HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Stewart, J. N. Articles by Rivera-Marrero, C. A. Search for Related Content PubMed PubMed Citation Articles by Stewart, J. N. Articles by Rivera-Marrero, C. A. Agricola Articles by Stewart, J. N. Articles by Rivera-Marrero, C. A.

Increased pathology in lungs of mice after infection with an -crystallin mutant of Mycobacterium tuberculosis: changes in cathepsin proteases and certain cytokines

¹ Department of Medicine, Division of Pulmonary, Allergy and Critical Care Medicine, Emory University School of Medicine, Athens, GA 30602, USA
² Department of Infectious Diseases, College of Veterinary Medicine, University of Georgia, Athens, GA 30602, USA
³ Atlanta VA Medical Center Research Service, Room 12C 106, 1670 Clairmont Rd, Decatur, GA 30033, USA

Correspondence
Carlos A. Rivera-Marrero
crivera{at}emory.edu

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

 
Latency and reactivation are a significant problem that contributes to the incidence, transmission and pathogenesis of tuberculosis. The mechanisms involved in these processes, at the level of both the bacillus and the host, are poorly understood. In Mycobacterium tuberculosis the -crystallin (acr) gene has been linked to latency, because it is highly expressed during hypoxic growth conditions. Deletion of the acr gene in M. tuberculosis H37Rv (acr strain) was previously shown to reduce the intracellular growth of bacilli in macrophages; however, its impact on pathogenesis in vivo was unknown. This study demonstrated that infection of C57BL6 mice with acr results in lung bacillary loads 1-2 log units higher in comparison to parental H37Rv. Haematoxylin/eosin staining of lungs revealed exacerbated pathology characterized by extensive obliteration of alveolar air spaces by granulomatous inflammation. RT-PCR analysis and immunostaining of lungs showed that infection with either H37Rv or acr results in the differential expression of lysosomal cathepsin proteases. A slight increase in the expression of the matrix-degrading acidic-type cathepsins B, D and H was noted in acr-infected mice and was associated with clusters of macrophages within lung granulomas. acr-infected mice also showed high serum levels of TNF-, IFN- and G-CSF, suggesting that Acr may play a role in modulating the host response to infection.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

 
Tuberculosis (TB) is the leading cause of death worldwide due to infectious disease, with an estimated 3 million deaths per year (Bloom & Murray, 1992; Dye et al., 1999). Mycobacterium tuberculosis infection occurs when aerosol-droplet nuclei containing a small number of bacilli are deposited in the alveoli of the lung and subsequently phagocytosed by alveolar macrophages. The bacteria replicate within macrophages, inducing the release of pro-inflammatory cytokines, which can lead to the formation of caseating granulomas, tissue destruction, liquefaction and cavity formation (Dannenberg & Rook, 1994). A major concern during TB pathogenesis is the potential of the bacillus to persist in human tissues for long periods in a clinically latent or dormant state without causing any overt disease symptoms. Live bacilli have been isolated from granulomas or tubercles in the lungs of patients with clinically inactive TB (Opie & Aronson, 1927; Robertson, 1933). It is generally thought that most M. tuberculosis-infected individuals have clinically latent infections, since they give a positive tuberculin skin test, but do not present with clinical symptoms and are not contagious (Flynn & Chan, 2001). It is estimated that in 5–10 % of latently infected persons the infection will reactivate and cause active TB disease (Selwyn et al., 1989). Therefore, understanding latency and reactivation, at the level of both the bacillus and the host, is extremely important in the control of this disease.

Some evidence suggests that latency, or the ability of bacilli to remain dormant, is due to reduced growth and/or metabolism, and can be induced by growth under hypoxic conditions within the host. M. tuberculosis infections are known to occur at the most oxygen-rich sites of the body (e.g. the upper lobes of the lung: Adler & Rose, 1996), while inhibition of bacillary growth in vivo is associated with the formation of hypoxic fibrous granulomas (Dannenberg, 1993). In vitro growth of bacilli under a variety of stress conditions such as by ageing growing cultures, or with limited oxygen, results in decreased metabolic activity and growth (Wayne & Diaz, 1967; Wayne & Hayes, 1996). Based on these observations, investigators have utilized hypoxic culture conditions to generate non-replicating but persistent mycobacteria as an in vitro model of latency (Imboden & Schoolnik, 1998; Wayne & Diaz, 1967; Wayne & Hayes, 1996; Yuan et al., 1998). These studies have facilitated the identification of mycobacterial factors that may confer in vivo growth and persistence advantages upon the pathogen. However, it has been difficult to establish the significance of these factors to latency and reactivation in the host.

In vitro growth of M. tuberculosis under hypoxic conditions results in the upregulation of a 16 kDa -crystallin (Acr) homologue, encoded by the acr gene (hspX, Rv2031). Acr protein is almost undetectable during exponential growth of M. tuberculosis, but is strongly induced in old and stationary-phase cultures (Yuan et al., 1996). It is considered a dominant antigen since antibodies are present in sera from most patients with pulmonary TB examined (Lee et al., 1992; Verbon et al., 1992). Acr belongs to a family of small heat-shock proteins that act as ATP-independent chaperones, and localize to the inner side of the cell membrane (Cunningham & Spreadbury, 1998). In vertebrates, Acr plays an important role in maintaining the transparency of the eye (Groenen et al., 1994; Horwitz, 1992); however, its role in M. tuberculosis has not been defined. Disruption of the acr gene in H37Rv was shown to not affect infectivity or survival in macrophages during early infection, but growth of the mutant was significantly impaired in both mouse bone-marrow-derived macrophages and THP-1 monocytes (Yuan et al., 1998). However, there is little information on the role of Acr in vivo. The present study was designed to determine the infectivity and pathogenicity of the acr mutant in the C57BL6 mouse infection model. We demonstrate that, in comparison to the parental wild-type strain H37Rv, infection of mice with acr results in higher bacillary burdens in the lung, exacerbated lung pathology, elevated expression of pro-inflammatory cytokines, and a slightly increased expression of lysosomal cathepsin proteases. We postulate that Acr in M. tuberculosis bacilli is an important modulator of the host response to infection.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

 
Bacterial cultures and growth conditions.
The M. tuberculosis mutant strain acr : : hpt (denoted as acr) was obtained from Dr Clifton E. Barry, III (Tuberculosis Research Section, NIH, Rockville, MD, USA). acr was generated by insertion of a hygromycin-resistance cassette by allelic exchange in the H37Rv strain that replaced the 1 kb acr gene (Yuan et al., 1998). The parental strain H37Rv (ATCC 27294) was obtained from The American Type Culture Collection (Rockville, MD, USA). Mycobacteria were grown in Middlebrook 7H9 broth supplemented with OADC enrichment (Difco) and containing 1 % (v/v) glycerol. acr was grown in 7H9/OADC broth plus hygromycin (50 µg ml^–1). Cultures were grown at 37 °C with slow shaking to mid-exponential growth phase (7–10 days) and bacterial clumps disrupted by repeated passage through syringes with 21, 25 and 27 gauge needles. Numbers of bacilli in the inoculum were determined by measuring OD₆₀₀ and using a linear regression equation generated from an OD₆₀₀ vs c.f.u. curve previously generated. Bacterial counts were determined by serial dilution of cultures in 7H9 medium, plating in triplicate on 7H11/OADC agar plates, and enumeration of c.f.u. after 3 weeks incubation at 37 °C.

Mouse infections and necropsies.
Eight-week-old female C57BL6 mice (Charles River Labs) free of common pathogens were used for these experiments. Mice (nine per group) were infected by inoculation in the tail vein with 0·2 ml (1x10⁶ bacilli) of a freshly grown suspension of M. tuberculosis H37Rv or the acr strain. Mice were humanely killed at weeks 2, 4 and 6 post-infection (three infected and two normal per time point). Blood was obtained by cardiac puncture and serum was separated. The lungs were removed, rinsed in sterile PBS, and the five lobes (one left, four right) divided as follows: (1) upper left lobe for bacillary load determinations, (2) lower left lobe for histopathology, and (3) all right lobes for RNA isolation and RT-PCR. All mice were kept in microisolator cages in a BSL3 facility and their health status monitored daily. Mice were humanely killed if they showed signs of pain or distress before the end point. All of the protocols were approved by the Emory University and Atlanta VA Institutional Animal Care and Use Committee.

Cytokine analysis.
Cytokine analysis of serum samples was done by three different methods: (1) Bioplex multiplex bead immunoassay (Bio-Rad), (2) BD Cytometric Bead Array (BD Biosciences), and (3) Quantikine ELISA (R&D Systems).

(1) Multiplex bead immunoassays were done in filter-bottom ELISA plates using the Bio-Rad Mouse 18-Plex panel kit and protocols for cytokines: IL-1, IL-1, IL-2, IL-3, IL-4, IL-5, IL-6, IL-10, IL-12(p40), IL-12(p70), IL-17, G-CSF, IFN-, GM-CSF, KC, MIP-1, RANTES and TNF-. Briefly, 50 µl of a mixture of the 18 anti-cytokine-conjugated bead families was added to a 96-well filter plate prewet with Assay Buffer A. Sera from control and M. tuberculosis-infected mice were diluted 1 : 4 in Mouse Diluent. To the filter plate containing the beads was added 50 µl diluted serum or serially diluted cytokine standards, followed by 30 min mixing at room temperature, and then at 4 °C overnight. The cytokine-bound beads were washed twice with Wash Buffer A using a filter manifold, and incubated for 1 h at room temperature with 50 µl biotin-conjugated detection antibodies. Bead–cytokine–antibody complexes were washed twice in Buffer A and incubated for 30 min at room temperature with phycoerythrin-conjugated streptavidin. Complexes were washed twice in Buffer A, resuspended in 125 µl Assay Buffer A, and cytokine levels were measured in a Bioplex instrument using Bioplex Manager Software (Bio-Rad). Assays were performed in duplicate and cytokine concentrations were reported in pg ml^–1.

(2) Determination of cytokines IL-2, IL-4, IL-5, IFN- and TNF- in serum samples was further carried out using the Mouse Th1/Th2 Cytokine Cytometric Bead Array kit and protocols from BD Biosciences. Serum samples (in triplicate) were tested undiluted and diluted at 1 : 10 and 1 : 100 in assay buffer, along with serially diluted cytokine standards (20–5000 pg ml^–1). In glass test tubes, 50 µl of the mixed cytokine capture beads was mixed with 50 µl test sample or standard, then 50 µl PE Detection reagent (phycoerythrin-labelled anti-mouse IgG) was added and incubated for 2 h at room temperature in the dark. After incubation, the beads were washed in 1 ml PBS wash buffer by centrifugation, resuspended in 300 µl wash buffer, and analysed in a BD FACS Caliber Flow cytometer after calibration with BD Calibrite beads. Data acquisition and analysis were done with BD CBA software.

(3) Quantification of G-CSF in mouse serum was done by ELISA using the Quantikine kit and protocol from R&D Systems. Serum samples in triplicate (50 µl, undiluted or diluted 1 : 10 in diluent buffer), and the G-CSF standard (serially diluted 14·1–900 pg ml^–1 in diluent buffer), were added to microtitre wells precoated with anti-G-CSF antibody, containing 50 µl assay diluent, mixed and incubated at room temperature for 2 h. After incubation, the wells were aspirated, washed (5x400 µl wash buffer) and 100 µl anti-mouse G-CSF–horseradish peroxidase conjugate added for 2 h. After aspiration and washing, 100 µl substrate solution was added and incubated for 30 min in the dark; the reactions were stopped by addition of 100 µl stop solution, and A₄₅₀ was measured in a Molecular Devices ThermoMax reader. G-CSF concentrations were determined based on the standard curve generated by a four parameter logistic (4-PL) curve fit.

C.f.u. determinations.
Lung tissue samples (0·05–0·1 g) were homogenized in 1·0 ml sterile PBS until no tissue clumps were visible. After a brief sonication (10 s pulse), serial dilutions (1 : 100, 1 : 500, 1 : 1000) were prepared in PBS, and 100 µl aliquots plated in triplicate on 7H10/OADC plates. The plates were sealed and incubated for 3 weeks at 37 °C. Colonies were counted and c.f.u. ml^–1 per g tissue determined. Two colonies from each plate were tested for acid-fast bacilli by Ziehl–Neelsen staining with TB stain kit ZN (Becton Dickinson).

Histopathology and immunohistochemical analysis.
The dissected lower left lobes were placed in 4 % paraformaldehyde for 2 h, transferred to 10 % buffered formalin and stored at 4 °C. The tissues were dehydrated, paraffin-embedded, and sectioned in 5 µm increments starting at the pleural surface. Sections were stained with haematoxylin/eosin for histopathological examination. Several sections were stained for acid-fast bacilli by the Ziehl–Neelsen technique.

Morphometric image analysis was done as described by Schacker et al. (2002), with multiple digital photomicrographs (Olympus) of sections (three mice per group) at x200 magnification. Photomicrographs were imported into Photoshop 7.0 (Adobe Systems) and a colour sampler tool was used to gate shades of white, which represent unaffected alveolar spaces. The remaining non-selected areas of the field were removed, the resulting image was loaded into Scion Image Beta 4.0.2 software (Scion Corporation), and the number of occupying pixels was quantified. The mean density and standard deviation were calculated for each section.

Immunostaining of paraffin-embedded lung sections was performed using goat and rabbit polyclonal antibodies directed against mouse CatG, CatB, CatD, CatH and G-CSF at a 1 : 100 dilution (San Cruz Biotechnology). Anti-Mac-3 (Biosciences) monoclonal antibody at 1 : 50 dilution was used to identify macrophages and monocytes. Biotin-conjugated rabbit anti-goat and goat anti-rabbit secondary antibodies (Vector Labs) were used at 1 : 200 dilution. The primary and secondary antibody concentrations were optimized for each application. Immunostaining reactions were visualized by the avidin–biotin complex method employing a Vectastain ABC alkaline phosphatase kit (Vector Labs), and 3,3'-diaminobenzidine as the substrate. The sections were counterstained with haematoxylin and mounted. The specificity of immunostaining was tested by substitution of the primary antibody with normal goat IgG and by preincubation of antibodies with blocking peptides. Whole sections were examined using a conventional microscope at x200–400 magnification and digitally photographed.

RT-PCR.
Analysis of gene expression levels was done by RT-PCR using manufacturer's reagents and protocols (Promega). RNA was isolated from mouse lungs by extraction in RNAzol and reverse transcribed with gene-specific reverse primers for mouse genes encoding CatG, B, D, H, cystatin C (CysC) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) control. Primers designed to amplify a 300 bp fragment based on GenBank-published sequences were as follows: catG (F, ACCCCTACATGGCATTTCTTC; R, ACATTTGGTCCATCTGCACTC), catB (F, CTTCCCATGTCGGCAATCAGAAC; R, AAGACATCTAGAGTACCCCCAAG), catD (F, CACGTCCTTTGACATCCACTACG; R, CAGCTCCTTCACCTCTTCCACAG), GAPDH (F, CAGCCGCATCTTCTTGTG; R, AGGGGGGCTAAGCAGTTG), catH (F, TGCCCAAGCCTTCAACAATCATG; R, AAGTACCCATTCTCCCCCCAC TG), cysC (F, CTGTGAGCGAGTACAACAAGG; R, GGAGCACAAGTAAGGAACAG). The first-strand cDNA synthesis reaction was carried out at 42 °C for 60 min in a 25 µl reaction mixture consisting of 1·0 µg RNA, 1 µl/20 µM reverse primer, 5 µl 5x M-MLV buffer, 5 µl dNTP 10 mM mix, 1 µl/200 U M-MLV reverse transcriptase, 25 U rRNasin ribonuclease inhibitor and nuclease-free H₂O. After cDNA synthesis, 5 µl cDNA was added to 44 µl PCR master mix (3 µl 25 mM MgCl₂, 5 µl 10x PCR buffer, 0·2 µl/1·0 U Taq DNA polymerase, 4 µl 10 mM dNTP mix and DEPC-treated water), and 1·0 µl forward and reverse primers (20 µM) was added to each reaction tube. The thermal cycling parameters were: 94 °C (5 min), 20 or 30 cycles of 94 °C (45 s), 55 °C (45 s), 72 °C (45 s), and a final extension at 72 °C (10 min). The RT-PCR products were separated on 1 % (w/v) TBE agarose gels, visualized by staining with ethidium bromide, and the intensity of bands was quantified by densitometry scanning.

Statistical analysis.
This consisted of simple descriptive statistical methods such as mean, standard deviation and two-tailed Student's t test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

 
Increased virulence and pathogenicity of the acr knockout strain
As shown in Fig. 1, infection with either the parental strain H37Rv or acr resulted in an exponential increase in bacillary burdens in the lung. An increase in recovered c.f.u. of approximately 0·5 log unit was observed after week 4, while a 1 log unit increase was found after week 6 with both strains. Interestingly, infection with the acr strain resulted in significantly higher lung burdens (by approx. 2 log units) than H37Rv at weeks 2, 4 and 6 post-infection. In other experiments, where c.f.u. were determined using whole lung and spleen homogenates at day 1 (24 h) and week 2 post-infection, it was found that both H37Rv and acr strains gave similar c.f.u. at day 1 (2·6x10⁴ and 3x10⁴, respectively), showing that mice received an equal-sized inoculum. However, at week 2 the acr strain was at least 1 log unit higher (2·6x10⁵) in comparision to H37Rv (3·6x10⁴), confirming the increased virulence of acr.

View larger version (12K): [in this window] [in a new window]   Fig. 1. Bacillary burdens in lungs of C57BL6 mice (n=3) after 2, 4 and 6 weeks of infection via tail vein inoculation (1x106 bacilli) with M. tuberculosis H37Rv or acr. Infection with acr results in a significant increase in lung bacillary c.f.u. in comparison to parental strain H37Rv. *, P<0·05 (Student's t test) in comparison to H37Rv infected mice.

View larger version (62K): [in this window] [in a new window]   Fig. 2. (a) Haematoxylin/eosin-stained paraffin-embedded lung sections of normal mice and mice infected for 6 weeks with M. tuberculosis H37Rv or acr. In comparison to the normal lung architecture observed in uninfected mice, infection with either H37Rv or acr results in marked infiltration of cells into the alveolar walls, resulting in extensive obliteration of alveolar air spaces and multifocal coalescing diffused granulomatous lesions. This pathology is more pronounced in acr mice. Bar: 100 µm. (b) Morphometric image analysis of lung sections showing a significant decrease in lung volume density (estimated by pixels) at 2, 4 and 6 weeks post-infection in mice (n=3 per group) infected with H37Rv (grey bars) or acr (black bars) compared with normal lungs (white bars) The acr-infected mice show a 2-fold decrease (weeks 2, 4) in lung volume density in comparison to H37Rv. *, P<0·05 (Student's t test) in comparison to normal lung. ^, P<0·05 (Student's t test) in comparison to H37Rv-infected lung.

The finding that infection with the acr mutant, in comparison to parental wild-type H37Rv, results in higher c.f.u. and a more severe pathology in the lung suggests that the acr deletion results in hypervirulence.

Infection with H37Rv and acr results in increased expression of cathepsins in the lung
We recently demontrated that M. tuberculosis infection of THP-1 monocytes results in the differential expression of lysosomal cathepsin proteases. CatG, a neutral serine protease, was shown to have tuberculocidal activity and to be downregulated after M. tuberculosis infection, while the acidic-type cathepsins (CatB and CatD) were upregulated (Rivera-Marrero et al., 2004). Here, we asked if this would also occur in the lungs of infected mice, and if infection with the hypervirulent acr strain would result in altered expression of cathepsins. By RT-PCR analysis we found that a 4 week infection with either H37Rv or acr (3 mice each) resulted in a significant downregulation of catG mRNA expression, when compared to non-infected controls (Fig. 3). However, the expression of catD, catB and catH mRNA was upregulated after infection, and was slightly higher in mice infected with acr. As estimated by densitometry, the expression of catD, catB and catH mRNA in the acr-infected mice was 1·2-, 1·1- and 3·0-fold greater, respectively, than in the H37Rv-infected mice. No changes in the expression of the housekeeping GAPDH gene were detected. We also tested for the expression of cysC, which encodes cystatin C, an inhibitor of cathepsin cysteine proteases such as catB and catH. A small increase (1·2-fold) in cysC mRNA expression was seen in the infected lungs, but no differences were detected between H37Rv and acr. These results demonstrate that M. tuberculosis infection of mice results in differential gene regulation of cathepsin proteases in lung tissue, with catG downregulated and catB, catD and catH upregulated.

View larger version (70K): [in this window] [in a new window]   Fig. 3. RT-PCR analysis to determine differences in the expression of cathepsin G, D, B and H genes in the lungs of mice after 4 weeks infection with H37Rv or acr. After infection with either strain, catG gene expression is decreased compared with that in normal lungs (N), while the expression of catD, catB and catH is slightly increased, the increase being slightly higher in the mice infected with acr. A small increase in cysC expression is seen in the infected lungs with either H37Rv or acr, while no changes were detected in the housekeeping gene GAPDH.

View larger version (193K): [in this window] [in a new window]   4 Immunohistochemical staining of M. tuberculosis-infected lungs to detect differential expression of cathepsins. Paraffin-embedded lung sections of normal versus mice infected with H37Rv and acr for 6 weeks were immunostained for (a) CatG and B, or (b) CatD and H, plus normal mouse IgG control. (a) Normal lung positively stained for CatG shows diffused brownish staining of alveolar epithelial cells, while lungs infected with H37Rv and acr show no CatG staining, but only the haematoxylin counterstain. Normal lung also stained mildly for CatB, but very strong staining is observed after infection with either M. tuberculosis strain. (b) Normal lung shows mild staining for CatD and CatH, whereas very strong staining is seen after infection with H37Rv or acr. Clusters of macrophages surrounding areas of granulomatous inflammation (arrows) are seen positively stained for (a) CatB and (b) CatD and CatH, while control IgG showed no cross-reactivity. (c) CatD expression in mouse lung granulomas from acr-infected mice co-localizes with macrophages expressing the macrophage-specific marker Mac-3. Bars: 100 µm (a, b); 50 µm (c).

Taken together, these results demonstrate that M. tuberculosis infection results in the differential expression of cathepsin proteases in the lung, particularly in areas of granulomatous inflammation.

Infection with acr results in increased serum levels of TNF-, IFN- and G-CSF
Since the proinflammatory cytokines TNF- and IFN- play such important roles in the pathogenesis of TB, we asked if their expression would be altered in mice infected with the acr strain vs H37Rv. Fig. 5 shows the levels of cytokines TNF- and IFN- in the serum of normal and infected mice after 2, 4 and 6 weeks of infection with H37Rv or acr. An increase in TNF- (2–3-fold) was detected in mice infected with H37Rv for 2, 4 and 6 weeks, in comparison with control mice. Also, IFN- was elevated (3–5-fold) in mice infected for 2 and 4 weeks. Most significantly, infection with the acr strain resulted in a substantial increase in TNF- and IFN- in all mice, in comparison to uninfected and H37Rv-infected mice. The highest levels of TNF- (120–160 pg ml^–1) and IFN- (150–170 pg ml^–1) were at weeks 2 and 4 post-infection and were at least 2–3-fold (TNF-) and 3–4-fold (IFN-) higher than those in the H37Rv-infected group. These results demonstrate that infection with the acr strain results in an exaggerated induction of proinflammatory cytokines TNF- and IFN-.

View larger version (15K): [in this window] [in a new window]   Fig. 5. Levels of TNF- and IFN- in the serum of C57BL6 mice after 2, 4 and 6 weeks of infection with H37Rv or acr, as determined by the Cytokine Cytometric Bead Array method (see Methods). As shown, infection with the acr strain results in a marked increase in TNF- and IFN- at weeks 2 and 4 in comparison with normal (N) and H37Rv-infected mice. The results shown are the mean of n=3 and the bars denote standard deviations. ^, P<0·05 (Student's t test) in comparison to uninfected control. *, P<0·05 (Student's t test) in comparison to H37Rv-infected mice.

View larger version (13K): [in this window] [in a new window]   Fig. 6. Serum levels of G-CSF in C57BL6 mice after 2, 4 and 6 weeks of infection with H37Rv and acr, as determined by the multiplex bead immunoassay (see Methods). Infection of mice with acr resulted in increased levels of G-CSF, with the most significant increase reaching 530 pg ml–1 (5-fold over normal, 3-fold over H37Rv-infected mice) at week 2 post-infection. G-CSF at weeks 4 and 6 was also higher in the acr group but not statistically different from the H37Rv group. Results are the mean of n=3 and bars denote standard deviations. ^, P<0·05 (Student's t test) in comparison to uninfected control. *, P<0·05 (Student's t test) in comparison to H37Rv-infected mice.

View larger version (54K): [in this window] [in a new window]   Fig. 7. G-CSF immunostaining of representative lung sections of normal vs infected mice after 2 weeks infection with H37Rv or acr. As shown, G-CSF is expressed by infiltrating macrophages and is more abundant in the acr-infected lung. Bar: 50 µm.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

 
There is little information on the role of the acr gene during the pathogenesis of TB and its implications in latency and reactivation. The purpose of this study was to determine the effects of the acr mutation in vivo, with particular attention to its role in infectivity, modulation of the host response, and pathogenicity in the mouse model. First, we demonstrated that intravenous tail infection of C57BL6 mice with acr resulted in exponential multiplication of bacilli with at least a 2 log increase in c.f.u. in lungs after 2, 4 and 6 weeks of infection, in comparison to its parental strain H37Rv. However, in this experiment, we noticed low bacillary loads in lung with both strains, since intravenous inoculation with 10⁶ bacilli typically results in at least a 1 log increase by 2 weeks (10⁵ c.f.u. ml^–1 g^–1 recovered) and a 2–3 log increase by 4-6 weeks (10⁶–10⁷ c.f.u. ml^–1 g^–1 recovered) (Copenhaver et al., 2004). We attribute the low numbers in Fig. 1 to the fact that only a small portion of the lung was utilized for c.f.u. counts and therefore the calculation of c.f.u. ml^–1 g^–1 may not be a true estimation of the bacillary load in the entire lung. Using the entire lung and spleen for c.f.u. counts we showed a bacillary load of >10⁴ at day 1 with both strains, a small increase of >10 000 c.f.u. at week 2 with H37Rv, and a 1 log increase (2·6x10⁵) with acr. The small increase in c.f.u. obtained with wild-type H37Rv at 2 weeks post-infection after such a large inoculum (10⁶) can be misleading and give the impression that the strain is not growing well. However, similar bacillary loads with H37Rv have been reported by another group in this model (Copenhaver et al., 2004).

Evidence for increased virulence of the acr strain also comes from an unpublished study by Smith et al. (2000). In that study, C57BL/6 mice were infected via aerosol with 25 c.f.u., and lung bacillary c.f.u. were determined at 1, 2, 4, 6 and 12 weeks. That study also showed an increase in c.f.u. of 1·1–1·2 log units after infection with acr. Therefore, based on these studies we can conclude that acr has increased virulence in mice equivalent to 1–2 log units in comparison to parental H37Rv. Thus, in contrast to the in vitro results obtained by Yuan et al. (1998), these in vivo studies demonstrate that the acr mutation does not impair the growth of M. tuberculosis bacilli in mouse tissues during an acute infection but, on the contrary, results in a state of hypervirulence. The observation that acr expression is upregulated in vitro during hypoxic and stationary-phase growth conditions (Schnappinger et al., 2003; Yuan et al., 1996, 1998), and in vivo in mouse (Shi et al., 2003; Timm et al., 2003) and human lungs (Timm et al., 2003), suggests that it is an important gene for survival of bacilli under stress and is possibly part of a genetic programme which allows adaptation to hypoxic microenvironments of the host. This adaptation may involve the shutdown of genes necessary for aerobic metabolic pathways, with the ultimate consequence of entering into a state of non-replicative stasis or latency. A study by whole-genome microarray analysis showed that the expression of more than 100 genes is altered in vitro by growth under hypoxic conditions and that many of the repressed genes are involved in aerobic metabolism (Sherman et al., 2001). In addition, many genes were induced under hypoxia, including members of the LuxR two-component response regulators such as Rv3133c, which when disrupted, resulted in the elimination of the hypoxic regulation of acr (Sherman et al., 2001). Therefore, the disruption of the acr gene may also affect the expression of other metabolic genes associated with growth and survival in vivo, allowing bacilli to replicate faster and preventing them from going dormant under stress conditions. Further studies are needed to determine the mechanism that affords the acr strain increased bacillary multiplication in vivo.

Latency is often associated with decreased bacillary burdens in tissues, and lack of disease symptoms and pathology (Flynn & Chan, 2001). The finding that infection with acr resulted in exacerbated lung pathology, in comparison with H37Rv, supports the hypothesis that disruption of acr can result in increased virulence and pathogenicity; however, its role in the highly complex events leading to latency and reactivation in vivo is not known. Since the acr gene has been associated with hypoxia-induced dormancy in vitro, events that are reminiscent of latency in vivo, we wanted to know if its disruption would have an effect on the host response to infection. To test this, we proceeded to determine the effects of acr infection on the host immune and tissue remodelling responses.

We demonstrated by RT-PCR and immunostaining of lungs that M. tuberculosis infection with both H37Rv and acr strains resulted in the differential expression of cathepsin proteases. The neutral serine protease CatG was downregulated, while the acidic-type cathepsins CatB, D and H were upregulated after infection and their expression associated with macrophages within granulomas. CatB, D and H expression was slightly more elevated in the acr-infected mice, suggesting that the increased pathogenicity observed with this strain could be the result of increased protease induction in the lung. In previous studies we have shown that tissue matrix proteases, such as metalloproteinases (Rivera-Marrero et al., 2000, 2002) and cathepsins (Rivera-Marrero et al., 2004), are important in M. tuberculosis infection. Cathepsins are a large family of lysosomal proteases that not only function in intralysosomal protein degradation, but participate in tissue remodelling responses by degrading extracellular matrix proteins. CatB and CatH are cysteine proteases, while CatD is an aspartyl protease (Wolters & Chapman, 2000). In particular, CatD has been shown in mature epithelioid macrophages surrounding the caseous and liquefied areas of pulmonary cavities in M. tuberculosis-infected rabbits (Converse et al., 1996). Therefore, our finding that cathepsins (B, D and H) are increased in lung after infection with both M. tuberculosis strains, but slightly higher with acr, suggests that these proteolytic enzymes are involved in the process of granuloma formation and play an important role in the pathogenesis of M. tuberculosis.

The downregulation of CatG in lungs after infection with both the H37Rv and acr was an interesting finding in this study. CatG is highly abundant in the azurophilic granules of neutrophils and monocytes (Senior & Campbell, 1984; Senior et al., 1982), and is synthesized during the promyelocytic and promonocytic stages of maturation, respectively (Bainton et al., 1971; van der Meer et al., 1981). In U937 monocytes, treatment with the phorbol ester 12-O-tetradecanoylphorbol 13-acetate (TPA) results in transcriptional downregulation of catG (Hanson et al., 1990; Ley et al., 1989; Welgus et al., 1986). Since there is no evidence that catG is expressed by alveolar epithelial pneumocytes, it is possible that the CatG detected in the normal lung is derived from circulating monocytes and/or neutrophils. Nonetheless, this is an important finding in view of our recent work showing that the downregulation of catG in THP-1 human monocytes after M. tuberculosis infection coincided with increased bacillary multiplication in cells and that CatG and its cationic peptide CG117-136 have tuberculocidal activity in vitro (Rivera-Marrero et al., 2004). The downregulation of catG in lung after infection may be advantageous to M. tuberculosis bacilli and represent an important mechanism for evasion of host innate immune defences. However, further studies are needed to fully define the role of CatG in TB pathogenesis.

To explore the mechanisms by which the acr-infected mice show increased pathology we determined their cytokine profile. We found that acr-infected mice had very high levels of IFN- and TNF- at weeks 2 and 4 post-infection, in comparison to H37Rv-infected mice. It is well documented that both IFN- and TNF- play important roles in the control of a persistent TB infection (Flynn & Chan, 2001). IFN- is involved in macrophage activation (Dalton et al., 1993; Flynn et al., 1993) and the production of reactive nitrogen intermediates that can kill intracellular bacilli (Chan et al., 1992). Knockout mice for IFN- are highly susceptible to M. tuberculosis infection, succumbing to disseminated TB infection (Cooper et al., 1993; Flynn et al., 1993). TNF- is also very important for the control of M. tuberculosis, with effects on macrophage activation, production of reactive nitrogen intermediates, granuloma formation and pathology (Bean et al., 1999; Flynn et al., 1995; Kindler et al., 1989). High levels of TNF- in lung cause extreme pathology, as shown in a murine model by infection with a recombinant BCG strain that secreted TNF- at the site of infection (Bekker et al., 2000). However, the effects of TNF- are also dose-dependent and determine whether the cytokine is protective or destructive. Mice functionally deficient in TNF- develop fatal acute M. tuberculosis infections (Bean et al., 1999; Flynn et al., 1995; Roach et al., 2002), characterized by extensive necrosis in the lungs and infected organs, and failure to form functional granulomas. Although it is widely accepted that IFN- and TNF- have a protective role during active TB infection, their roles in latency and reactivation are not completely understood. Our finding that the acr strain causes extreme pathology, which is concomitant with increases in INF- and TNF-, may suggest that these cytokines are elevated during reactivation as a protective mechanism to prevent disseminated infection. However, their high levels in the lung may also contribute to the exacerbated pathology either directly (Bekker et al., 2000) or via the activation of proteolytic pathways involving the action of cathepsins (Wang et al., 2000).

We also showed that mice infected with acr had higher levels of G-CSF in serum than those infected with H37Rv and that this could be attributed to the elevated expression by lung macrophages. G-CSF is involved in the proliferation, survival, maturation and functional activation of cells from the neutrophilic granulocyte lineage (Basu et al., 2002). Serum G-CSF levels rapidly increase in response to bacterial infection and cell-mediated immune responses, at times when granulocyte levels become elevated, suggesting that G-CSF is a crucial regulator of an emergency response involving granulocyte production (Cheers et al., 1988; Demetri & Griffin, 1991; Nicola, 1989). Bacterial products such as endotoxin, or inflammatory cytokines induced during infections, such as TNF, interleukin (IL-1) and IFN-, are the major stimulators of G-CSF production in vivo and result in a rapid but transient elevation in serum G-CSF levels. G-CSF is produced mainly by haematopoietic cells, such as monocytes/macrophages, and lymphocytes (Nicola et al., 1983; Sallerfors, 1994). Other cells, such as fibroblasts (Kaushansky et al., 1988), endothelial cells (Zsebo et al., 1988), astrocytes (Aloisi et al., 1992) and bone marrow stromal cells (Fibbe et al., 1988), can also produce G-CSF following activation by LPS, IL-1 or TNF-. Therefore, the rapid increase in G-CSF observed in acr-infected mice (at week 2) could be caused directly by the bacillus, or indirectly by the action of cytokines TNF- and IFN-. G-CSF could be involved in triggering a rapid mobilization of granulocytes to sites of granulomatous inflammation in the acr-infected mice that results in the exacerbated pathogenesis.

In conclusion, this work provides tantalizing new information about the in vivo pathogenicity of the acr mutant of M. tuberculosis. We demonstrate that the acr strain is hypervirulent in mice and causes exacerbated lung pathology, and that this effect could be the result of increased induction of pro-inflammatory cytokines (TNF-, IFN-, G-CSF) and lysosomal cathepsin proteases (CatB, D, H) in the lung. Future studies are designed to define the molecular mechanisms by which Acr affects the pathogenesis of M. tuberculosis and its role in latency and reactivation.

	ACKNOWLEDGEMENTS

 
This work was supported by a Veterans Affairs Merit Review Award (to C. A. Rivera-Marrero) and National Institute of Allergy and Infectious Diseases Grant 1RO1 AI-37937 (J. Roman). We thank James B. Harten for the excellent technical assistance in the processing of lung samples for histochemical analysis. We thank Dr William Shafer, Division of Infectious Diseases, Emory University School of Medicine, and Atlanta VA Medical Center, for helpful discussions and critical reading of this manuscript.

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

 
Adler, J. J. & Rose, D. N. (1996). Transmission and pathogenesis of tuberculosis. In Tuberculosis, pp. 129–140. Edited by W. N. Rom & S. M. Gray. Boston, MA: Little, Brown.

Aloisi, F., Care, A., Borsellino, G. & 7 other authors (1992). Production of hemo-lymphopoietic cytokines (IL-6, IL-8, colony-stimulating factors) by normal human astrocytes in response to IL-1 beta and tumor necrosis factor-alpha. J Immunol 149, 2358–2366.[Abstract]

Bainton, D. F., Ullyot, J. L. & Farquhar, M. G. (1971). The development of neutrophilic polymorphonuclear leukocytes in human bone marrow. J Exp Med 134, 907–934.[Abstract]

Basu, S., Dunn, A. & Ward, A. (2002). G-CSF: function and modes of action. Int J Mol Med 10, 3–10.[Medline]

Bean, A. G., Roach, D. R., Briscoe, H., France, M. P., Korner, H., Sedgwick, J. D. & Britton, W. J. (1999). Structural deficiencies in granuloma formation in TNF gene-targeted mice underlie the heightened susceptibility to aerosol Mycobacterium tuberculosis infection, which is not compensated for by lymphotoxin. J Immunol 162, 3504–3511.[Abstract/Free Full Text]

Bekker, L. G., Moreira, A. L., Bergtold, A., Freeman, S., Ryffel, B. & Kaplan, G. (2000). Immunopathologic effects of tumor necrosis factor alpha in murine mycobacterial infection are dose dependent. Infect Immun 68, 6954–6961.[Abstract/Free Full Text]

Bloom, B. R. & Murray, C. J. L. (1992). Commentary on a reemergent killer. Science 257, 1055–1064.[Abstract/Free Full Text]

Chan, J., Xing, Y., Magliozzo, R. S. & Bloom, B. R. (1992). Killing of virulent Mycobacterium tuberculosis by reactive nitrogen intermediates produced by activated murine macrophages. J Exp Med 175, 1111–1122.[Abstract/Free Full Text]

Cheers, C., Haigh, A. M., Kelso, A., Metcalf, D., Stanley, E. R. & Young, A. M. (1988). Production of colony-stimulating factors (CSFs) during infection: separate determinations of macrophage-, granulocyte-, granulocyte-macrophage-, and multi-CSFs. Infect Immun 56, 247–251.[Abstract/Free Full Text]

Converse, P. J., Dannenberg, A. M., Jr, Estep, J. E., Sugisaki, K., Abe, Y., Schofield, B. H. & Pitt, L. M. (1996). Cavitary tuberculosis produced in rabbits by aerosolized virulent tubercle bacilli. Infect Immun 64, 4776–4787.[Abstract]

Cooper, A. M., Dalton, D. K., Stewart, T. A., Griffin, J. P., Russell, D. G. & Orme, I. M. (1993). Disseminated tuberculosis in interferon gamma gene-disrupted mice. J Exp Med 178, 2243–2247.[Abstract/Free Full Text]

Copenhaver, R. H., Sepulveda, E., Armitige, L. Y., Actor, J. K., Wagner, A., Norris, S. J., Hunter, R. L. & Jagannath, C. (2004). A mutant of Mycobacterium tuberculosis H37Rv that lacks expression of antigen 85A is attenuated in mice but retains vaccinogenic potential. Infect Immun 72, 7084–7095.[Abstract/Free Full Text]

Cunningham, A. F. & Spreadbury, C. L. (1998). Mycobacterial stationary phase induced by low oxygen tension: cell wall thickening and localization of the 16-kilodalton alpha-crystallin homolog. J Bacteriol 180, 801–808.[Abstract/Free Full Text]

Dalton, D. K., Pitts-Meek, S., Keshav, S., Figari, I. S., Bradley, A. & Stewart, T. A. (1993). Multiple defects of immune cell function in mice with disrupted interferon-gamma genes. Science 259, 1739–1742.[Abstract/Free Full Text]

Dannenberg, A. M., Jr (1993). Immunopathogenesis of pulmonary tuberculosis. Hosp Pract (Off Ed) 28, 51–58.[Medline]

Dannenberg, A. M. & Rook, G. A. W. (1994). Pathogenesis of pulmonary tuberculosis: an interplay of tissue-damaging and macrophage-activating immune responses – dual mechanisms that control bacillary multiplication. In Tuberculosis: Pathogenesis, Protection and Control, pp. 459–483. Edited by R. Blood. Washington, DC: American Society for Microbiology.

Demetri, G. D. & Griffin, J. D. (1991). Granulocyte colony-stimulating factor and its receptor. Blood 78, 2791–2808.[Free Full Text]

Dye, C., Scheele, S., Dolin, P., Pathania, V. & Raviglione, M. C. (1999). Consensus statement. Global burden of tuberculosis: estimated incidence, prevalence, and mortality by country. WHO Global Surveillance and Monitoring Project. JAMA 282, 677–686.[Abstract/Free Full Text]

Fibbe, W. E., van Damme, J., Billiau, A., Goselink, H. M., Voogt, P. J., van Eeden, G., Ralph, P., Altrock, B. W. & Falkenburg, J. H. (1988). Interleukin 1 induces human marrow stromal cells in long-term culture to produce granulocyte colony-stimulating factor and macrophage colony-stimulating factor. Blood 71, 430–435.[Abstract/Free Full Text]

Flynn, J. L. & Chan, J. (2001). Tuberculosis: latency and reactivation. Infect Immun 69, 4195–4201.[Free Full Text]

Flynn, J. L., Chan, J., Triebold, K. J., Dalton, D. K., Stewart, T. A. & Bloom, B. R. (1993). An essential role for interferon gamma in resistance to Mycobacterium tuberculosis infection. J Exp Med 178, 2249–2254.[Abstract/Free Full Text]

Flynn, J. L., Goldstein, M. M., Chan, J., Triebold, K. J., Pfeffer, K., Lowenstein, C. J., Schreiber, R., Mak, T. W. & Bloom, B. R. (1995). Tumor necrosis factor-alpha is required in the protective immune response against Mycobacterium tuberculosis in mice. Immunity 2, 561–572.[CrossRef][Medline]

Groenen, P. J., Merck, K. B., de Jong, W. W. & Bloemendal, H. (1994). Structure and modifications of the junior chaperone alpha-crystallin. From lens transparency to molecular pathology. Eur J Biochem 225, 1–19.[Medline]

Hanson, R. D., Connolly, N. L., Burnett, D., Campbell, E. J., Senior, R. M. & Ley, T. J. (1990). Developmental regulation of the human cathepsin G gene in myelomonocytic cells. J Biol Chem 265, 1524–1530.[Abstract/Free Full Text]

Horwitz, J. (1992). Alpha-crystallin can function as a molecular chaperone. Proc Natl Acad Sci U S A 89, 10449–10453.[Abstract/Free Full Text]

Imboden, P. & Schoolnik, G. K. (1998). Construction and characterization of a partial Mycobacterium tuberculosis cDNA library of genes expressed at reduced oxygen tension. Gene 213, 107–117.[CrossRef][Medline]

Kaushansky, K., Lin, N. & Adamson, J. W. (1988). Interleukin-1 stimulates fibroblasts to synthesize granulocyte-macrophage and granulocyte colony-stimulating factors. Mechanism for the hematopoietic response to inflammation. J Clin Invest 81, 92–97.[Medline]

Kindler, V., Sappino, A. P., Grau, G. E., Piguet, P. F. & Vassalli, P. (1989). The inducing role of tumor necrosis factor in the development of bactericidal granulomas during BCG infection. Cell 56, 731–40.[CrossRef][Medline]

Lee, B. Y., Hefta, S. A. & Brennan, P. J. (1992). Characterization of the major membrane protein of virulent Mycobacterium tuberculosis. Infect Immun 60, 2066–2074.[Abstract/Free Full Text]

Ley, T. J., Connolly, N. L., Katamine, S., Cheah, M. S., Senior, R. M. & Robbins, K. C. (1989). Tissue-specific expression and developmental regulation of the human fgr proto-oncogene. Mol Cell Biol 9, 92–99.[Abstract/Free Full Text]

Nicola, N. A. (1989). Haematopoietic cell growth factors and their receptors. Annu Rev Biochem 58, 45–77.[Medline]

Nicola, N. A., Metcalf, D., Matsumoto, M. & Johnson, G. R. (1983). Purification of a factor inducing differentiation in murine myelo- monocytic leukemia cells. Identification as granulocyte colony- stimulating factor. J Biol Chem 258, 9017–9023.[Abstract/Free Full Text]

Opie, E. L. & Aronson, J. D. (1927). Tubercle bacilli in latent tuberculous lesions and in lung tissue without tuberculous lesions. Arch Pathol Lab Med 4, 1.

Rivera-Marrero, C. A., Schuyler, W. & Roman, J. (2000). Induction of MMP-9 mediated gelatinolytic activity in human monocytic cells by cell wall components of M. tuberculosis. Microb Pathog 29, 231–244.[CrossRef][Medline]

Rivera-Marrero, C. A., Schuyler, W., Roser, S., Ritzenthaler, J. & Roman, J. (2002). Mycobacterium tuberculosis induction of matrix metalloproteinase-9: the role of mannose and receptor-mediated mechanisms. Am J Physiol 282, 546–555.

Rivera-Marrero, C. A., Stewart, J. N., Shafer, W. M. & Roman, J. (2004). The down-regulation of cathepsin G in THP-1 monocytes after infection with M. tuberculosis is associated with increased intracellular survival of bacilli. Infect Immun 72, 5712–5721.[Abstract/Free Full Text]

Roach, D. R., Bean, A. G., Demangel, C., France, M. P., Briscoe, H. & Britton, W. J. (2002). TNF regulates chemokine induction essential for cell recruitment, granuloma formation, and clearance of mycobacterial infection. J Immunol 168, 4620–4627.[Abstract/Free Full Text]

Robertson, H. E. (1933). Persistence of tuberculous infection. Am J Pathol 9, 711.

Sallerfors, B. (1994). Endogenous production and peripheral blood levels of granulocyte-macrophage (GM-) and granulocyte (G-) colony-stimulating factors. Leuk Lymphoma 13, 235–247.[Medline]

Schacker, T. W., Nguyen, P. L., Beilman, G. J., Wolinsky, S., Larson, M., Reilly, C. & Haase, A. T. (2002). Collagen deposition in HIV-1 infected lymphatic tissues and T cell homeostasis. J Clin Invest 110, 1133–1139.[CrossRef][Medline]

Schnappinger, D., Ehrt, S., Voskuil, M. I. & 8 other authors (2003). Transcriptional adaptation of Mycobacterium tuberculosis within macrophages: insights into the phagosomal environment. J Exp Med 198, 693–704.[Abstract/Free Full Text]

Selwyn, P. A., Hartel, D., Lewis, V. A., Schoenbaum, E. E., Vermund, S. H., Klein, R. S., Walker, A. T. & Freidland, G. H. (1989). A prospective study of the risk of tuberculosis among intravenous drug users with human immunodeficiency virus infection. N Engl J Med 320, 545–550.[Abstract]

Senior, R. M. & Campbell, E. J. (1984). Cathepsin G in human mononuclear phagocytes: comparisons between monocytes and U937 monocyte-like cells. J Immunol 132, 2547–2551.[Abstract]

Senior, R. M., Campbell, E. J., Landis, J. A., Cox, F. R., Kuhn, C. & Koren, H. S. (1982). Elastase of U-937 monocyte like cells. Comparisons with elastases derived form human monocytes and neutrophils and murine macrophagelike cells. J Clin Invest 69, 384–393.[Medline]

Sherman, D. R., Voskuil, M., Schnappinger, D., Liao, R., Harrell, M. I. & Schoolnik, G. K. (2001). Regulation of the Mycobacterium tuberculosis hypoxic response gene encoding -crystallin. Proc Natl Acad Sci U S A 98, 7534–7539.[Abstract/Free Full Text]

Shi, L., Jung, Y.-J., Tyagi, S., Gennaro, M. L. & North, R. J. (2003). Expression of Th1-mediated immunity in mouse lungs induces a Mycobacterium tuberculosis transcription pattern characteristic of nonreplicating persistence. Proc Natl Acad Sci U S A 100, 241–246.[Abstract/Free Full Text]

Smith, S., Tan, S., Jeromsky, E., Liao, R., Yuan, Y., Wilson, C. B. & Sherman, D. S. (2000). Enhanced in vivo growth of -crystallin deficient M. tuberculosis. In TB 2000 – Past, Present, and Future, ASM Conference on Tuberculosis, New York. Washington, DC: American Society for Microbiology.

Timm, J., Post, F. A., Bekker, L.-G. & 9 other authors (2003). Differential expression of iron-, carbon-, oxygen-responsive mycobacterial genes in lungs of chronically infected mice and tuberculosis patients. Proc Natl Acad Sci U S A 100, 14321–14326.[Abstract/Free Full Text]

van der Meer, J. W., van de Gevel, J. S., Beelen, R. H., Fluitsma, D., Hoefsmit, E. C. & van Furth, R. (1981). Culture of human bone marrow in the Teflon culture bag: identification of the human monoblast. J Reticuloendothel Soc 32, 355–369.

Verbon, A., Hartskeerl, R. A., Schuitema,