The authors have declared no competing interests exist.
Conceived and designed the experiments: DJO IMO. Performed the experiments: AOH MHT. Analyzed the data: AOH MHT. Contributed reagents/materials/analysis tools: DJO. Wrote the manuscript: AOH DJO IMO. Revised the manuscript critically for important intellectual content: DJO IMO.
Tuberculosis is one of the world’s leading killers, stealing 1.4 million lives and causing 8.7 million new and relapsed infections in 2011. The only vaccine against tuberculosis is BCG which demonstrates variable efficacy in adults worldwide. Human infection with
We compared the bacterial burden, lung pathology and Gr1intCD11b+ myeloid-derived suppressor cell immune responses in
Tuberculosis is the primary cause of death from a bacterial disease, and is further exacerbated by the very extensive incidence of latent disease, as well as the emergence of drug-resistant forms of the bacillus [
That is not to say however that some more recently described inbred mouse models do not develop necrosis. Chronically infected mice on the C3Heb/FeJ background, for instance, gradually develop degenerating lesions. Mice in which genes have been deleted for gamma interferon, B and T cells [
There is increasing evidence that Gr1+ neutrophils are present in the granuloma and play a key role in the process of necrosis[
In the studies reported here, we used flow cytometry and cell sorting techniques to monitor the influx of granulocytes into the lungs of necrosis prone mouse strains NOS2-/-, RAG-/-, C3HeB/FeJ strains and wildtype C57BL/6 mice devoid of necrosis. We tracked the influx of Gr1+ cells to the lungs and demonstrate two distinct populations: one Gr-1hi, and a second, quite substantial Gr-1int population. Gr-1int cells have been described primarily in cancer, and are thought to be a more immature cell type, but there is evidence they can modulate both innate and acquired responses [
Flow cytometric analysis was performed on gene disrupted and wild-type (WT) mice in order to analyze the influx of leukocytes into the lungs. From day 30 onwards, a time at which studies have previously demonstrated [
Flow cytometry analysis was performed on single cell suspensions obtained from lungs of WT C57/BL6 or NOS2-/- mice. In these mice increased numbers of CD11b+ cells but not CD11b+CD11c+ cells could be observed as the infection progressed (A, B, C). In contrast, differences for CD11c+ and CD4+ cells only became evident at later time points (B, D). No significant differences were observed for CD8+ cells (E). Results are expressed as the mean values of log mean cell number (± SEM, n=5) in the lung. Bacterial burden (F) was enumerated at different time points after a low dose aerosol infection with
In contrast to C57BL/6, flow cytometric analysis clearly identified two CD11b+ populations in the lungs of
As suggested by FSC analysis, the increased number of CD11b+ cells present in NOS2 -/- mice could be separated in two populations based on cell size (A, B). Both populations of CD11b+ FSChigh cells and CD11b+FSClow cells are significantly increased in NOS2 -/- mice compared to C57/BL/6 mice (A, B). Contour plot is representative of day 45 after infection. FSC analysis clearly identified two CD11b+ populations in the lungs of NOS2 -/- mice (C). Results are expressed as the mean values fluorescence intensity (± SEM, n=5) in the Lung, Student t-test, *** p<0.001. Necrotic granulomas are only present in the lungs of NOS2 -/- mice (D). At different time points after infection, lungs were harvested and stained with H&E after fixation and paraffin embedding. For both murine strains, lesions are not present at day 15 after infection. Starting day 30, necrotic granulomas become evident in the lungs of NOS2 -/- mice but not WT C57BL/6. As observed at day 45 and 60, necrotic granulomas coalesce leading to severe lung consolidation at day 60 when most NOS2 -/- are moribund.
Further analysis of CD11b+FSClow and CD11b+FSChigh cells was performed by staining with anti-Gr1 antibody. As determined by mean fluorescence intensity, CD11b+FSChigh had significantly higher Gr1 expression than CD11b+FSClow (A). Starting at early time points NOS2-/- mice but not WT C57BL/6 mice have a significant number of cells giving an intermediate staining pattern for Gr-1 (B). At later time points, the influx of CD11b+FSChigh cells into the lungs of NOS2 -/- mice is also enhanced (C). Results are expressed as the Log mean cell number (± SEM, n=5) in the lung. Gr1high CD11b+ FSChigh cells express markers compatible with a monocytic lineage. Both populations of Gr1 expressing cells were further analyzed by flow cytometry using the monocytic markers Ly6C, CD14 and F4/80 (D). As determined by the mean fluorescence intensity for these markers, a more precise phenotype for these two cellular populations would be Gr1high CD11b+ FSChigh Ly6Chigh CD14+ F4/80+ and Gr1int CD11b+ FSClow Ly6Clow CD14low F4/80low, compatible with a monocytic and granulocytic lineage, respectively. Results are expressed as the log mean cell number (± SEM, n=5) in the Lung. ***Student t test, * p<0.05, ** p<0.01, *** p<0.001.
As shown in
To evaluate the kinetics of migration, localization and morphology of Gr1+ cells in the lungs of infected NOS2 -/- mice, immunohistochemical analysis was performed in tissues 45 days after infection. As shown in
Anti-Gr1 immunohistochemistry was performed to determine the location of Gr1+ cells surrounding necrotic lesions in the NOS2 -/- mice. As observed in low magnification (4x), anti-Gr1 labeling was predominantly localized in the periphery or rim of these structures (red, arrows, A). At higher magnification (100x), two distinct cellular morphologies could be observed for cells staining with anti-Gr1. In the left panel, neutrophils with multi-lobed nucleus could readily be seen. However, particularly in close apposition to the fibrotic capsule, cells were seen that had abundant and vacuolated cytoplasms as well as a unilobed nucleus, consistent with the morphology of MDSCs. Arginase activity was biochemically detected in the lungs of NOS2 -/- mice but minimally WT controls (B). As disease progressed in NOS2 -/-, more arginase activity could be detected compared to wild-type mice (B).Results are expressed as the mean values of arginase activity (± SEM, n=5) in the Lung. **Student t-test, p<0.01 and ***Student t-test, p<0.001. Anti-arginase-1 immunohistochemistry was performed to determine its expression by inflammatory cells in the lungs of NOS2-/- mice. Similar to the anti-Gr1 staining, at low magnification (4x), robust arginase-1 expression was detected in cells located in the rim of necrotic granulomas (C, left panel). At higher magnification (100x), arginase-1 expression was limited to highly vacuolated cells containing abundant cytoplasm (C, right panel). Panel D shows co-localized staining of both Gr1+ and arginase-1 in NOS2 -/- mouse lung tissues after 45 days of infection. At higher magnification (100x), co-localized staining of Gr1+ and arginase-1 demonstrates that Gr1+ cells (D lower right panel, arrow, red) and arginase-1 staining (D, arrow, brown) are clearly co-localized on the same cell.
Arginase-1 activity plays a pivotal role in MDSCs function [
Given these observations, we posed the question as to whether these observations were unique to the NOS2 -/- model, or were a common feature in models in which lung necrosis occurs. To address this question, we used immunocompetent C3HeB/FeJ mice and immunodeficient Rag2 -/- mice as additional possible examples, given the knowledge these both develop severe lung necrosis [
The influx of Gr1intCD11b+ cells was evaluated in immunocompromised RAG-/- (A), and immunocompetent C3HeB/FeJ mice (C) and NOS-/- (F). Again, large numbers of Gr1intCD11b+ cells were observed in these three mouse strains and significant differences were observed in the Gr1+ MFI (B, D, G). Similar to NOS2 -/- mice, arginase activity was also increased in C3HeB/FeJ undergoing lung necrosis (E). In contrast, C3H/HeOuJ did not have major arginase activity. Similar to NOS2 -/- mice, arginase activity was also increased in C3HeB/FeJ undergoing lung necrosis (E). Results are expressed as the mean values of mean fluorescence intensity (MFI) or arginase activity (± SEM, n=5) in the Lung. **Student t-test, p<0.001.
L-arginine is catabolized either by arginase or nitric oxide synthase [
In an attempt to further characterize Gr1int and Gr1hi cell function during tuberculosis infection we conducted cell sorting of Gr1int and Gr1hi cells from naïve and infected NOS-/- mice.
Specific populations of Gr1int or Gr1hi cells were sorted from infected NOS-/- mice in an attempt to further characterize Gr1int and Gr1hi cell function during tuberculosis using flow cytometry. Panel A, demonstrates pre-sorted Gr1+CD11b+ cells which were then further characterized. Gr1hi and Gr1int sorted cells (B, C) both demonstrated high expression of arginase I and IL-17. Panel D shows the isotype controls for arginase-1 and IL-17A.
We then evaluated if sorted Gr1+ cells obtained from
Specific populations of Gr1int or Gr1hi cells were sorted from infected NOS-/- mice and placed into
Studies were then conducted to evaluate if the sorted Gr1int or Gr1hi cells from NOS-/- mice were able to induce proliferation.
The results of this study provide further evidence that cells expressing intermediate levels of the Gr1 marker are a predominant subset of cells accumulating in the lungs of mice developing lung necrosis due to their inability to control infection with
We strongly suspect that neutrophils are the main cause of necrosis in both the mouse and guinea pig models of tuberculosis [
Whereas some of the Gr1+ cells are clearly neutrophils, an additional population of Gr1+ subsets, fall into the category of myeloid derived suppressor cells. These recognized as either a monocytic or granulocyte-like population that are implicated in the failure to control the growth of tumors [
MDSCs appear to be able to modulate T cell functions [
Even though most studies have reported that MDSCs suppress the immune response, particularly T cells, in some recent models it has emerged that MDSCs could actually activate and induce proliferation of T cells. Interestingly and in accordance with our results, in the experimental autoimmune encephalomyelitis (EAE) model it was reported that MDSCs activated Th17 cells. In turn, Th17 negatively affected the outcome of EAE by increasing pathology. Mechanistically, it was shown that Th17 induction in EAE was mediated by IL-1β. Furthermore, a complex relationship between TH17 cells producing IL-17 and MDSCs was recently unveiled in several cancer models. IL-17 was shown to increase the MDSCs-dependent induction of chronic inflammation by production of IL-1 in the tumor microenvironments. In our model, it could be envisioned that autocrine or paracrine IL-17 produced by Gr1+ or T cells, respectively could lead to inflammation and tissue damage, as well as increased bacterial burden. This uncontrolled tissue damage could also cause Gr1int cells directly suppress other T cell subsets and to recruit regulatory T cells. However, this would need further validation with either IL-17 or IL-17R knockout mice or alternatively, with anti-IL-17 therapy. Our studies denoted as the splenocyte to Gr1high or Gr1int ratio increased the percentage of CD4+ proliferation reduced. These results suggest that Gr1+ subsets are either capable of suppression under these circumstances or alternatively the induced proliferation present at lower ratios was diluted out. Finally, it was recently reported that the origin of MDSCs could impact their immunomodulatory function. Whereas MDSCs obtained from a chronic peritonitis model activated T cells, peritoneal tumors induced MDSCs with suppressive functions.
MDSC have been widely reported in human patients, mostly in the context of cancer, and in fact as new data accumulates there are probably subsets of MDSC themselves. The importance of neutrophils in the human immune response to tuberculosis is becoming more widely recognized, but what proportion of these cells is actually MDSC instead has not to our knowledge been investigated. Unfortunately, human neutrophils do not express Gr1, but other markers such as CD15, CD66b, CD115 and CD124 [
Six to 9 weeks old C57BL/6, NOS2 -/-, RAG2 -/-, C3HeB/FeJ and C3H/HeOuJ were purchased from the Jackson Laboratories (Bar Harbor, ME). This work was approved by the IACUC of Colorado State University. Animals were maintained in a BSL-3 facility at Colorado State University and had
At the indicated time points, mice were humanely euthanized via CO2 inhalation and organs harvested for CFU enumeration, histology, and flow cytometry analysis. Organs were homogenized in saline and serial dilutions were plated on 7H11 agar plates supplemented with OADC (BD Biosciences, San Jose, CA). After 3-4 weeks incubation at 37°C, CFUs were counted and the data expressed as the log10 numbers per target organ.
Cells were harvested from the lungs as described before [
Organs were perfused with 4% formaldehyde. Paraffin embedded tissues were stained with Hematoxylin and Eosin [
Paraffin embedded slides were processed and after de-waxing with Histo-Clear (National Diagnostics, Atlanta, GA) and decreasing concentrations of ethanol, antigen retrieval was performed in a pressure cooker with DakoCytomation Target Retrieval Solution (DakoCytomation, Carpinteria, CA). Endogenous peroxidases and alkaline phosphatases were inactivated for 10 min with BLOXALL™ (Vector Laboratories, Burlingame, CA) and slides were blocked with 2.5% normal goat serum (Vector Laboratories) for 1 hr. Thereafter, slides were incubated O/N at 4°C with 1/40 rat anti-Gr1 (clone RB6-8C5, eBioscience) or rat IgG2b (isotype control, clone eB149/1OH5 eBioscience) diluted in 2.5% goat serum. After washing slides with TBS, they were incubated for 1 hr at RT with 1/1200 alkaline phosphatase-labeled, goat anti-rat IgG (Santa Cruz Biotechnology, Dallas, TX) diluted in 2.5% normal goat serum. The reaction was developed for 20 min with Vector® Red Alkaline Phosphatase Substrate (Vector Laboratories) and slides were counterstained with Hematoxylin QS (Vector Laboratories) for 30 sec. For arginase-1 expression, slides were processed similarly except for minor modifications. In this case, slides were blocked with ProteinBlock (DakoCytomation). The primary antibody was sheep anti-arginase-1 IgG (R&D, Minneapolis, MN), the isotype control was sheep IgG (R&D) and the secondary antibody was peroxidase-labeled donkey anti-sheep/goat IgG (AbD Serotec). Finally, the reaction was developed for 10 min with Liquid DAB substrate (X Biogenex, Freemont, CA).
Lung lysates were centrifuged at 4°C for 10 min at 13,000 x g to remove debris and supernatants were quantified using the BCA assay (Pierce, Rockford, IL).
The level of arginase activity in lung lysates was performed using the QuantiChrom™ Arginase Assay Kit (BioAssays Systems, Hayward, CA). Briefly, protein concentration was normalized to100 µg/ml. As a control, an aliquot of normalized protein was heat-inactivated at 90°C for 10 min. Forty µl of untreated or heat-inactivated sample was incubated at 37°C for 60 min with 10 µl of 5X reagent. Thereafter, the reaction was processed as recommended by the manufacturer and analyzed in a plate reader (BioRad, Hercules, CA) at an absorbance of 430 nm. The absorbance of the heat-inactivated sample was subtracted from the respective untreated sample and then compared to urea standards. Activity is expressed as mU/mg of protein.
To sort Gr1int and Grhigh, single cell suspensions were obtained from lungs as described above. After incubating with 5 µg/ml of Fc block (eBioscience) for 20 min at 4°C, cells were stained with PeCy7-labeled anti-CD11b (clone M1/70, eBioscience) and Alexa 700 labeled anti-Gr1 (clone RB6-8C5, eBioscience) as described above. Thereafter, cells were stained with 0.5 µg/ml 7-AAD (Invitrogen) for 5 min and sorted using a FACS Aria III (BD Biosciences) using the following strategy: doublets were gated out by FSC-A vs FSC-H and 7-AAD+ dead cells were excluded from the singlets population using FL3. Live singlets expressing CD11b+ were then analyzed based on FSC-A vs Gr1, in order to sort the two populations of Gr1high and Gr1int. Sorted cells were analyzed for purity in a FACSAria. Additionally, cytokine production and arginase-1 expression was evaluated for both populations of freshly sorted Gr1+ cells, as follows: cells were incubated for 1 h at 37°C in a CO2 incubator with 2 µM monensin and 3 µg/ml brefeldin (eBioscience). After fixation and permeabilization using Fix/Perm and 1x permeabilization buffer, respectively (eBioscience), intracellular staining was performed with Alexa488-labeled anti-IL17A and F (clones TC11-18H10.1 and 9De.1C8, Biolegend) or FITC-labeled anti-arginase (R&D, same clone as for immunohistochemistyr), APC-labeled anti-IL-10 (clone JES5-16E3, eBioscience) or Alexa647-labeled anti-IL-12 (clone C17.8, eBioscience). As controls, cells were also labeled with the respective isotype control.
Cytokines and activation markers: T cells were obtained from naïve or
T cells were obtained as described above and stained with CFSE (Invitrogen), following their protocol. Briefly, T cells were stained with 10 µM CFSE for 10 min at 37°C and after quenching the reaction with 3 volumes of medium, cells were washed 3 times with PBS. Stained T cells were incubated with increasing ratios of T cells per Gr1high or Gr1int cell for 7 days at 37°C in a CO2 incubator. Cells were then surfaced stained with PerCP-labeled anti-CD4 (same clone as above), Pe-Cy7-labeled anti-CD8 (same clone as above) and APC-labeled anti-CD3 e (clone 145-2C11, eBioscience) and analyzed using an LSR-II after gating on CD3+ CD4+ or CD3+CD8+.Proliferation was evaluated as a decrease in FL1 fluorescence over the basal conditions (in the absence of Gr1+ cells).
Data are presented using the mean values from 5 mice per group and from values from replicate samples and duplicate or triplicate assays. Two-tailed, unpaired Student t-test or ANOVA was performed using GraphPad Prism 4 (GraphPad software, San Diego, CA). Significance was considered below p<0.05.