Bronchiolitis obliterans syndrome (BOS) is characterized by persistent alloreactive, infective and non-specific epithelial injury, loss of epithelial integrity and dysregulated repair. We have reported increased apoptosis of epithelial cells collected from the large airway in lung transplant recipients. As part of the alloreactive response, T cells induce apoptosis of target epithelial cells by secreting granzyme b. We hypothesized that granzyme b would be increased in lung transplant patients with acute rejection and BOS and that commonly used immunosuppressive agents would fail to suppress this serine protease adequately. We investigated intracellular T cell granzyme b in blood, bronchoalveolar lavage (BAL) and large airway brushing (23 controls, 29 stable transplant, 23 BOS, 28 acute rejection, 31 infection) using flow cytometry and assessed the effect of clinically relevant concentrations of cyclosporin A, tacrolimus, methylprednisolone and a protease inhibitor, gabexate mesilate, on in vitro granzyme b production. Granzyme b was increased significantly in all compartments of all transplant groups compared to controls. Surprisingly, granzyme b was even higher in patients with BOS than in patients with acute rejection. In longitudinal analysis in three patients, blood granzyme b increased prior to or at the onset of BOS. In vitro, methylprednisolone and gabexate mesilate had no effect and cyclosporin A and tacrolimus only a moderate effect on production of granzyme b by CD8+ T cells. Increased T cell granzyme b production may contribute to BOS pathogenesis and is not curtailed by current immunosuppressants. Longitudinal investigation of granzyme b in blood may provide an adjunctive non-invasive method for predicting BOS/OB.
Long-term survival following human lung transplantation is limited by progressive deterioration in lung function due to obliterative bronchiolitis (OB) and its clinical manifestation, bronchiolitis obliterans syndrome (BOS). OB is an essentially irreversible process, characterized by persistent alloreactive, infective and non-specific epithelial injury and dysregulated repair. These processes culminate in remodelling and fibrotic obstruction of small airways, leading directly to patient morbidity and mortality [1,2]. There has been no substantial improvement in the reported incidence of OB over the last 10 years, despite improvements in immunosuppression and patient management [1,2]. The strongest risk factor for OB is acute rejection, with repeated episodes of even minimal acute rejection increasing the risk of subsequent OB significantly . This association suggests that immune mechanisms may in part underlie OB and have thus led to the hypothesis that the disorder is a form of chronic rejection.
As part of the alloreactive response T cells secrete granzyme b, which can induce apoptosis in target epithelial cells by processing and activating members of the caspase family. Elevated levels of granzyme b have been noted in acute rejection of kidney and heart allografts and granzyme b mRNA expression has been shown previously to correlate with acute lung transplant rejection [4–6]. Further, our group has shown previously an increase in apoptosis of airway epithelial cells in transplanted lungs , and a significant correlation between the percentage of T cells expressing granzyme b and apoptosis of bronchial epithelial cells in the airway . Granzymes may also be released extracellularly, particularly during cytotoxic T cell degranulation , and it has been reported that the secreted granzyme b can contribute to tissue destruction by degrading various extracellular proteins .
The aim of the present study was to evaluate the utility of measuring granzyme b as a predictor of BOS/OB following lung transplantation.
Intracellular granzyme b was measured in T cells from large airway brushing, peripheral blood and bronchoalveolar lavage (BAL), collected from stable transplant patients and those with evidence of acute rejection, BOS or infection. Released granzyme b was measured in BAL.
The most effective transplantation immunosuppressive strategies are based on interruption of interleukin (IL)-2 signalling by calcineurin inhibitors, cyclosporin A and tacrolimus. However, intensification of immunosuppressive therapies has not led to any consistent improvement in BOS [1,2]. We therefore also investigated the ability of these commonly used immunosuppressive agents, and a novel protease inhibitor, gabexate mesilate, to inhibit T cell granzyme b production in vitro.
Materials and methods
CD45 and CD3 [PC-5] (Immunotech/Coulter, Marseille, France) were used to identify leucocyte contamination and identify T cells in BAL and airway brushing. The following were also employed: CD8 [fluorescein isothiocyanate (FITC)] (BD Biosciences, San Jose, CA, USA), granzyme b [phycoerythrin (PE)] (Serotec, Oxford, UK), IgG1/IgG1 negative control (FITC/PE) (BD Biosciences), red blood cell lysing agent [fluorescence activated cell sorter (FACS)lyse], cell membrane permeabilizing agent [FACSperm] (BD Biosciences), methylprednisolone (David Bull Laboratories, Melbourne, Australia), cyclosporin A (Novartis Pharmaceuticals, North Ryde, NSW, Australia), tacrolimus (Janssen-Cilag, North Ryde, NSW, Australia) and gabexate mesilate (Sigma, St Louis, MO, USA).
Currently, our lung transplant patients undergo frequent bronchoscopic evaluation, either in the context of surveillance to detect preclinical histological signs of rejection or in the context of acute deteriorations, where samples are required for a specific diagnosis. These routine procedures have provided an ideal opportunity to evaluate the role of key mediators in both stable subjects (including longitudinally in the same subject) and those with BOS, and to assess less invasive approaches to diagnosis and surveillance. Transplant subjects undergoing routine surveillance bronchoscopy (according to international guidelines) as described previously  were recruited from three centres: the Royal Adelaide Hospital, South Australia, the Royal Perth Hospital, Western Australia and the Prince Charles Hospital, Brisbane, Australia. Patient rejection status was categorized both clinically and histologically on transbronchial biopsies according to standard criteria . Histological assessment was performed by pathologists who were blinded to the results of flow cytometric testing. A further group of healthy never-smoker volunteers with no history of lung disease were recruited as controls. Ethics approval was obtained from all institutions and informed consent obtained. Patient demographic details are presented in Table 1. Patients were maintained on standard immunosuppressive therapy (cyclosporin or tacrolimus, mycophenolate or azathioprine and prednisolone). Treatment for acute rejection consisted of therapy with prednisolone (10 mg/kg daily for 3 days followed by a PO taper). BAL, large airway brushing and peripheral blood was obtained from 23 healthy controls, 29 stable transplant subjects, 23 patients with evidence of BOS, 28 with acute rejection and 31 patients with proven infection. All subjects underwent spirometry as part of their routine clinical assessment.
Table 1. Demographic characteristics of the population studied.
FEV1 (% pred)
FVC (% pred)
Time post-transplant (months)
Data presented as mean ± standard error of the mean. BOS, bronchiolitis obliterans syndrome; FEV1, forced expiratory volume in 1 s; n.a., not applicable.
49·7 ± 18
97·7 ± 14
101·4 ± 13
50·8 ± 11
77·9 ± 23
79·9 ± 18
11 ± 19
54·3 ± 10
57·7 ± 21
64·3 ± 17
20 ± 21
Transplant: acute rejection
49·9 ± 11
71·9 ± 29
70·4 ± 24
10 ± 8
42·5 ± 13
69·6 ± 22
71·3 ± 19
15 ± 11
Bronchoscopy procedure and preparation of samples
BAL and large airway brushings (to obtain intraepithelial T cells) was obtained via flexible bronchoscopy as described previously [7,8]. Samples were processed within 20 h. We have shown previously that granzyme b levels are unchanged during this time period (unpublished observations). For each collection from an individual patient the first BAL aliquot was processed for microbiology and the second and third aliquots were pooled, kept on ice and processed within 24 h of collection. BAL-derived cells were washed and resuspended in RPMI-1640 + 10% fetal calf serum (GIBCO, BRL, Germany) and 1% weight per volume penicillin/streptomycin (GIBCO) (culture medium) as described previously [7,8] at a concentration of 4 × 105/ml. Total cell counts in BAL were performed using a modified Neubauer haemocytometer.
Venous blood was collected into 10 U/ml preservative free sodium heparin (DBL, Sydney, Australia). Blood films were stained by May–Grunwald–Giemsa using an automated staining machine (Shandon Veristat Southern Products, Astmore, UK) and differential cell counts were performed using a CELL DYN 4000 (Abbott Diagnostics, Sydney, Australia).
Detection of intracellular granzyme b by flow cytometry
Two hundred ml aliquots of prepared BAL or large airway brushing were added to FACS tubes. For blood, FACSlyse was added for 10 min. All cells were washed with a buffer containing 0·5% BSA in Isoflow, centrifuged at 500 g for 90 s, and the supernatant discarded. Cell membranes were permeabilized with FACSperm for 10 min, then washed. Cells were then incubated for 20 min with directly conjugated monoclonal antibody (mAb) to granzyme b. Data were acquired using a FACScalibur flow cytometer and analysed using CellQuest software. Representative plots are presented in Fig. 1.
Enzyme-linked immunosorbent assay (ELISA) measurement of released granzyme b
BAL and peripheral blood were prepared as described above from 10 patients in each group. BAL supernatant and plasma were frozen at −80°C. Thawed samples were activated through the addition at 1:5 of 0·1 ml 1 N HCl and neutralized with 1:5 1·2 N NaOH/0·5 M HEPES. BAL supernatants were then concentrated using Amicon Ultra-4 10 kDa centrifugal filter devices (Millipore, Billerica, MA, USA). Granzyme b levels were quantified by ELISA (Diaclone, Besançon, France), according to the manufacturer's instructions. The range for detection using this kit is 31·25–1000 pg/ml. A quality control sample of BAL from one of the lung transplant patients was run on each ELISA plate with the interassay coefficient of variation for the plates remaining within ± 17%.
The effects of immunosuppressive agents on production of granzyme b
To evaluate the effects of commonly applied immunosuppressive agents on granzyme b production, a novel in vitro assay was utilized. Mononuclear cells were isolated from peripheral blood of healthy human donors using Lymphoprep (Axis-Shield, Oslo, Norway), according to the manufacturer's instructions. The mononuclear cell layer was washed in culture medium at a concentration of 4 × 105 cells/ml. Mononuclear cells were incubated with gabexate mesilate (0·1–250 mg/ml) or various physiologically relevant concentrations of 10−6 M methylprednisolone, 5 ng/ml cyclosporin A or 25 ng/ml tacrolimus (diluted with RPMI-1640), or RPMI-1640 as a control, for 24 h at 37°C in an atmospheric pressure of 5% CO2. Cells were then stimulated for a further 48 h with Dynabeads CD3/CD28 T cell expander (Dynal Biotech, Oslo, Norway). The pellet containing mononuclear cells was then resuspended and flow cytometry staining for granzyme b performed immediately. Apoptosis was assessed using 7-aminoactinomycin D (7AAD) as described previously .
Statistical analysis was performed using the non-parametric Kruskall–Wallis, Mann–Whitney and Spearman's correlation tests. spss software and statistically significant differences between groups of P < 0·05 were applied.
Intracellular granzyme b
The percentage of CD3+ T cells expressing granzyme b was increased in blood, BAL and large airway (intraepithelial compartment) of all transplant groups compared to healthy controls, although statistically significant only in the BOS group (Fig. 2a). The increase was significant in CD8+ T cells from the BOS group and in CD4+ T cells from all transplant groups (see Fig. 2b). Granzyme b expression by CD3+ T cells was increased significantly in blood, intraepithelial compartment and BAL in patients with BOS versus those with stable graft function, acute rejection or infection (all P < 0·031) (Fig. 2a). Similar differences in granzyme b expression were noted for CD4+ and CD8+ T cells, although the differences between BOS and acute rejection or infection did not reach statistical significance (blood data presented in Fig. 2b).
Soluble granzyme b
Levels of soluble granzyme b were below the limits of detection in plasma. BAL from patients with BOS or infection contained more soluble granzyme than BAL from patients with stable graft function or acute rejection (BOS: mean ∼900 ± 102; infected mean 664 ± 210; stable; 157 ± 102; acute rejection 349 ± 197). The differences did not reach statistical significance (Fig. 2c).
Longitudinal analyses of intracellular and soluble granzyme b
For one patient, blood levels of intracellular granzyme b were consistently high over a 12-month period prior to diagnosis of BOS by histopathological examination of biopsy (Fig. 3a). A further two patients demonstrated increased production of granzyme b by blood T cells coincident with a decrease in lung function and diagnosis of BOS (Fig. 3b and c).
The effects of immunosuppressive agents on production of granzyme b
The production of granzyme b was investigated in T cells exposed to physiologically relevant concentrations of methylprednisolone, tacrolimus, cyclosporin A and gabexate mesilate in an in vitro model of T cell activation. Fewer than 5% of unstimulated T cells produced granzyme b. The effectiveness of the model was confirmed by a significant increase in granzyme b expression following stimulation with CD3/CD28 T cell expander compared to unstimulated cells (Fig. 4). None of the agents tested reduced granzyme b production by more than 48% over the experimental time-frame. Cyclosporin A (5 ng/ml) and tacrolimus (25 ng/ml) were the most effective for reducing granzyme b production, and this reduction was significant for both CD4+ and CD8+ T cells (Fig. 4). Methylprednisolone (10−6 M) effectively reduced overall T cell production of granzyme b; however, this drug had no significant effect on granzyme b production by CD8+ T cells. Dose–response experiments were not possible for methylprednisolone or cyclosporin due to increased non-specific cytotoxicity of T cells with increasing concentration as determined with 7AAD staining (data not shown). Dose–response experiments showed that tacrolimus exerted a dose-dependent reduction in granzyme b production (the maximum dose tested was 25 ng/ml, which is slightly higher than the therapeutic range of 5–20 ng/ml) (Fig. 5).
Gabexate mesilate, a protease inhibitor, decreased granzyme b production by both CD4+ and CD8+ T cells in a dose-dependent manner (Fig. 6). The decrease was significant for CD4+ T cells at a concentration of 25 mg/ml, and for CD4+ and CD8+ T cells at higher concentrations of 50 and 250 mg/ml. High levels of non-specific T cell cytotoxicity (>40% 7AAD positive apoptotic cells) were noted in the presence of 250 mg/ml gabexate mesilate.
Lung transplantation is being accepted and utilized increasingly as a therapy for end-stage lung disease. Chronic graft failure, due mainly to BOS and its pathological correlate OB, is high, and the 5-year survival following transplant is less than 50% [1,2]. BOS is characterized by loss of airway epithelial integrity and dysregulated repair. The causes are not completely understood, but may include persistent alloreactive, infective and non-specific epithelial injury resulting in increased epithelial cell apoptosis . As part of the alloreactive response, T cells induce apoptosis of target epithelial cells by secreting granzyme b.
We noted a significantly increased proportion of T cells expressing intracellular granzyme b protein in BAL, large airway brushings and peripheral blood and increased released granzyme b in BAL from all transplant groups compared to healthy controls. Surprisingly, granzyme b was even higher in patients with BOS than in patients with acute rejection. Our data suggest that increased T cell-induced apoptosis leading to dysregulated epithelial repair may have a strong immunological role in BOS, and that measurement of granzyme b may have potential as an adjunct investigational tool for accurate prediction of the onset of the disease. The ‘gold standard’ for the diagnosis of acute rejection is the histological evaluation and grading of transbronchial biopsy according to the International Society for Heart and Lung Transplant guidelines . To this end, the practice of surveillance bronchoscopy has been adopted widely in an effort to detect important changes prior to signs of physiological dysfunction so that early therapeutic intervention can be instigated. However, the inadequacy of bronchoscopy and the difficulty in applying this approach to clinical management is evident by the wide range of acute rejection reports (22–73%) that may result from sampling errors and low sensitivity [14,15]. A recent study showed that acute lung transplantation rejection can be predicted with a sensitivity of 94% and specificity of 67% using a cut-off value of 3·1 fg of granzyme b RNA per microgram of total RNA (derived from cells obtained by BAL) . However, reliable quantitative measurement of RNA is difficult in routine practice due to rapid RNA degradation. Further, RNA measurement is only at best a guide to the levels of active protein present. Measurement of granzyme b protein may relate better to clinical disease and be more applicable to routine practice. Interestingly, levels of intracellular granzyme b protein were higher in peripheral blood than in BAL, possibly reflecting the increased release of granzyme b by cytotoxic T cells in this compartment. The lack of significance in our measurements of soluble granzyme b in BAL may reflect the small numbers tested (n = 10 in each group) compared to the relatively large numbers included in our analyses of intracellular granzyme b.
We performed preliminary evaluation of the potential diagnostic value of serial monitoring of granzyme b in peripheral blood. Importantly, we were able to detect the onset of BOS/OB in three patients prior to histological confirmation. This finding supports the potential utility of monitoring blood levels of granzyme b as a relatively non-invasive tool for prediction of BOS in lung transplant recipients, although microbiological and clinicopathological correlation is warranted for accurate interpretation of data.
Commonly used immunosuppressive agents fail to halt the progression of BOS/OB. To investigate the effectiveness of these agents for controlling granzyme b we applied physiologically relevant concentrations of methylprednisolone, tacrolimus and cyclosporin in an in vitro model of T cell activation. None of the agents tested reduced granzyme b production by more than 48% over the experimental time-frame. Cyclosporin A and tacrolimus were the most effective for reducing granzyme b production by both CD4+ and CD8+ T cells. Methylprednisolone (10−6 M) reduced overall T cell production of granzyme b effectively; however, this drug had no significant effect on granzyme b production by CD8+ T cells. These results are consistent with our previous findings that current immunosuppression protocols have a limited effect on CD8+ T cell production of IFN-γ in the peripheral blood  and T helper type 1 (Th1) proinflammatory cytokines in BAL from lung transplant patients . Taken together, our findings indicate that currently used immunosuppressive agents may not be sufficient to suppress high CD8+ T cell production of the proinflammatory Th1 cytokines and the cytotoxic mediator granzyme b in lung transplant recipients. Of particular concern was the inability of methylprednisolone, an agent used commonly for treatment for acute rejection, to reduce granzyme b production by CD8+ T cells.
We thus investigated further a novel protease inhibitor, gabexate mesilate. This agent is administered commonly for the treatment of acute pancreatitis . It has also been shown to reduce ischaemia/reperfusion-induced hepatic injury and prevent the ischaemia/reperfusion-induced decrease in bile flow in rats . Further, gabexate mesilate significantly lowered neutrophil infiltration, lung injury score and wet/dry ratio BAL protein and improved gas exchange in an animal model of lung transplantation . In the present study, gabexate mesilate was very effective at reducing granzyme b production by CD4+ T cells; however, significant suppression of this mediator by CD8+ T cells was achievable only at a concentration that also caused increased non-specific cytotoxicity of T cells. The relatively high doses required for significant inhibitory effects and the lack of effect on CD8+ T cells may thus preclude gabexate mesilate from clinical use in the context of lung transplantation; however, further studies are warranted.
Taken together, our results indicate that increased granzyme b production by T cells may contribute to the pathogenesis of BOS/OB by contributing to epithelial cell apoptosis and abnormal airway repair and that currently used immunosuppressive agents may not be adequate to completely control granzyme b production in these patients. Longitudinal investigation of granzyme b in peripheral blood T cells may provide a novel and non-invasive method for predicting BOS following lung transplantation.
The authors have no conflict of interest to declare in relation to this manuscript.