of Thoracic Surgery of the Central Tuberculosis Research Institute (CTRI). At the time point of elective surgical intervention, all TB patients were under multidrug combination therapy of at least 3-4 first-and/or second-line drugs. Immediately after surgery, removed lung tissue was fixed in 4% paraformaldehyde. All procedures involving patients were reviewed and approved by the appropriate ethical review boards in both Moscow and Berlin. Informed consent was obtained from each patient in this study.

Measurement of TNF-a and CXCL10 Secretion
MH-S cells (2610 6 ) were treated with 0.8 mM MG in 4 ml culture medium. The growth medium was removed at different time points after MG treatment and assessed for TNF-a and CXCL10 content using Mouse TNF-alpha and CXCL10 DuoSet ELISA Development Systems, respectively, (R&D Systems) according to manufacturer's instructions. The samples from each time point were measured in triplicate. The error bars represent standard deviation from three measurements. Relative OD values were calculated by dividing each OD value with that of untreated sample (0 h).

Gene Expression Profiling
After treatment with MG for 0, 0.5, 4.0, and 8.0 h, MH-S cells were harvested, pelleted, resuspended in Trizol (Invitrogen), and vortexed for 1 min to lyse the cells. Total RNA was extracted with chloroform and precipitated in isopropanol. The RNA pellet was washed twice with cold 70% ethanol.
The GeneChipH Mouse Genome 430 2.0 oligonucleotide microarray (Affymetrix, Santa Clara, CA, USA) comprises 45,101 probe sets, which analyze the expression level of over 39,000 transcripts and variants from over 34,000 well-characterized mouse genes. Labeling of RNA targets, hybridization, and post-hybridization procedures were performed according to protocols provided by Affymetrix. Following washing and staining, arrays were scanned twice at 1.56 mm resolution using a Gene-ChipH Scanner 3000 (Affymetrix, Santa Clara, CA), controlled by GCOS 1.3 software (Affymetrix). Photoemission was detected by a photomultiplier tube through a 570-nm longpass filter. Computer-generated array images were overlaid with a virtual grid, controlled by GCOS 1.3 software. This step allowed definition of each feature and alignment within known array dimensions. About 40 pixels within each feature were averaged after discarding outliers and pixels close to feature boundaries. Gene expression levels were calculated according to the average hybridization intensities of perfectly matched versus mismatched oligonucleotide probes. Arrays were scaled by GCOS 1.3 software to an average hybridization intensity of 2,500 per gene and analyzed independently. After hybridization and image analysis we excluded probe sets with absent calls in all 12 samples or with nonsignificant hybridization intensities.
The Data Mining Tool 3.1 (Affymetrix) and GeneSpring software package 7.2 (Silicon Genetics, Redwood City, CA) were used to compare the untreated control sample (0 h time point) with MG-treated samples at different time points. The data were normalized to compensate for variability in hybridizations and hybridization artefacts. The normalization consisted of three sections: 1. Data transformation: set measurements,300 to 300. 2. Normalize each chip to 50 th percentile of the measurements taken from that chip. 3. Normalize each gene to the median of the measurements for that gene.
The GeneSpring software package 7.2 was used to identify coexpressed genes within a set of gene expression data typically obtained using the microarray technology. The coexpression information was in turn used to help identify the biological function of the genes. To identify potential biological functions associated with gene clusters, we analyzed the Gene Ontology (GO) annotations of coexpressed genes. Microarray data are deposited in GEO (http://www.ncbi.nih. gov/geo/; accession number: GSE3837).
Quantitative Real-time RT-PCR of MH-S cells MH-S cells were treated with 0.8 mM MG for 4 h with or without pretreatment with GSH. After treatment, MH-S cells were harvested, pelleted, resuspended in Trizol (Invitrogen), and vortexed for 1 min to lyse the cells. Total RNA was extracted with chloroform and precipitated in isopropanol. The RNA pellet was washed once with cold 70% ethanol. Five mg total RNA was reverse-transcribed using Reverse Transcriptase Superscript III (Invitrogen) according to the manufacturer's protocol. Quantitative real-time RT-PCR was performed using SYBR Green kit according to the manufacturer's instructions (ABI Prism 7000, Applied Biosystems) and primer pairs listed in Table S8. Relative expression levels were normalized to b-actin levels.