Hypoxia and tissue destruction in pulmonary TB

Background It is unknown whether lesions in human TB are hypoxic or whether this influences disease pathology. Human TB is characterised by extensive lung destruction driven by host matrix metalloproteinases (MMPs), particularly collagenases such as matrix metalloproteinase-1 (MMP-1). Methods We investigated tissue hypoxia in five patients with PET imaging using the tracer [18F]-fluoromisonidazole ([18F]FMISO) and by immunohistochemistry. We studied the regulation of MMP secretion in primary human cell culture model systems in normoxia, hypoxia, chemical hypoxia and by small interfering RNA (siRNA) inhibition. Results [18F]FMISO accumulated in regions of TB consolidation and around pulmonary cavities, demonstrating for the first time severe tissue hypoxia in man. Patlak analysis of dynamic PET data showed heterogeneous levels of hypoxia within and between patients. In Mycobacterium tuberculosis (M.tb)-infected human macrophages, hypoxia (1% pO2) upregulated MMP-1 gene expression 170-fold, driving secretion and caseinolytic activity. Dimethyloxalyl glycine (DMOG), a small molecule inhibitor which stabilises the transcription factor hypoxia-inducible factor (HIF)-1α, similarly upregulated MMP-1. Hypoxia did not affect mycobacterial replication. Hypoxia increased MMP-1 expression in primary respiratory epithelial cells via intercellular networks regulated by TB. HIF-1α and NF-κB regulated increased MMP-1 activity in hypoxia. Furthermore, M.tb infection drove HIF-1α accumulation even in normoxia. In human TB lung biopsies, epithelioid macrophages and multinucleate giant cells express HIF-1α. HIF-1α blockade, including by targeted siRNA, inhibited TB-driven MMP-1 gene expression and secretion. Conclusions Human TB lesions are severely hypoxic and M.tb drives HIF-1α accumulation, synergistically increasing collagenase activity which will lead to lung destruction and cavitation.

of patient motion during PET acquisition, the last 8 frames of the dynamic PET image series were non-rigidly registered to the first frame using symmetric normalization (SyN) with mutual information as the similarity metric [3]. All PET frames were then registered to CT using affine transformation. The last three of the registered frames (120-135mins post injection) were converted to a TBR image through normalization with the radioactivity concentration in a bilateral lateral muscle region assumed to represent normoxic tissue. All 9 frames were used to produce a Ki map using Patlak plot analysis. Regions of interest (ROI) were defined on the TBR images and ROI values were determined by placing the ROIs on the TBR and Ki maps.
Cultured cells were infected with M.tb at mid-logarithmic growth at an optical density of 0.6 (Biowave cell density meter, WPA, Cambridge, UK). To count intracellular bacilli, infected MDMs were lysed with 0.1% Triton-X. Aliquots of cell lysates were serially diluted 10-fold in 7H9 liquid and 4 dilutions of each sample were plated on 7H11 agar and incubated for 14 days befoe determining the number of CFU.

Monocyte Purification and Maturation
Monocytes were isolated from single donor leucocyte cones (National Blood Transfusion Service, London, UK) by density gradient centrifugation (Amersham Biosciences) and re-suspended in RPMI. Monocytes were purified by adhesion for 1 hour followed by washing 6 times in HBSS to remove non-adherent cells. Monocytes were matured for 4 days in 100ng/ml M-CSF and then rested for 1 day in RPMI containing no M-CSF. The resultant monocyte-derived macrophages (MDMs) were infected with M.tuberculosis H37RV at an MOI of 1. MMP gene expression was analyzed at 24 hours and MMP secretion measured at 3 days after infection. Conditioned media from M.tb infected monocytes (CoMTb) was produced by infecting monocytes at a multiplicity of infection (MOI) of 1. After 24 hours, cell culture medium was harvested, centrifuged at 13,000rpm for 5 minutes and filtered to remove infectious material using a 0.2μm Anopore filter (Whatman, UK) [4].

Hypoxia workstation
A custom designed hypoxia workstation was commissioned for the biological safety level 3 facility (Coy Laboratories, USA). The workstation comprised a sealed environment with an inbuilt humidified incubator accessed by air-tight gloves. Materials and reagents were passed in and out of the workstation via a sealed airlock, such that the atmosphere remained constant at all times. Temperature and CO2 were maintained at 37 0 C and 5% respectively and O2 levels (by N2 displacement) adjusted according to the experimental requirements. All parameters were monitored by digital sensors throughout and adjustments made automatically. In experiments comparing multiple oxygen levels, a hypoxic incubator was also used for 5% O2 (Galaxy 14S, New Brunswick, UK)

Casein Zymography
To analyze MMP-1 activity, samples were electrophoresed on 0.05% casein gels (Invitrogen) and incubated in collagenase buffer at 37 o C for 48h. All gels were run with a 5ng recombinant MMP-1 (Calbiochem, Merck Biosciences, UK) to standardise between gels.
Caseinolytic activity was revealed by Coomassie blue staining (Pharmacia) for one hour followed by destain in methanol: acetic acid: water [5].

Real-time PCR
Macrophages were lysed using Tri-Reagent (Sigma-Aldrich, Dorset, UK), and total RNA was extracted using PureLink RNA® Mini Kit (Invitrogen, Paisley, UK). 1µg RNA was reverse transcribed using QuantiTect Reverse Transcriptase Kit (Qiagen, Manchester, UK). qPCR reactions were performed in an ABI Prism 7700 (Applied Biosystems, Paisley, UK). MMP-1 cycle thresholds were quantified by comparison to an MMP-1 standard curve generated using known MMP-1 concentrations generated from a plasmid standard and then standardised to 18S rRNA. HIF-1α mRNA samples were analysed using the comparative threshold method. The Ct values of both calibrator and samples were normalized to 18S rRNA.

Measurement of MMP and TIMP concentrations
Total MMP and TIMP secretion in cell culture supernatants was measured by ELISA (R&D Systems, Abdingdon, UK) or on the Luminex 200 platform using MMP Luminex multiplex array (R&D Systems, Abingdon, UK) according to the manufacturer's instructions. The minimum level of detection for MMP-1 was 10 pg/ml.
Luminescence was detected with the ECL system (Amersham, UK) according to manufacturer's protocol.

Confocal microscopy
Primary human monocytes were matured to macrophages in 4-well Permanox plastic chamber slides (Thermo Fisher Scientific Inc. UK). Cells were fixed for 3 days following infection with 4% paraformaldehyde in PBS for 30 minutes. Cells were permeabilized with 0.5% saponin and blocked with 1% BSA for 1 hr. Cells were incubated with anti-MMP-1 primary Ab (1:200, Abcam, UK) overnight, washed and incubated with goat anti-mouse secondary Ab (1:100, Abcam, UK) and labelled with FITC for 2 hrs. Nuclei were visualized using DAPI. The images were scanned under an x60 oil immersion objective and to avoid bleed-through effects, each dye was scanned independently by a Leica TCS SP5 equipped with an argon-krypton mixed gas laser. Image analysis was with Leica confocal software (Leica Microsystems, Germany) and Image J v1.48a.

Promoter-reporter assay
A549 cells were transfected with either wild-type, truncation or deletion constructs of the MMP-1 promoter DNA inserted into the firefly luciferase expression plasmid pGL2-basic (gift from Dr. I. Clarke, University of East Anglia, U.K.) and control reporter plasmid, pRL-TK, constitutively expressing Renilla luciferase activity, with FuGene 6 (Roche) [6]. Cells were stimulated with CoMTb in the presence or absence of hypoxia then lysed after 24 h.
Luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega, Southampton, UK) analyzed with the Bio-Orbit 1253 luminometer (Labtech International). Renilla luciferase activity was used to normalize firefly activity to control for transfection efficiency.

Immunohistochemistry
HIF-1α immunohistochemistry was performed on formalin-fixed, paraffin-embedded lung biopsies from six patients with culture-proved M.tb infection and six non-infected control samples. Immunohistochemistry staining was performed using Bond III fully automated staining system with the Bond Polymer Refine Detection system and associated reagents (Leica Microsystems, Newcastle-Upon-Tyne, UK). Antigen retrieval was carried out using Epitope retrieval pH 6 (ER1) at 100°C for 30mins. The antibody HIF1-α (Abcam AB1) was applied for 30mins at a dilution of 1/500. The detection system uses DAB as chromogen and section were counterstained with haematoxylin.

Statistics
Statistical analysis was performed using GraphPad PRISM 6. Data was analyzed with oneway ANOVA and Tukey's post-test performed to account for multiple comparison of variables. A P value of 0.05 was considered significant. For all experiments, bars represent mean values of 3 samples ± SD. All graphs are representative of at least duplicate experiments.  n/a n/a n/a n/a n/a Consolidation -*The binary hypoxic scale denotes whether the lesion ROI Ki value is above a hypoxia threshold (0.00037 min -1 ) determined from the mean + 3SD of the Ki values within the lateral muscle background ROIs, assumed to represent normoxic tissue. Figure S1. Hypoxia does not cause cytotoxicity in NHBE cells NHBE cells were stimulated with CoMTb and incubated in normoxia or hypoxia (1% or 5% pO2; 5% CO2). Supernatant were harvested at 72 hrs. Hypoxia did not cause an increase in LDH release in either control or CoMTb stimulated cells.