Principal Findings: Here we show that ER stress is activated in macrophages during TB infection. Our results show that ER stress markers such as CHOP, phosphorylated-Ire1ï¡ and eIF2ï¡, and ATF3 are expressed in macrophage-rich areas of granulomas in mice infected with virulent Mycobacterium tuberculosis (Mtb). These areas were also positive for numerous apoptotic cells as assayed by TUNEL. Microarray analysis of human caseous granulomas isolated by laser capture microdissection reveal that 85% of genes that play a role in the Unfolded Protein Response ranked in the top 33% in relative transcript abundance. The expression of two ER stress markers, ATF3 and CHOP, were also increased in macrophages of human TB granulomas when assayed by immunohistochemistry. CHOP has been causally associated with ER stress-induced macrophage apoptosis. We found that apoptosis was more abundant in granuloma tissue as compared to non-granuloma tissue isolated from patients with pulmonary TB, and apoptosis correlated with CHOP expression in areas surrounding the centralized areas of caseation.
Conclusions: In summary, ER stress is induced in macrophages of TB granulomas in areas where apoptotic cells accumulate in mice and humans. Although macrophage apoptosis is generally thought to be beneficial in initially protecting the host from Mtb infection, death of infected macrophages in advanced granulomas might favor dissemination of the bacteria. Therefore future work is needed to determine if ER-stress is causative for apoptosis and plays a role in the host response to infection.
Introduction:
Induction of the endoplasmic reticulum (ER) stress pathway known as the unfolded protein response (UPR) is activated in a number of disease processes such as neurodegenerative diseases, obesity, diabetes, atherosclerosis, heart disease, cancer, and viral infection [1-3]. The UPR is an adaptive survival pathway that becomes activated by through accumulation of misfolded proteins in response to a variety of cellular insults that result in ER stress [4]. However, if ER stress is prolonged, or is combined with additional insults, this pathway can lead to apoptosis. ER stress-induced apoptosis is highly dependent on the upregulation of the UPR-inducible transcription factor CHOP (C/EBP homologous protein also known as GADD153). CHOP-deficiency has been shown to protect mice from cholestasis-induced liver fibrosis by reducing hepatocyte apoptosis [5]. Mice deficient in CHOP are markedly protected from ï¢-cell apoptosis and diabetes [6]. Recently, CHOP deficiency was also found to protect from ER stress-induced apoptosis in cultured macrophages in vitro, and macrophage apoptosis was markedly reduced in advanced atherosclerotic lesions of Chop-deficient mice [7-9] .
Both atherosclerotic lesions and tuberculous granulomas accumulate apoptotic macrophages and macrophage foam cells [10-13]. In atherosclerosis, the foamy appearance is caused by the accumulation of cholesterol ester lipid droplets [14]. Mycobacteria-derived lipids such as oxygenated mycolic acids from virulent M. tuberculosis (Mtb) or Mycobacterium bovis Bacillus Calmette-Guerin (BCG) induce a foam cell phenotype in vitro, and cholesterol ester accumulation occurs in Mtb-infected mouse lungs [15-18]. Importantly, recent work has shown that the primary lipids that accumulate in the caseum of a TB granuloma are unesterified cholesterol, cholesterol ester, and triacylglycerol supporting a concept that mycobacterial infection also causes the accumulation of host-derived (Kim et al., submitted manuscript)[19]. Host-derived lipids such as 7-ketocholesterol [20] , saturated fatty acids [21], and unesterified cholesterol accumulation from the uptake of modified, aggregated, and remnant lipoproteins are known inducers of ER stress and contribute to macrophage apoptosis in vitro [7,8,22] (Seimon et al., submitted manuscript). Because ER stress has been causally associated with macrophage apoptosis in advanced atherosclerosis, and macrophage apoptosis is also associated with TB, we tested the hypothesis that ER stress may also be induced in macrophages of granulomas during Mtb infection.
Results:
ER stress is induced in macrophages of granulomas during Mtb infection in mice.
To examine if ER stress is induced during TB infection, we infected five 8 week old female C57BL6 mice by aerosol with Mtb strain H37Rv. Eight weeks post infection, lungs were collected, sectioned, and stained by immunohistochemistry for various ER stress markers. Granulomas were characterized histologically by an abundance of lymphocytes and other cell types surrounding a macrophage-rich core. We found that the ER stress marker CHOP (C/EBP homologous protein also known as GADD153) was expressed in the majority of the granulomas of all five Mtb-infected mice examined. A representative image is shown in Figure 1A. The CHOP-positive staining occurred in the center of the granuloma where macrophages typically reside. Staining was not observed in the lymphocyte-rich area of the granuloma or when we stained a serial section with the normal rabbit IgG control antibody (Figure 1B). To test for macrophage co-localization, we stained a serial sections using the macrophage marker Mac-3 (Figure 1C, left panel). Corresponding boxed areas of Figure 1B (upper panel) and Figure 1C show that the CHOP-positive area was abundant in Mac3-positive macrophages. Interestingly, some areas with the highest Mac-3 staining had relatively low levels of CHOP staining indicating that not all macrophages are expressing CHOP, or expressing very low levels. CHOP is a pro-apoptotic ER stress factor that is necessary for ER stress-induced apoptosis under many types of conditions. Recent data from a number of laboratories suggest that macrophage apoptosis may play a protective role in the host response to infection [13]. We therefore assayed for apoptosis by TUNEL analysis. We co-stained the Mac-3 stained sections by TUNEL to identify apoptotic cells and DAPI to confirm that the TUNEL stain was specific for DNA. We found numerous TUNEL-positive cells that were both Mac3 and DAPI-positive (Figure 1C, middle panel). Moreover, TUNEL positive apoptotic cells (indicated by the red arrows) were abundant in the Mac-3 and CHOP-positive areas of the corresponding boxes (Figure 1C, right panel). .
We also tested for other markers of the ER stress pathway. Sections were stained by immunohistochemistry with antibodies that recognize phosphorylated Ire-1ï¡, phosphorylated eIF2ï¡, or the transcription factor ATF3. All were similarly increased in the granulomas of Mtb-infected mice and colocalized with areas that were positive for the macrophage-specific antibodies AIA31240 and Mac 3 (corresponding dotted boxes, Figure 2A and B). Staining was not observed with an IgG control antibody. Similar to Figure 1, we found many cells that costained with TUNEL, Mac3, and DAPI (red arrows) in areas that were also positive for our ER stress markers ATF3, phospho-Ire-1ï¡, and phospho-eIF2ï¡ (corresponding dotted boxes, Figure 2A and B). These results indicate that several markers associated with the ER stress pathway are induced in macrophages of granulomas from Mtb-infected mice.
ER stress is induced in granulomas during Mtb infection in humans. To determine if ER stress was also induced in human tuberculosis, we analyzed a genome-wide microarray on caseous granulomas from three independent lung tissue samples to explore which genes were highly expressed in response to human tuberculosis (Kim et al., submitted manuscript). Caseous granulomas obtained from lung resection cases were isolated by laser capture microdissection (LCM) and captured RNA was amplified by PCR and hybridized on a GeneChip ® Human X3P array. Because TB granulomas have markedly different cell types than that of non-infected lung parenchymal tissue, an arbitrary hierarchical scale was assigned to rank the genes in relative transcript abundance. A ranking of greater than 10 is considered high for gene expression and represents the approximately the top 33% of genes in relative transcript abundance. We mined the database for genes involved in the ER stress response and compared them with the expression of genes involved in innate immunity (TLRs and scavenger receptors) and resident ER proteins that are not regulated by ER stress. As shown in Figure 3, the classic ER stress-induced proteins including protein folding chaperones such as calreticulin (CALR), calnexin (CANX), heat shock proteins, and transcription factors such as GADD153 (CHOP), XBP1, ATF3, and ATF4, ranked among the most abundant transcripts in this database. Approximately 85% of the ER stress-regulated genes had a relative transcript abundance ranking over 10 (indicated in red). In contrast, the types A and B scavenger receptors SRA (MSR) and CD36, the majority of the toll-like receptors (TLR's), and many of the ER resident proteins that are not induced by ER stress were not as highly expressed (indicated in green). These results suggest that the UPR is induced in human tuberculosis at the mRNA transcript level.
Induction of the ER stress markers CHOP and ATF3 occur in macrophages of human granulomas. To determine if the induction of the ER stress markers could also be detected at the protein level, we performed immunohistochemistry on granulomatous tissue collected from lung resections of 3 patients with TB. In all cases CHOP was highly expressed around the edges of the caseous granuloma, with some staining extending into the centralized area of caseation (Figure 4A). In contrast, no staining was observed using the rabbit IgG control antibody. We also stained serial sections of these lung samples for ATF3 and the macrophage marker CD68. Again, in all cases, the CHOP-positive areas (left panels) were also positive for ATF3 (middle panels) and CD68 (far right panels) (Figure 4B). Lower magnification revealed the entire granuloma and showed a very similar staining pattern between CHOP, ATF3, and CD68 (short red arrow), as well as co-staining in multinucleated giant cells surrounding the granuloma (long red arrows). In addition, there were numerous cells that expressed ATF3 and CHOP that were not CD68 positive indicating that other cell types involved in granuloma formation may also be undergoing ER stress. These results suggest that ER stress is induced in macrophages and potentially other cell types during Mtb infection in humans.
Induction of apoptosis in CHOP-positive areas of human granulomas. We next assayed for apoptosis in the CHOP-positive areas of human granulomas. Lung sections taken from three patients with pulmonary TB were stained for CHOP, DAPI, and TUNEL. Two of the specimens contained granulomas and the third did not contain granulomatous tissue and was used as a negative control. Similar to the data in Figure 4, CHOP staining was most abundant surrounding the centralized areas of caseation in the granuloma-positive tissues (Figure 5A). When the same sections were analyzed for apoptosis, we observed numerous TUNEL-positive cells that were also positive for the DNA stain (DAPI) within the CHOP-positive region surrounding the centralized area of necrosis (Figure 5B). When the number of CHOP-positive cells was quantified and expressed as a percent of total cells that stained with DAPI, we found a significant increase in the percent of CHOP-expressing cells in the granuloma tissue versus the control patient (Figure 5C). A corresponding increase in the percent of TUNEL-positive cells was similarly observed (Figure 5D). Moreover, over 80% of the TUNEL-positive cells were also CHOP-positive in the granulomatous tissue of both patients, whereas less than 20% were CHOP-positive in the negative control (Figure 5E). When we quantified apoptosis specifically in the CHOP-expressing population of cells, we observed a significant increase in apoptosis in the granuloma tissue versus the negative control (Figure 5F). These results suggest that apoptosis correlates with CHOP expression in the granulomatous tissue of human TB patients.
ER stress enhances Mycobacterium bovis bacillis Calmette-Guerin (BCG)-induced apoptosis in cultured macrophages. Although CHOP expression has been shown to be necessary for ER stress-induced apoptosis, it is often not sufficient for inducing apoptosis [7,23]. Rather, one or more additional cell insults are often needed in combination with UPR-CHOP activation to trigger apoptosis, perhaps to ensure that macrophage apoptosis only occurs when cell recovery is unlikely. In this regard, we have recently discovered a multi-hit mechanism of macrophage apoptosis that involves the combination of pattern recognition receptor (PRR) engagement by ligands for the types A and B scavenger receptors (SRA and CD36) and ER stress [24-26]. Mycobacteria have been shown to engage many PRRs including the scavenger receptors SRA, CD36, and MARCO [27-30]. Our data showing that ER stress is correlated with macrophage apoptosis in granulomas during Mtb infection suggest that ER stress may promote macrophage apoptosis during mycobacterial infection. To test this hypothesis, we examined if the addition of an ER stress-inducing agent could enhance macrophage apoptosis induced by Mycobacterium bovis bacillis Calmette-Guerin (BCG). Thapsigargin is an ER stress-inducing agent that, when given at a low dose, does not induce apoptosis by itself (Figure 6). Likewise, we found that infection of peritoneal macrophages with BCG at a low multiplicity of infection (MOI < 10) also resulted in low levels of apoptosis. However, when BCG was combined with thapsigargin, a synergistic increase in apoptosis was observed even when given at very low levels (MOI = 4).
Discussion:
TB is a major contributor to mortality, and a third of the world's population is thought to be infected with Mtb. Mtb gains entry into the host by hijacking macrophages and suppressing macrophage death [31,32]. In spite of the evidence that apoptosis can be suppressed by Mtb, macrophage apoptosis is induced during Mtb infection in vitro, and apoptotic macrophages are found inside granulomas of Mtb-infected lung [33,34].
Based on the previous work in our laboratory on the role of ER stress in macrophage apoptosis in atherosclerosis, we set out to test the hypothesis that ER stress may also be induced in TB; a disease in which macrophage apoptosis has been hypothesized to be beneficial. Our finding that ER stress is induced in macrophages residing in mouse and human granulomas during Mtb infection, notably where apoptotic cells accumulate, reveals a striking commonality to what we and others have observed in advanced atherosclerosis [9,20,35]. Similar to the situation with atherosclerosis, the exact cause of ER stress during TB infection is unknown. Some of these insults may arise from the accumulation of ER stress-inducing host-derived lipids such as 7-ketocholesterol [20], long chain saturated fatty acids [21], and unesterified cholesterol [8], which also contribute to macrophage apoptosis in vitro (Seimon et al., submitted manuscript). Indeed, one of the primary lipids that accumulate in the caseum of a TB granuloma is unesterified cholesterol (Kim et al., submitted manuscript) [36]. In addition to host-derived lipids, other potential causes of ER stress in granulomas that could promote apoptosis are mycobacteria themselves, peroxynitrite, or hypoxia, which are associated with mycobacteria infection [37-39].
Our cell culture data show that the addition of an ER stress-inducing agent synergizes with BCG to trigger apoptosis, whereas neither of the agents induced death on their own when given at low concentrations. We have recently shown that ER stress in combination with additional insults, for example when an ER stressor is combined with atherosclerosis-relevant PRR ligands that recognize either the type A scavenger receptor (SRA) in combination with TLR4, or the type B scavenger receptor (CD36) in combination with TLR2, a synergistic induction of macrophage apoptosis ensues [40]. Previous studies have revealed that mice deficient in CHOP or in the PRR's SRA, CD36, TLR4, or TLR2, are markedly protected from macrophage apoptosis by this two-hit pathway [9,25,40] and (Seimon et al., submitted manuscript). Based on our data presented herein, our findings suggest an intriguing hypothesis whereby ER stress contributes to PRR-induced apoptosis of macrophages taking up mycobacteria within the granuloma. There are many endogenous inducers of ER stress in vivo, e.g., unesterified cholesterol from host-derived lipoproteins, oxysterols, and peroxynitrite that may coexist with various SRA, TLR4, CD36, and TLR2 ligands in granulomas, e.g., mycobacteria and their lipid products and oxidized-phospholipids [28,29,41-43]. The combination of these events could therefore contribute to macrophage apoptosis in granulomas during TB infection.
Both the mechanisms and functional consequences of macrophage apoptosis in TB have been the subject of intense investigation and debate over the last several years. For example, in vitro data have shown that inducing apoptosis in macrophages infected with mycobacteria results in killing of the pathogen [44]. Apoptosis, but not necrosis of infected monocytes is coupled with killing of intracellular BCG[44,45]. Although all mycobacteria are capable of inducing apoptosis to some degree, less virulent strains such as BCG and avirulent Mtb (H37Ra) are better inducers of macrophage apoptosis when directly compared with virulent strains of Mtb (H37Rv) [11,32,46,47]. From these data a hypothesis has emerged that mycobacteria pathogenicity is inversely correlated with the ability of macrophages to undergo apoptosis [13]. This hypothesis is supported by results that suggest two distinct advantages in promoting macrophage apoptosis during early mycobacteria infection. First, the increased bacterial killing that is known to occur in apoptotic macrophages in combination with clearance of the infected corpses may lead to sterilization of the infected host and prevent the spread of infection [12,34,45,48,49]. Second, the uptake of infected apoptotic macrophages by dendritic cells (DCs) leads to the breakdown and processing of Mtb antigen [50], which is then presented by MHC class 1 to T-cells stimulating the adaptive immune response [50]. Two groups also showed that delivery of mutated proapoptotic versions of Mtb, or pro-apoptotic DNA from Mtb, are more effective at eliciting a host response to Mtb challenge [50,51]. In late-stage infection, however, the consequence of cell death likely changes. A late-stage tuberculosis granuloma has little vascularization and a restricted supply of both macrophages and lymphocytes [52]. The induction of cell death in infected macrophages and the foamy macrophages that are abundant in late-stage disease will lead to accumulation of caseum and the progression of pathology. Mtb takes advantage of this pathology, the liquifaction and cavitation of the granuloma, to release infectious bacilli into the airways. Therefore while apoptosis might be beneficial to the host at low bacterial numbers early in infection, it is more likely to be detrimental once the disease has progressed and fibrocaseous granulomas have formed [13,36].
Although we found that ER stress is induced during TB infection, our data linking ER stress to the pathogenesis of TB is correlative. Therefore, future in vivo studies using, for example, CHOP-deficient mice are needed to determine whether ER stress is causative for macrophage apoptosis in vivo and whether this pathway is ultimately beneficial or detrimental in the host response to Mtb infection.
Materials and Methods:
Reagents. Falcon tissue culture plastic was purchased from Fisher Scientific. Tissue culture media, cell culture reagents, and heat-inactivated fetal bovine serum (FBS) were from GIBCO. Thapsigargin (Tg), concanavalin A, DAPI, and Meyers hematoxylin were purchased from Sigma. Antibodies were purchased from the following sources: rabbit polyclonal antibody to CHOP, ATF3, and normal rabbit IgG control-Santa Cruz Biotechnology; Mac-3 antibody and biotinylated anti-rat IgG-BD Biosciences; mouse monoclonal antibody to actin-Chemicon; CD68 antibody-DAKO; antibody against Ser724-Ire1ï¡, Ser51-eIf2ï¡-Abcam; horseradish peroxidase-conjugated donkey anti-rabbit IgG secondary antibodies-Jackson ImmunoResearch; donkey anti-rabbit Alexa-488-Invitrogen/Molecular Probes; Thr183 /Tyr185-JNK, and JNK-Cell Signaling Technology.
Mouse infection. Eight to ten weeks old female C57BL/6J mice were infected with Mtb using an aerosol chamber (Glas-Col Inc.). Animals were exposed to an aerosol produced by nebulizing 5 ml bacterial single cell suspension in PBS at ~2x108 bacilli/ml resulting in 17598 colony forming units per lung as determined by plating homogenized lungs 24h post infection on enriched Middlebrook 7H11 agar plates. Eight weeks post infection, the upper left lung lobe was fixed in 10% formalin and embedded in paraffin.
Human tissue specimens. Human tuberculosis tissues were processed as detailed in Kim et al 2010. All protocols were reviewed and approved by the ethical review boards at Cornell University and University of Cape Town. Informed consent was obtained from all of patients. We have obtained human TB lung tissues which were surgically excised when TB patients had extensive lung cavitation and tissue degeneration, with no response to the antibiotics. Frozen samples were embedded in O.C.T. compound (Tissue-Tek®), cut into 10 µm-thick sections at −25°C on a Cryocut 1800 cryostat (Leica Microsystems) and attached onto PET-membrane frame slides (Leica Microsystems) for laser capture microdissection or Superfrost Plus Micro slides (Erie Scientific) for histologic examination.
Laser Capture Microdissection. Tissue sections mounted onto PET-membrane slides were fixed in graded ethanol (70%, 75%, 96%, and 100%) containing 0.5% sodium azide for 30 sec per each step (Kim et al, 2010). Slides were air dried briefly and areas of interest were dissected by using Leica AS LMD system (Leica Microsystems).
Microarray. Total RNA was isolated from LCM-derived materials using PicoPure™ RNA Isolation Kit (Arcturus) and treated with DNase using TURBO DNA-free™ (Ambion). Purity, quality, and quantity of RNA were assessed by the Agilent NanoChip Bioanalyzer assay, spectrophotometer, and reverse-transcription PCR (RT-PCR) of β-actin (234 bp), glyceraldehyde-3-phosphate dehydrogenase (GAPDH, 189 bp), 18S rRNA (236 bp), and ubiquitin (198 bp). To obtain sufficient quantities of mRNA for microarray, two-rounds of linear amplification of RNA were conducted using RiboAmp HS™ Amplification Kit (Arcturus) and BioArray High Yield™ RNA Transcript Labeling Kit (T7) (Enzo Life Sciences, Inc.). Biotinylated aRNA was fragmented and hybridized onto GeneChip® Human X3P Array (Affymetrix), and the chip was scanned by GeneArray 300 scanner (Affymetrix). We performed three independent arrays on one LCM-derived granuloma sample, which confirmed the reliability of the technique. Two additional granulomas were further analyzed by microarray for biological replicates.
Microarray data analysis. Affymetrix X3P array raw CEL files were analyzed at the probe level using R and Bioconductor (Kim et al 2010). Only PM probes were included in the estimation for expression level of probe sets. Raw PM intensities from each array were subjected to local background correction using MAS5 method, log2-transformed and median-centered. Informative PM probes from each probe set were selected as having the highest mean across all arrays. In this study, the mean signal of top five informative PM probes of each probe set was used to represent its expression level. The relative levels of gene expression within each of three independent caseous granulomas were averaged to yield a single list with standard deviation.
Mouse peritoneal macrophages and bacteria. Peritoneal macrophages from adult female C57BL/6J mice and all mutant mice used in this study were harvested three days after i.p. injection of concanavalin A [7]. All macrophages were grown in full medium containing DMEM (25mM glucose), 10% FBS, and 20% L-cell conditioned medium solution with no antibiotic (GIBCO). The medium was replaced every 24 h until cells reached 90% confluency. One day before the experiment, the cells were dissociated and counted. Macrophages were plated on 48 well tissue culture plates at a density of 75,000 cells per well in the same medium without the antibiotic. Mycobacterium bovis bacillis Calmette-Guerin (BCG Pasteur strain, American type Culture Collection) were grown to early log phase in Middlebrook 7H9 medium with ADN enrichment and 40mM Sodium Pyruvate, and 0.05% Tween to prevent clumping. The bacteria were pelleted and resuspended in PBS and 0.05% Tween, vortexed, and prepared as a single cell suspension by a series of low speed centrifugations (800rpm for 10 minutes). Macrophages were infected with BCG at increasing multiplicity of infection based on optical density (0.1OD = 5 X 107 bacterial/ml). The input bacteria was verified by plating in triplicate on 7H9 agar supplemented with OADC and colonies were counted two weeks later to obtain the MOI.
Antibody staining of mouse and human sections. Lungs were immersion-fixed in 10% neutral-buffered formalin overnight followed by embedding in paraffin. 7-µM sections on glass slides were deparaffinized in xylene and hydrated in water. Antigen retrieval was obtained by exposing the slides for 20 min to 1 mM EDTA, pH 8.0, in a steamer. The sections were incubated with the appropriate antibody overnight, and stained for immunofluorescence of immunohistochemistry. For immunohistochemistry, the sections were then treated with 3% H2O2 to inactivate the endogenous peroxidase. The sections were stained using the rabbit ABC staining system from Santa Cruz Biotechnology according to the manufacturer's protocol. The sections were then counterstained with Mayers Hematoxylin and examined by light microscopy.
For immunofluorescence assays, the sections were permeabilized with 0.1% Triton X-100 and then incubated with anti-Mac-3 antibody overnight, followed by incubation with a biotinylated anti-rat IgG, and a streptavidin-conjugated Alexa 488-labeled secondary antibody. Sections were then stained for TUNEL as described below, followed by counter staining with DAPI to identify nuclei.
Apoptosis assays. For the in vitro experiments, apoptosis was assayed in cultured macrophages by staining with Alexa 488-conjugated Annexin V (green) and propidium iodide (PI; red), as described previously [7,23]. The number of Annexin V- and PI-positive cells were counted and expressed as a percent of the total number of cells in at least 4 separate fields from duplicate wells. Representative fields (4-6 fields containing ~1000 cells) were photographed. Apoptotic cells in tissues were labeled by TUNEL (TdT-mediated dUTP nick-end labeling) using the in situ cell death detection kit TMR-red (Roche Diagnostics) according to the manufacturer's protocol. Only TUNEL-positive cells that co-localized with DAPI-stained nuclei and Mac3 were counted as being positive. TUNEL and annexin staining were viewed using an Olympus IX-70 inverted fluorescent microscope equipped with an Olympus DP71 CCD camera. For TUNEL analysis, DAPI, TUNEL, and/or Mac3 images were merged using Photoshop analysis software (Adobe Systems). The number of cells positive for TUNEL, DAPI, and CHOP (human sections only) were counted from four fields obtained from each section.
Statistics. Values are given as means  S.E.M. unless noted otherwise in the figure legend (n is noted in the figure legends); absent error bars in the bar graphs signify S.E.M. values smaller than the graphic symbols. Comparison of mean values between groups was evaluated by an ANOVA followed by a post-hoc Student-Newman-Keuls. For all statistical methods a P value less than 0.05 were considered significant.
Figure Legends:
Figure 1. Apoptotic macrophages in CHOP-positive regions of Mtb-infected mouse lung. Mice were infected by aerosol with Mtb (strain H37Rv). Eight weeks later lungs were fixed and embedded in paraffin. Granulomas are identified in the center of each image and have rings of lymphocytes (dark blue) surrounding a central macrophage-rich core. A and B. Immunohistochemistry was performed using an antibody against CHOP and a normal IgG control. Slides were counter-stained with hematoxylin. A. Low power image of CHOP staining (brown) in several granulomas present (4X magnification, red arrows). B . IgG and CHOP staining (10X magnification, bar represents 1mm). C. A serial section was stained for macrophages using an anti-Mac-3 antibody (green), and for apoptotic cells by TUNEL (red). Nuclei were stained using DAPI. Left and right panels (20X and 40X magnification respectively) are merged images of Mac3, TUNEL, and DAPI stain. Black and white dashed boxes indicate corresponding areas of the granuloma of adjacent sections. Red arrows indicate TUNEL positive cells that were also positive for Mac3 and DAPI. Far right panel is the same area in a serial section, also shown in A and B, stained for CHOP by immunohistochemistry (Magnification 40X).
Figure 2. UPR markers are induced in macrophages residing in granulomas during Mtb infection in mice. Mice were infected as described in Figure 1. A and B. Immunohistochemistry was performed on lung sections using an antibody against Ser724 Ire1ï¡ (P- Ire1ï¡), Ser51 eIF2ï¡ (P- eIF2ï¡), ATF3, the macrophage (Mï¦) marker AIA31240, and normal rabbit IgG. Slides were counter-stained with hematoxylin. B. A serial section was stained for macrophages using an anti-Mac-3 antibody (green), and for apoptotic cells by TUNEL (red). Nuclei were stained using DAPI. Dashed boxes in A and B indicate similar areas of the granuloma in adjacent sections that were enlarged in the corresponding images for clarity. Red arrows indicate similar areas of the lesion from serial sections that are positive for the various ER stress markers, macrophages, and apoptotic cells.
Figure 3. ER stress-induced genes are upregulated in human TB granulomas. RNA isolated from caseous granulomas by laser capture microdissection from 3 TB patients was subjected to microarray analysis. All genes in the database were ranked and a hierarchical list of each gene in relative transcript abundance was created. Shown is a comparision of ER resident genes that that have not been shown to be regulated by ER stress, common innate immune receptors such as scavenger receptors, TLR's, and macrophage markers, and genes known to participate in the Unfolded Protein Response or ER stress pathway. Error bars represent the standard deviation from 3 independent caseous granulomas compared in the microarray. Genes represented by a red bars had a ranking above 10 and were represented in the top third of all genes in relative transcript abundance, and the genes represented by green bars fell below that threshold.
Figure 4. Induction of ER stress markers in human caseous TB granulomas. A and B. Lung sections from TB patients containing granulomas were stained by immunohistochemistry using an antibody against CHOP, normal IgG control, ATF3, and the macrophage marker CD68. Slides were counter-stained with hematoxylin. A. CHOP staining (brown), but not the IgG control, is seen around the central area of caseation. B. Low and high power magnification shows CHOP staining in areas that are also positive for ATF3 and CD68 (thick red arrow). Also shown are multinucleated giant cells surrounding the granuloma also positive for CHOP, ATF3, and CD68 (thin red arrows).
Figure 5. Increased apoptosis in CHOP-positive regions of human TB granulomas. Lung sections from TB patients containing granulomas were stained for CHOP (green), TUNEL (red), and DAPI to stain the nuclei (blue). Patients 1 and 2 were positive for granulomas and the negative control was from a patient with no observable lung granulomas. A. CHOP staining (green) is seen around the central area of caseation. B. High power magnification shows TUNEL staining in areas that are also positive for CHOP and the DNA stain DAPI (red arrows). C. Quantification of the percent of CHOP-positive cells from 4 random fields of view. The number of CHOP-positive cells were expressed as a percent of total DAPI-positive cells in each field (n=4 fields/patient). D. Quantification of the percent of TUNEL-positive cells from 4 random fields of view. The number of cells positive for both TUNEL and DAPI were expressed as a percent of total DAPI-positive cells in each field (n=4 fields/patient). E. Quantification of the percent of total TUNEL-positive cells that express CHOP. The number of cells positive for TUNEL, DAPI, and CHOP were expressed as a percent of total number of TUNEL-positive cells in each field (n=4 fields/patient). F. Quantification of the percent TUNEL-positivity in the CHOP expressing population. The number of cells positive for TUNEL, DAPI, and CHOP were expressed as a percent of total number of CHOP-positive cells in each field (n=4 fields/patient). Differences between values with symbols and no symbols, or between values with different symbols, are statistically significant by ANOVA followed by post-hoc Student-Newman-Keuls (P < 0.05).
Figure 6. BCG-induced macrophage death is enhanced by thapsigargin. A. Murine peritoneal macrophages were infected with the indicated MOI of BCG in the presence or absence of 0.5 ïM thapsigargin (an ER stressor). After 24 hours the macrophages were analyzed for apoptosis, and the data are expressed as the percent of total cells that stained with annexin V and propidium iodide (mean ± SEM; n=4). Representative fluorescent images are shown. The presence of symbols or differing symbols indicate P < 0.05; identical symbols or the absence of symbols indicate differences that are not significant by ANOVA post-hoc Student-Newman-Keuls.
Acknowledgements. This work was supported by funds from the National Institutes of Health NHLBI 055936 and NIAID 057086 to DGR and RO1AI05338 to GK.
