Joint Senior Authorship: Paul N Reynolds and Sandra Hodge contributed equally to this work.
Altered sputum granzyme B and granzyme B/proteinase inhibitor-9 in patients with non-eosinophilic asthma
Version of Record online: 23 DEC 2013
© 2013 The Authors. Respirology © 2013 Asian Pacific Society of Respirology
Volume 19, Issue 2, pages 280–287, February 2014
How to Cite
Simpson, J. L., Gibson, P. G., Yang, I. A., Upham, J., James, A., Reynolds, P. N., Hodge, S. and AMAZES Study Research Group (2014), Altered sputum granzyme B and granzyme B/proteinase inhibitor-9 in patients with non-eosinophilic asthma. Respirology, 19: 280–287. doi: 10.1111/resp.12213
(Associate Editor: Claire Wainwright).
- Issue online: 14 JAN 2014
- Version of Record online: 23 DEC 2013
- Manuscript Revised: 4 OCT 2013
- Manuscript Accepted: 4 OCT 2013
- Manuscript Revised: 30 AUG 2013
- Manuscript Received: 2 JUL 2013
- NHMRC. Grant Number: 569246
- granzyme B;
- non-eosinophilic asthma;
- proteinase inhibitor-9
Background and objective
The non-eosinophilic phenotype of asthma (NEA) is associated with chronic airway inflammation and airway neutrophilia. An accumulation of apoptotic airway epithelial cells, if not efficiently cleared by efferocytosis, can undergo secondary necrosis, with the potential for inflammation of surrounding tissues. Apoptosis may occur via the T cell granzyme B pathway. The role of granzyme B in NEA is not known. The aim of this study was to investigate production of granzyme B and its inhibitor proteinase inhibitor (PI)-9 by T cells from induced sputum and compare expression between eosinophilic, NEA and healthy controls.
We investigated T cell intracellular granzyme B and its inhibitor, PI-9, in sputum from healthy control subjects (n = 10), and patients with NEA (n = 22) or eosinophilic asthma (EA) (n = 15) using flow cytometry.
Granzyme B expression and the ratio of granzyme B to PI-9 positive cells were highest in those with NEA for both CD3+ and CD4+ T cells. The expression of granzyme B was not statistically different between patients with NEA and EA; however, the ratio of granzyme B to PI-9 positive cells for CD3+ T cells was significantly higher in those with NEA compared with EA.
Induced sputum provides a non-invasive tool for investigating T cell cytotoxic mediators in the various asthma subtypes. Granzyme B expression is increased in NEA and the contribution of granzyme B to chronic inflammation requires further study.
chronic obstructive pulmonary disease
non-eosinophilic phenotype of asthma
natural killer cells
protease inhibitor 9
Non-eosinophilic phenotype of asthma (NEA) is a distinct inflammatory subtype of asthma, with relevance to disease mechanisms and treatment responsiveness.[1, 2] In fact, eosinophilic inflammation is absent in up to 50% of all asthma patients, and the absence of eosinophils in NEA has been confirmed in bronchial tissue by both endobronchial biopsy[3, 4] and post-mortem examination. Like patients with chronic obstructive pulmonary disease (COPD), patients with NEA often display chronic airway inflammation associated with significant airway neutrophilia and are poorly responsive to inhaled corticosteroids.[6, 7] Importantly, patients with persistent inflammatory asthma experience more non-eosinophilic exacerbations than eosinophilic exacerbations; these exacerbations are not prevented by corticosteroid treatment. Therefore, new therapeutic options are needed, based on a greater understanding of the pathogenesis of this form of asthma.
Increasing evidence implicates perturbations in programmed cell death (apoptosis) in a variety of chronic lung diseases including asthma, COPD, cystic fibrosis and bronchiolitis obliterans syndrome following lung transplantation.[8-14] We have previously shown that increased numbers of apoptotic airway epithelial cells in COPD are associated with airway epithelial apoptosis and secondary necrosis. In severe asthma, a condition associated with increased airway neutrophilia, tissue damage and increased rates of apoptosis of airway epithelial cells or smooth muscle cells have also been demonstrated.[9-11] The potential stimuli for increased apoptosis in asthma have not been identified.
Cytotoxic T cells and natural killer (NK) cells can induce apoptosis of target cells including bronchial epithelial cells via mechanisms that include the granzyme-mediated pathway. Granzyme B and perforin are stored in cytoplasmic secretory granules of cytotoxic cells, and released into the intercellular space following adhesion to the target cell. In the presence of Ca2+, perforin pores in the cell membrane enables entry of granzyme B and induction of caspase-dependent apoptosis. It is possible that intracellular levels of granzyme B may reflect cytotoxic T cell activity in the airways in NEA. The main inhibitor of granzyme B is proteinase inhibitor-9 (PI-9), a 42 kDa intracellular protein belonging to the serpin superfamily.[16, 17] PI-9 irreversibly binds to granzyme B in vitro and can abrogate cytotoxicity and is considered as a self-protecting mechanism against the apoptotic effects of endogenous granzyme B.[16, 18, 19]
We thus hypothesized that increased T cell granzyme B and PI-9 would be detected in the airways in induced sputum from patients with NEA, potentially contributing to increased apoptosis of bronchial epithelial cells and associated tissue damage. We assessed intracellular granzyme B and PI-9, in sputum from well-characterized people with asthma with eosinophilic asthma (EA) or NEA and healthy controls.
Granzyme B and PI-9 was investigated in sputum-derived T cells from subjects who were either non-smoking controls recruited by advertisement, with no history of cancer, respiratory or allergic disease and normal spirometry (n = 10) or participants with stable asthma (15 with eosinophilic and 22 with non-eosinophilic). All participants with asthma had a physician diagnosis of asthma, current asthma symptoms, demonstrated evidence of variable airflow obstruction, were taking inhaled corticosteroids and no patients were receiving oral corticosteroids. Subjects underwent a clinical assessment that included a smoking and allergy history, and a hypertonic saline sputum induction. All samples were collected from participants with stable asthma without a recent (past 4 weeks) viral or bacterial infection. Any participant who had reported a recent or current infection was delayed before their visit was conducted.
Subject demographics are presented in Table 1. Ethical approval was granted by both the Royal Adelaide Hospital Ethics Committee and the Hunter New England Human Research Ethics Committee. Written informed consent was obtained for each subject recruited for the study. The diagnosis of asthma was confirmed by the presence of symptoms with airways hyperresponsiveness to hypertonic saline or a clinically significant bronchodilator response (>12% improvement in expiratory volume in 1 s). EA was defined using a cutpoint of >3% eosinophils in induced sputum.
|FEV1 (% predicted)||1014||704*||726*||<0.001|
|Pack years, median (q1, q3)||0.0 (0.0, 1.5)||0.0 (0.0, 15.4)||0.0 (0.0, 2.3)||0.658|
KCO % predicted, mean (SD)
If ex-smoker with more than 1 10 year smoking history
|83.5 (3.9)||85.4 (7.6)||0.539|
|Total cell count (106/mL)|| |
3.6 (2.3, 5.0)
n = 4
5.8 (3.8, 8.3)
n = 12
6.5 (4.1, 7.6)
n = 19
|Neutrophils %|| |
38.5 (32.2, 51.5)
n = 4
43.4 (22.8, 76.3)
n = 14
43.8 (33.8, 52.0)
n = 22
|Neutrophils, × 104/mL||134.1 (86.81, 235.8)||212.1 (54.81, 720.5)||212.1 (139.9, 301.7)||0.717|
|Eosinophils %||0.8 (0.4, 2.1)||9.3 (5.8, 21.3)*||0.9 (0.3, 1.5)||<0.001|
|Eosinophils, × 104/mL||3.15 (0.88, 10.49)||4.16 (1.49, 10.95)*||0.82 (0.00, 4.20)||<0.001|
|Lymphocytes %||0.65 (0.50, 1.15)||0.50 (0.00, 1.00)||0.13 (0.00, 0.75)||0.267|
|Lymphocyte (104/mL)||3.0 (1.9, 3.7)||0.8 (0, 4.1)||2.1 (0, 10.9)||0.330|
|CD4+ cells %||79.0 (71.4, 91.8)||70.2 (63.5, 82.4)||61.7 (55.2, 77.0)*||0.032|
|CD8+ cells %||21.0 (8.2, 28.6)||29.8 (17.6, 36.5)||38.3 (23.1, 44.8)*||0.032|
Spirometry (KoKo PD Instrumentation, Louisville, CO, USA) and sputum induction with hypertonic saline (4.5%) were performed as previously described. A fixed sputum induction time of 15.5 min was used for all subjects to reduce any bias due to differing induction times.
Processing of induced sputum
Samples were collected from two sites and all testings performed at 18–24 h following collection. Induced sputum was processed as we have previously reported. Briefly, a minimum of 100 μL of sputum selected from saliva was dispersed using dithiothreitol, prepared for differential cell counts and cells re-suspended in RPMI1640 (with 1% FCS, 0.5% 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, 2% Penicillin-Streptomycin, 1% Fungizone (Amphotericin B)) as previously reported. When the sputum sample had sufficient volume a total cell count, cell viability (35 of 47 samples) and sputum differential cell count (40 of 47 samples) were also performed.
To assess intra-subject variability in sputum parameters in samples collected at different time points, repeat sputum was collected from 12 of the subjects (six NEA and six EA) 2 weeks following the first collection. Granzyme B and PI-9 were compared. Repeatability was assessed by calculation of bias from mean difference Bland–Altman plots.
The following monoclonal antibodies and immunological reagents were employed: CD3[PC5] (Immunotech/Coulter, Marseille, France), and CD8 [FITC] (BD Biosciences, San Jose, CA, USA), granzyme B [PE] (Serotec, Oxford, UK), rabbit anti-mouse PE, FACSPerm, FACSLyse and IsoFlow (BD Biosciences, North Ryde, NSW, Australia). PI-9-7D8 was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA).
Intracellular staining with monoclonal antibodies to granzyme B and PI-9
Granzyme B and PI-9 expression was determined as previously described. Briefly 200 μL of cells were added to FACS tubes. Cell membranes were permeabilized with FACSPerm, cells were then stained with directly conjugated monoclonal antibody to granzyme B or monoclonal antibody to PI-9 7D8 followed by a secondary antibody (rabbit anti-mouse PE). Data were acquired using a FACSCalibur flow cytometer (BD Biosciences) and analysed using CellQuest software (BD Biosciences). T cells were identified based on CD3 PC-5 versus side-scatter and CD4+ and CD8+ T cells enumerated using a monoclonal antibody to CD8. Negative controls were applied using fluorescence thresholds set so that <1% of lymphocytes showed non-specific antibody binding. Due to limitations of cell numbers, our gating technique did not exclude the small population of NKT-like cells (CD3+CD56+).
The Kruskal–Wallis and Mann–Whitney U-tests were applied to analyse the non-normally distributed data. Correlations were performed using Spearman's rank correlation. Intra-subject variability was assessed by calculation of bias from mean difference Bland–Altman plots. Analysis was performed using Stata 11 (College Station, TX, USA). P-values <0.05 were considered significant.
Identification and differential counting of sputum
There was no significant difference in total leukocyte or lymphocyte counts in sputum from any patient group versus controls (Table 1). As expected, sputum eosinophil proportions were significantly higher in patients with EA compared with those with NEA and healthy controls. Similar neutrophil proportions were observed between NEA and EA groups.
CD4+ and CD8+ T cells
The proportion of CD4+ T cells was significantly reduced while the proportion of CD8+ T cells was significantly increased in induced sputum from patients with NEA versus healthy controls (Table 1).
Intracellular staining of granzyme B and PI-9
The analysis approach and examples of flow cytometric dot plots of granzyme B and PI-9 staining has been previously published (for bronchoalveolar and lung tissue T cells).[23-25]
The percentage of sputum derived CD3+ and CD4+ T cells expressing intracellular granzyme B was significantly higher in NEA than controls (Fig. 1a,b). The median granzyme B expression on CD3+ and CD4+ T cells was highest in NEA compared with EA (CD3+ median NEA 18.50 vs median EA 11.91; CD4+ median NEA 9.81 vs median EA 4.49), although these results did not reach statistical significance in post-hoc testing.
There were no changes in expression of PI-9 in any T cell subsets from any patient group compared with controls.
Whenever smokers were compared with ex-smokers, there was no difference in the granzyme B or PI-9 expression in any T cell type (data not shown).
Correlation between granzyme B and PI-9, and smoking pack year, age and disease severity
There were no significant correlations between T cell granzyme B in induced sputum and age or pack years (Spearman rank correlation P > 0.5, data not shown).
Sputum PI-9 expression on CD8+ T cells was significantly negatively correlated to expiratory volume in 1 s/forced vital capacity ratio (Spearman rho −0.38; P = 0.024).
We observed a significant positive correlation between granzyme B and PI-9 on CD3+ T cells (Spearman r = 0.517; P = 0.009), and a weaker correlation on CD4+ T cells (Spearman r = 0.316; P = 0.053). There was no association between granzyme B and PI-9 on CD8+ T cells (Spearman r = 0.107; P = 0.522).
Granzyme B and PI-9 ratio
The ratio of granzyme B/PI-9 was increased for CD3+ T cells from patients with NEA compared with those with EA but not different to healthy controls (Fig. 2a). In addition, the ratio of granzyme B/PI-9 was increased in CD4+ T cells from patients with NEA compared with healthy controls but not significantly different compared with those with EA (Fig. 2b). There was no difference in granzyme B/PI-9 ratio in CD8+ T cells between groups (Fig. 2c).
The intraclass correlation coefficient was 0.92, 0.90 and 0.83 respectively for granzyme B expression on CD 3+, 4+ and 8+ T cells indicating substantial agreement in the measurements. For PI-9 expression, the intraclass correlation coefficient was 0.86, 0.57 and 0.63 for CD 3+, 4+ and 8+ T cells respectively, which suggests a moderate to substantial agreement in the measurement.
The bias was small with equal scatter for the proportion of cells staining positive for granzyme B irrespective of cell type; the bias was largest for CD4+ T cells staining positive for PI-9, and there tended to be a larger difference at higher average %CD4+PI-9+ (Fig. 3).
Granzyme B can induce apoptosis in target epithelial cells by processing and activating members of the caspase family. In this study, we found elevated levels of granzyme B in NEA, consistent with a role leading to increased apoptosis in NEA. We have recently also shown that alveolar macrophages from patients with NEA also exhibit a reduced phagocytic function that is similar to that found in COPD. The dual changes of increased granzyme B and impaired efferocytosis may lead to persistent inflammation in NEA.
Our previous methods for assessing T cell phenotype and function in the airway have relied largely on investigating cells obtained from flexible bronchoscopy.[27-29] This method has proved to be reliable and has produced several key findings with regard to the pathogenesis of COPD; however, it is relatively invasive and not suited to large scale studies of asthma pathophysiology and treatments. In this study, we assessed the utility of measuring granzyme B and PI-9 in T cells obtained from induced sputum from people with asthma as a non-invasive surrogate for bronchoalveolar lavage. The technique was reliable with acceptable reproducibility.
Using induced sputum cells, we showed that intracellular granzyme B was highest on both CD3+ and CD4+ cells subjects with NEA and significantly elevated versus healthy controls. Levels of granzyme B (or PI-9) were not affected by asthma severity, smoking history or age. Although not significantly different, granzyme B expression was higher in those with NEA compared with EA, and these differences could account for differences in asthma pathogenesis.
Since the percentage of granzyme positive cells was higher for CD8+ than CD4+ cells in all groups, the additional increased proportion of CD8+ cells in NEA (as we have found in COPD) could potentially contribute to greater damage. The lower proportion of CD4+ cells may be less important, even with a higher percentage of granzyme B positive staining as these two findings may negate each other. Future studies are needed investigate the relative cytotoxic potential of the granzyme B+ cells.
Due to the limited cell numbers available, we did not measure perforin (an important mediator of granzyme-targeted apoptosis) or granzyme B production from NK and NK like T cells. A previous study showed that CD8+ T cells from induced sputum from COPD patients expressed increased levels of perforin and had higher cytotoxic activity; and it is possible that similar mechanisms are at play in NEA. We and others have shown have enhanced cytotoxic function with increased production of granzyme B in NK and NK-like cells in the airway in COPD.[31, 32]
The association of the granzyme B inhibitor, PI-9, with the surface of the cytotoxic granule allows it to deactivate granzyme B as soon as it leaks out of the granule. This mechanism is thought to protect cells from damage from their own misguided granzyme B.[17, 19, 33] It has been shown that during in vitro stimulation of cytotoxic lymphocytes, both granzyme B and PI-9 production were upregulated in the same cell with a direct relationship between the two mediators. We observed that patients with NEA had significantly higher granzyme B/PI-9 ratio on both CD3+ and CD4+ T cells compared with the other groups.
To our knowledge, this is the first comprehensive report of T cell granzyme B expression in induced sputum obtained from patients with well-categorized asthma. A few studies have reported increased levels of granzyme B in airway T cells in other chronic inflammatory lung diseases including septic acute respiratory distress syndrome. A further study investigated granzyme B in bronchoalveolar lavage from patients with asthma following segmental allergen challenge and found an increase in granzyme B+ lymphocytes and soluble granzyme B in bronchoalveolar lavage of allergen-challenged patients with atopic asthma. Granzyme B expression in induced sputum has been less studied, although one study reported increased intracellular perforin in sputum CD8+ lymphocytes in COPD.
The precise role of cytotoxic T cell damage in NEA has not been fully elucidated, in particular the cytotoxic T cell-mediated cytolysis of airway epithelial cells. Importantly, in our previous COPD studies using bronchoalveolar lavage, a significant correlation was found between intracellular granzyme B in the airway and bronchial epithelial apoptosis. As it has been convincingly demonstrated that uncleared apoptotic material may then undergo secondary necrosis with the potential for inflammation of surrounding tissues,[8, 12] our finding of increased granzyme expression in NEA suggest that upregulation of the granzyme apoptotic pathway may at least partially explain the prevalence of chronic inflammation and defective repair processes that are the hallmark of NEA pathogenesis.
While the role of CD4+ T cells has been well described in asthma, the relative expression of CD4+ and CD8+ T cells in induced sputum from patients with NEA has not been characterized. We found the highest proportion of CD8+ cells in NEA, similarly increased CD8+ cells have been reported in occupational asthma where neutrophils and not eosinophils are the dominant granulocyte. A crucial cytotoxic role of CD4+ T cells has been shown in viral disease, coronary artery disease, cardiac allograft rejection and renal disease.[37-41] It is thought that these cells may be CD28(null) T cells that expand and have deleterious effects including their ability to infiltrate into tissues and cause tissue damage. Interestingly, in COPD, we have recently shown an increase in the percentage of CD28(null) T cells in the airway in COPD subjects, and that these cells are potent producers of pro-inflammatory mediators including tumour necrosis factor-α and granzyme B. Given the similarities between NEA and COPD highlighted by our previous study, further investigation of this subset in NEA is warranted. We are also currently investigating NKT-like T cells in our asthma groups, as they are a major source of pro-inflammatory cytokines and cytotoxic mediators (granzymes) and numbers of these cells are increased in the airway in COPD/emphysema, accompanied by a reduction in expression of the inhibitory receptor CD94 (Kp43) and increased cytotoxic potential.[31, 43] We further showed that production of pro-inflammatory cytokines and granzyme B by these cells is poorly suppressed by corticosteroids, indicating their potential importance particularly in the non-eosinophilic form of asthma.
In conclusion, we have shown that induced sputum is a suitable, non-invasive surrogate for bronchoalveolar lavage for investigating T cell expression of granzyme B and PI-9. Our data suggest that the granzyme B pathway is upregulated in T cells obtained from induced sputum from patients with NEA to a similar extent as we have previously found in COPD, raising the possibility that granzyme-mediated apoptosis may be one mechanism of lung injury in asthma.
The authors gratefully acknowledge the technical assistance of Jessica Ahern, Sarah Richards, Kirtsy Herewane, Calida Moller, Gabrielle LeBrocq, Kelly Steel, Kellie Fakes, Michelle Gleeson and Bridgette Ridewood. The work was supported by an NHMRC Project Grant ID: 569246.
- 20Epidemiological association of airway inflammation with asthma symptoms and airway hyperresponsiveness in childhood. Am. J. Respir. Crit. Care Med. 1998; 158: 136–145., , , , , , .