Bronchiolitis obliterans syndrome is associated with increased p‐glycoprotein expression and loss of glucocorticoid receptor from steroid‐resistant proinflammatory CD8⁺ T cells

Summary

Immunosuppressive therapy fails to suppress the production of proinflammatory cytokines, particularly by CD8⁺ T cells, in stable lung transplant recipients and those undergoing chronic rejection, suggesting that some patients may become relatively resistant to immunosuppressants such as glucocorticoids (GC). We have shown loss of GC receptor (GCR) from the CD8⁺ cells, and we hypothesized that the drug membrane efflux pump, p‐glycoprotein‐1 (Pgp), may also be involved in lymphocyte steroid resistance following lung transplant. Pgp/GCR expression and interferon (IFN)‐γ/tumour necrosis factor (TNF)‐α proinflammatory cytokine production was measured in blood lymphocytes from 15 stable lung transplant patients, 10 patients with bronchiolitis obliterans syndrome (BOS) and 10 healthy aged‐matched controls (± prednisolone ± Pgp inhibitor, cyclosporin A ± GCR activator, Compound A) using flow cytometry. Both Pgp⁺ and Pgp^– lymphocyte subsets from all subjects produced IFN‐γ/TNF‐α proinflammatory cytokines. Pgp expression was increased in CD8⁺Pgp⁺ T cells and correlated with IFN‐γ/TNF‐α expression and BOS grade. Reduced GCR was observed in CD8⁺Pgp^– T, natural killer (NK) T‐like and NK cells from stable patients compared with controls, and reduced further in CD8⁺Pgp^– T cells in BOS. The addition of 2·5 ng/ml cyclosporin A and 1 µM prednisolone inhibit IFN‐γ/TNF‐α production significantly by CD8⁺Pgp⁺ T cells from BOS patients. The addition of 10 µM Compound A and 1 µM prednisolone inhibit IFN‐γ/TNF‐α production significantly by CD8⁺Pgp^– T cells from BOS patients. BOS is associated with increased Pgp expression and loss of GCR from steroid‐resistant proinflammatory CD8⁺ T cells. Treatments that inhibit Pgp and up‐regulate GCR in CD8⁺ T cells may improve graft survival.

Introduction

Immunosuppression therapy following lung transplant is inadequate at preventing chronic graft failure in many patients, with 5‐year survival rates less than 60%, and incidence of bronchiolitis obliterans syndrome 50% at 5 years 1. T cell T helper type 1 (Th1) proinflammatory cytokines are involved in transplant rejection and are a target of current immunosuppression strategies 2. Our previous studies of stable lung transplant patients showed inadequate immunosuppression in peripheral blood T cells [evident by lack of suppression of interferon (IFN)‐γ and tumour necrosis factor (TNF)‐α by CD8⁺ T cells], even though systemic drug levels were within ‘therapeutic range’, suggesting that measurement of intracellular T cell cytokine levels may provide a more physiological indication of immunosuppression than drug levels 3.

In support of this, we showed further that bronchiolitis obliterans syndrome (BOS) and lymphocytic bronchiolitis (an important risk factor for the development of obliterative bronchiolitis 4 were associated with lack of suppression of peripheral blood T cells IFN‐γ and TNF‐α 5, 6, 7. Importantly, these proinflammatory mediators were shown to increase with time post‐transplant in otherwise stable patients before clinical signs of declining lung function 8.

These findings suggest the development of a resistance to immunosuppressants, such as glucocorticoids (GC), following lung transplantation. This may be explained partially by our recent findings of down‐regulated GC receptor (GCR) 9 in peripheral blood T cells from stable lung transplant recipients, as binding of GC to its receptor is required to regulate gene transcription. Another probable contributor to the GC resistance is increased levels of the drug efflux pump, p‐glycoprotein‐1 (Pgp), which has been well documented in drug‐resistant cancer cells 10 and described recently in steroid‐resistant T cell subsets in patients with chronic obstructive pulmonary disease (COPD), another chronic inflammatory lung disorder 11.

We hypothesized that increased expression of Pgp in GC‐resistant T cells following lung transplant would be increased further and accompanied by a loss of GCR in patients with BOS.

To investigate this hypothesis, intracellular cytokine production by CD8⁺ and CD8^– (CD4⁺) T, natural killer (NK) T‐like cell and NK subsets, GCR and Pgp expression were determined in cultured peripheral blood from cohorts of stable lung transplant patients, patients with BOS and aged‐matched healthy control subjects, using multi‐parameter flow cytometry.

Materials and methods

Patient and control groups

For all participants, fully informed consent and institutional ethics approval was obtained. Ten lung transplant recipients with clinical and physiological criteria for BOS were invited to participate in the study. Where possible, patient rejection status was categorized histologically on transbronchial biopsies, according to standard criteria 12. Demographic details of these patients are shown in Table 1. Fifteen lung transplant recipients with no clinical or histopathological evidence of acute rejection or BOS, scheduled for routine surveillance assessment, were also recruited. Demographic details of the patients are presented in Table 1. All patients were submitted to the same protocol. All patients were tested to exclude cytomegalovirus [CMV, histopathologically, rapid viral culture and CMV polymerase chain reaction (PCR) of bronchoalveolar lavage (BAL], mycoplasma (enzyme immunoassay of BAL), bacterial and fungal infection (BAL culture). All transplant recipients were classified as A0B0 or A0BX and were 29·1 ± 26·9 months [mean ± standard deviation (s.d.)] post‐transplant (Table 2).

Table 1.

Demographic details of the populations studied

Subjects	Healthy controls	Stable	BOS
No. of subjects	10	15	10
SLT/DLT		6/9	4/6
Age (years)	44 (± 10)	42 (± 12)	43 (± 10)
FEV1, % pred	100·6 (± 21)	72·8 (± 21)	46·9 (± 18)
FVC, % pred	99·5 (± 15)	77·5 (± 15)	54·3 (± 11)
FEV1, % FVC	96·2 (± 24)	77·6 (± 14)	48·4 (± 14)

BOS = bronchiolitis obliterans syndrome; DLT = double lung transplant; FEV₁ = forced expiratory volume in 1 s; FVC = forced vital capacity; SLT = single lung transplant; Stable = lung function recipients with stable lung function.

Table 2.

Lung transplant patients and previous acute rejection (ACR) episodes

Patient	Predisposing pathology	Time post‐transplant (months)	Grade ACR	Number of prior ACR episodes	Grade of prior ACR episodes	*CsA/Tac levels
1	Cystic fibrosis	18	A0B0	2	A1	Tac 15
2	Emphysema †	108	A0B0	1	A2	Tac 13
3	Bronchiectasis	14	A0B0	0		Tac 12
4	Pulmonary hypertension	19	A0BX	0		CsA 150
5	† Emphysema	29	A0B0	0		Tac 11
6	Congenital bronchial webbs	112	A0BX	2	A3B0 A2B0	Tac 17 Tac 18
7	Pulmonary hypertension	10	A0B0	2	A3B0 A2B0	Tac 9 Tac 11
8	Emphysema †	50	A0B0	1		Tac 8
9	Pulmonary hypertension	107	A0BX	0		Tac 12
10	Emphysema	45	A0B0	1	A2B0	Tac 17

Therapeutic range for cyclosporin A (CsA) (80–250 μg/l) and tacrolimus (Tac) (5–20 μg/l).

*All patients received the same calcineurin inhibitor post‐transplant. ^†Patients receiving low‐dose steroids prior to transplant.

Immunosuppression therapy comprised combinations of either cyclosporin A (CsA) or tacrolimus (Tac) with prednisolone (12 ± 3·6 mg/day (mean ± s.d.) and azathioprine or mycophenolate mofetil. A cumulative dose of prednisolone was calculated for each patient, including low‐dose steroid prior to transplant in three patients (Table 2). Trough plasma drug levels of either CsA or Tac were within or above recommended therapeutic ranges [ranges for CsA (80–250 μg/l) and Tac (5–15 μg/l)] (Table 2). Ten healthy age‐matched volunteers were recruited as controls. Venous blood was collected into 10 U/ml preservative‐free sodium heparin (DBL, Sydney, Australia) and blood samples were maintained at 4°C until processing.

Leucocyte counts

Full blood counts, including white cell differential counts, were determined on blood specimens using a CELL‐DYN 4000 (Abbot Diagnostics, Sydney, Australia). Blood films and BAL cytospins were stained by the May–Grunwald–Giemsa method and white cell differential counts checked by morphological assessment microscopically.

IFN‐γ/TNF‐α and GCR/Pgp expression in CD8⁺ and CD8^– T, NK T‐like and NK cell subsets

To determine IFN‐γ/TNF‐α production and expression of GCR/Pgp in CD8⁺ and CD8^– T and NK T‐like cells and NK cells, aliquots of blood were stimulated as reported previously 3, 6 with phorbol myristate (25 ng/ml) (Sigma, Sydney, Australia) and ionomycin (1 μg/ml) (Sigma) in the presence of brefeldin A (1 μg/ml) (Sigma). The culture tubes were incubated in a humidified 5% CO₂/95% air atmosphere at 37°C 3. At 16 h, 100 μl 20 mM ethylenediamine tetraacetic acid/phosphate‐buffered saline (EDTA/PBS) was added and tubes vortexed vigorously for 20 s to remove adherent cells. Red blood cells were lysed and cells were permeabilized as described previously 3, 6. Two ml of 0·5% bovine serum albumin (BSA) (Sigma/Aldrich, Sydney, Australia)/Isoflow (Beckman Coulter, Sydney, Australia) was then added and the tubes centrifuged at 300 g for 5 min. After decanting supernatant, Fc receptors were blocked with 10 μl human immunoglobulin (Ig) (Intragam, CSL, Parkville, Australia) for 10 min in the dark at room temperature. Five μl appropriately diluted anti‐GCR (clone 5E4; Serotec, Sydney, Australia, raised against a conserved sequence of the regulatory part of the receptor), was added for 15 min in the dark at room temperature (RT), as reported previously 9. Cells were washed with wash buffer as described above and 5 μl appropriately diluted rat anti‐mouse IgG1 V450 (BD Bioscience, Sydney, Australia) was added for 15 min. Following cell washing, appropriately diluted monoclonal antibodies to IFN‐γ fluorescein isothiocyanate (FITC) (BD Bioscience), TNF‐α FITC (BD Bioscience), Pgp phycoerythrin (PE) (BD Bioscience), CD3 peridinin chlorophyll (perCP) cyanin (Cy) 5·5, CD56 allophycocyanin (APC) (Beckman Coulter, Sydney, Australia), CD8 APC Cy7 (BD) and CD45 V500 (BD Bioscience) were added for 15 min in the dark at room temperature. Cells were washed with wash buffer. After decanting, cells were analysed within 1 h on a fluorescence activated cell sorter (FACS)Canto II flow cytometer using FACSDiva software (BD Bioscience). Samples were analysed by gating lymphocytes using CD45 staining versus side‐scatter (SSC). A minimum of 350 000 low SSC events were acquired in list‐mode format for analysis. T cells were identified as CD3⁺CD56^–CD45⁺, NK T‐like cells identified as CD3⁺CD56⁺CD45⁺ and NK cells identified as CD3^–CD56⁺CD45⁺ low FSC/SSC events.

IFN‐γ/TNF‐α, Pgp and GCR expression by CD8⁺ T cells in BAL

To determine whether Pgp⁺CD8⁺ proinflammatory T cells are present in the lungs of patients with BOS, these cells were enumerated in bronchoalveolar lavage specimens from three patients with BOS, five stable transplant patients and four control subjects, as performed previously 11, 13.

Effect of Pgp inhibition and GCR activation on IFN‐γ/TNF‐α expression in CD8⁺ and CD8^– T, NK T‐like and NK cell subsets

To determine the effects of inhibiting Pgp on IFN‐γ/TNF‐α expression in CD8⁺ and CD8^– T, NK T‐like and NK cell subsets, 1‐ml aliquots of blood from patients with BOS were incubated with a range of different concentrations of known Pgp inhibitor, CsA ± 1 µM prednisolone for 24 h and culture tubes were incubated in a humidified 5% CO₂/95% air atmosphere at 37°C 3. IFN‐γ/TNF‐α production in Pgp⁺CD8⁺ and CD8^– T and NK T‐like cells and NK cells was then determined as described above. Preliminary experiments showed that significant inhibition of IFN‐γ/TNF‐α production occurred in Pgp⁺CD8⁺ and CD8^– T and NKT‐like cells and NK cells at concentrations of CsA ≥ 2·5 ng/ml in the presence of 1 µM prednisolone. Similarly, to determine the effects of GCR activation on IFN‐γ/TNF‐α expression in Pgp^–(GCR low)CD8⁺ and CD8^– T, NK T‐like and NK cell subsets, 1‐ml aliquots of blood from patients with BOS were incubated with 10 µM of known GCR activator Compound A 13 (Santa Cruz Biotechnology, MetaGene Pty Ltd, Redcliffe, Qld, Australia) ± 1 µM prednisolone for 24 h and culture tubes were incubated in a humidified 5% CO₂/95% air atmosphere at 37°C 3. IFN‐γ/TNF‐α production in CD8⁺ and CD8^– T and NK T‐like cells and NK cells was then determined as described above.

Statistical analysis

Statistical analysis was performed using the Friedman test with Wilcoxon's signed‐rank tests for post‐hoc pairwise comparisons. Correlations were performed using Spearman's rho correlation tests. spss software was applied and differences between groups of P < 0·05 was considered significant.

Results

Increased Pgp expression in CD8⁺ T cells in BOS

In patients with BOS, there was a significant increase in the percentage of peripheral blood CD8⁺ T cells expressing Pgp compared with healthy controls (Fig. 1). In these patients there was also a significant decrease in the percentage of peripheral blood CD8^– T cells, CD8^– NK T‐like cells and NK cells expressing Pgp compared with healthy controls (Fig. 1). There was no change in the percentage of CD8⁺ NK T‐like cells expressing Pgp between any groups (Fig. 1).

Figure 1.

Box‐plots showing the percentage of P‐glycoprotein‐positive (Pgp⁺) CD8⁺ and CD8^– T cells, CD8⁺ and CD8^– natural killer (NK) T‐like cells and NK cells in control (clear bars), stable transplant patients (light grey bars) and patients with bronchiolitis obliterans syndrome (BOS) (dark grey bars). There was a significant increase in the percentage of CD8⁺ T expressing Pgp in patients with BOS compared with control subjects. There was a significant decrease in the percentage of CD8^– T, CD8^– NK T‐like and NK cells expressing Pgp in BOS patients compared with control subjects and in CD8^– NK T‐like cells in stable patients compared with BOS patients. There was no change in the percentage of CD8⁺ NK T‐like cells expressing Pgp between groups.

Decreased GCR expression in CD8⁺ T cells and NK cells in BOS

In stable transplant patients, there was a significant decrease in the percentage of CD8⁺ and CD8^– T and NK T‐like cells expressing GCR compared with control subjects, consistent with previous results 9 (Fig. 2). In BOS patients, there was a significant decrease in the percentage of peripheral blood CD8⁺ T cells expressing GCR compared with stable patients and healthy controls (Fig. 2). There was also a significant decrease in the percentage of NK cells expressing GCR in BOS patients compared with healthy controls (Fig. 2). There was no change in the percentage of other lymphocyte subsets expressing GCR between any groups (Fig. 2).

Figure 2.

Box‐plots showing the percentage of glucocorticoid receptor‐positive (GCR⁺) CD8⁺ and CD8^– T cells, CD8⁺ and CD8^– natural killer (NK) T‐like cells and NK cells in control (clear bars), stable transplant patients (light grey bars) and patients with bronchiolitis obliterans syndrome (BOS) (dark grey bars). There was a significant decrease in the percentage of CD8⁺ T expressing GCR in patients with BOS compared with stable patients and stable patients compared with control subjects. There was a significant decrease in the percentage of CD8^– T, CD8⁺ and CD8^– NK T‐like and NK cells expressing GCR in patients with BOS and stable patients compared with control subjects.

Increased IFN‐γ/TNF‐α production by CD8⁺ T cells in BOS

In BOS patients, a significant increase in the percentage of peripheral blood CD8⁺ T cells producing IFN‐γ and TNF‐α was observed, compared with stable patients and healthy control subjects, and consistent with previous findings 8 (data not shown). In stable transplant recipients, there was a significant decrease in the percentage of peripheral blood CD8^– T cells and NK T‐like cells and NK cells producing IFN‐γ and TNF‐α compared with healthy control subjects (data not shown). There were no changes in the percentage of other lymphocyte subsets producing IFN‐γ and TNF‐α between any groups (data not shown).

Increased IFN‐γ/TNF‐α production by Pgp⁺ and Pgp^–CD8⁺ T cells in BOS

Both IFN‐γ and TNF‐α proinflammatory cytokines were produced by Pgp⁺ and Pgp^– lymphocytes. In BOS patients, there was a significant increase in the percentage of Pgp⁺ and Pgp^– CD8⁺ T cells producing IFN‐γ and TNF‐α compared with stable patients and healthy control subjects (Fig. 3). In both stable transplant recipients and those with BOS, there was a decrease in the percentage of Pgp⁺CD8^– T, CD8⁺ and CD8^– NK T and NK cells producing IFN‐γ and TNF‐α compared with control subjects (e.g. % CD8^–Pgp⁺IFN‐γ cells for healthy, stable, BOS: 15 ± 5, 9 ± 3, 6 ± 5). Combined staining with GCR identified no change in the Pgp⁺IFN‐γ⁺ T, NK T or NK lymphocyte subsets expressing GCR between groups. However, there was a significant decrease in GCR expression in Pgp^–IFN‐γ⁺ T, NK T or NK subsets in patients with BOS compared with other groups. Representative plots showing the percentage of Pgp⁺IFN‐γ⁺⁺ and Pgp^–IFN‐γ⁺⁺ CD8⁺ T cells expressing GCR from a patient with BOS and a healthy control subject are presented in Fig. 4.

Figure 3.

Box‐plots showing the percentage of interferon (IFN)‐γ and tumour necrosis factor (TNF)‐α‐producing P‐glycoprotein‐positive (Pgp⁺) and Pgp^–CD8⁺ T cells in control (clear bars), stable transplant patients (light grey bars) and patients with bronchiolitis obliterans syndrome (BOS) (dark grey bars). There was a significant increase in the percentage of Pgp⁺ and Pgp^–CD8⁺ T cells producing IFN‐γ and TNF‐α in patients with BOS compared with stable patients and control subjects.

Figure 4.

Representative plots showing the percentage of P‐glycoprotein‐positive (Pgp⁺) interferon (IFN)‐γ⁺ and Pgp^–IFN‐γ⁺ CD8⁺ T cells expressing glucocorticoid receptor (GCR) from a patient with bronchiolitis obliterans syndrome (BOS) and a healthy control subject. There was a significant decrease in GCR expression in Pgp^–IFN‐γ⁺CD8⁺ T cells from BOS patients compared with control subjects, but no change in GCR expression in Pgp⁺IFN‐γ⁺ subsets. [Colour figure can be viewed at http://wileyonlinelibrary.com]

There was a significant decrease in GCR expression in Pgp^–CD8⁺ T cells from BOS patients compared with stable patients and control subjects and between stable patients and controls (Fig. 5), but no change in GCR expression in Pgp⁺GCR⁺CD8⁺ T cells between groups (P > 0·05 for all). There was a significant decrease in GCR expression in Pgp^–CD8⁺ NK T‐like cells from BOS patients and stable patients compared with control subjects (Fig. 5). There was a decrease in GCR expression in Pgp^– NK cells from BOS patients compared with stable patients and controls (Fig. 5), but not Pgp⁺ NK cells (data not shown).

Figure 5.

Box‐plots showing the percentage of glucocorticoid receptor‐positive (GCR⁺) P‐glycoprotein‐positive (Pgp⁺) and Pgp^–CD8⁺ T, natural killer (NK) T‐like and NK cells in control (clear bars), stable transplant patients (light grey bars) and patients with bronchiolitis obliterans syndrome (BOS) (dark grey bars). There was a significant decrease in GCR expression in Pgp^–CD8⁺ T cells from BOS patients compared with stable patients and control subjects and between stable patients and controls, but no change in GCR expression in Pgp⁺GCR⁺CD8⁺ T cells between groups (P > 0·05 for all). There was a significant decrease in GCR expression in Pgp^–CD8⁺ NK T‐like cells from BOS patients and stable patients compared with control subjects. There was a decrease in GCR expression in Pgp^–NK cells from BOS patients compared with stable patients and controls.

Correlation between Pgp‐positive CD8 T cells and IFN‐γ/TNF‐α expression in patients with BOS

There was a significant correlation between the percentage of Pgp⁺CD8⁺ T cells in the peripheral blood and expression of proinflammatory cytokines IFN‐γ (r = 0·764, P = 0·023) and TNF‐α (r = 0·641, P = 0·031).

Correlation between Pgp⁺CD8 T cells and BOS grade

There was a significant correlation between the percentage of Pgp⁺CD8⁺ T cells in the peripheral blood and BOS grade (r = 0·764, P = 0·041).

Correlation between the percentage of GCR⁺Pgp^–CD8⁺ cells expressing IFN‐γ/TNF‐α in BOS patients

There was a significant negative correlation between the percentage of GCR⁺Pgp^–CD8⁺ cells expressing IFN‐γ (r = –0·583, P = 0·029) and TNF‐α (r = –0·612, P = 0·031) in patients with BOS, but no correlations between these parameters in stable patients or controls (data not shown).

Correlation between GCR‐positive T cells and time post‐transplant for stable patients

There was a significant correlation between the percentage of GCR^– T cells in the peripheral blood and time post‐transplant for stable transplant patients (r = 0·526, P = 0·031), consistent with previous findings 9, but not for patients with BOS (data not shown).

Correlation between cumulative dose of prednisolone and time post‐transplant for stable patients

There was a significant correlation between the cumulative dose of prednisolone that the stable transplant patients had received and time post‐transplant (r = 0·789, P = 0·031), consistent with previous findings 9. No correlations were observed for patients with BOS (data not shown).

IFN‐γ/TNF‐α, Pgp and GCR expression by CD8⁺ T cells in BAL

The percentage of CD8⁺ T in BAL for BOS, stable and control subjects was 65 ± 5·7, 47 ± 12·9 and 33 ± 22·1 (mean ± s.d.), respectively. The percentage of CD8⁺ T cells producing IFN‐γ in BAL for BOS, stable and control subjects was 74 ± 20·3, 42 ± 28·1 and 20 ± 18·5, respectively. The percentage of CD8⁺ T cells expressing Pgp in BAL for BOS, stable and control subjects was 42 ± 8·5, 30 ± 10·8 and 27 ± 7·3, respectively. The percentage of CD8⁺ T cells expressing GCR in BAL for BOS, stable and control subjects was 20 ± 10·3, 35 ± 8·5 and 52 ± 12·6, respectively. Due to the small number of BOS patients and control subjects, appropriate statistics could not be performed.

Effect of Pgp inhibition and GCR activation on IFN‐γ/TNF‐α expression in CD8⁺ and CD8^– T, NK T‐like and NK cell subsets

The addition of 1 µM prednisolone failed to inhibit IFN‐γ/TNF‐α production significantly by CD8⁺Pgp⁺ and CD8⁺Pgp^– T cells from BOS patients compared to those from stable patients and controls. Prednisolone also failed to inhibit IFN‐γ/TNF‐α production significantly by CD8⁺Pgp^– T cells from stable patients and controls (although a trend was noted for decreased TNF‐α production by these cells) (Fig. 6).

Figure 6.

Graph showing the inhibitory effect of 1 µM prednisolone on production of interferon (IFN)‐γ and tumour necrosis factor (TNF)‐α in CD8⁺ P‐glycoprotein‐positive (Pgp⁺) and CD8⁺Pgp^– T cells from cells in control (clear bars), stable transplant patients (light grey bars) and patients with bronchiolitis obliterans syndrome (BOS) (dark grey bars) (median ± standard error of the mean, n = 3). Addition of 1 µM prednisolone failed to inhibit IFN‐γ and TNF‐α significantly in CD8⁺Pgp⁺ and CD8⁺Pgp^– T cells from BOS patients compared with stable patients and controls and in CD8⁺Pgp^– T cells between stable patients and controls (trend for TNF‐α).

The addition of 2·5 ng/ml CsA⁺ 1 µM prednisolone resulted in significant inhibition of IFN‐γ/TNF‐α production by CD8⁺Pgp⁺ T cells from BOS patients compared with 1 µM prednisolone alone (Fig. 7). The addition of 10 µM GCR activator, Compound A, resulted in significant inhibition of IFN‐γ/TNF‐α production by CD8⁺Pgp^– T cells from BOS patients compared with 1 µM prednisolone alone (Fig. 7).

Figure 7.

Addition of 2·5 ng/ml cyclosporin A⁺ 1 µM prednisolone (horizontal striped bars) resulted in a significant inhibition of interferon (IFN)‐γ/tumour necrosis factor (TNF)‐α production by CD8⁺ glycoprotein‐positive (Pgp⁺) T cells from bronchiolitis obliterans syndrome (BOS) patients compared with 1 µM prednisolone alone (speckled bars). Addition of 10 µM glucocorticoid receptor (GCR) activator, Compound A, resulted in a significant inhibition of IFN‐γ/TNF‐α production by CD8⁺Pgp^– T cells from BOS patients (vertical striped bars) compared with 1 µM prednisolone alone.

Discussion

This is the first study to show an increase in the drug efflux pump, Pgp, in CD8⁺ T in patients with BOS. Pgp is the most studied membrane protein of the large mammalian ABC transporter family 14. ABC transporters may be therapeutic targets in organ transplantation, but the importance of Pgp in immune function remains unclear 15. Due to its drug efflux properties, it is likely that up‐regulation of Pgp in lymphocyte populations could result in decreased intracellular drug concentrations, rendering these cells resistant to immunosuppressive therapy despite appropriate plasma drug exposure 15. In this regard, Pgp has been shown to be increased in CD8⁺ T cells during acute rejection episodes; however, to our knowledge, Pgp expression has not been studied in patients with BOS 16. In the present study we noted that Pgp expression was associated with an increase in the proinflammatory nature of CD8⁺ cells, findings consistent with previous studies 3, 6. Importantly, the percentage of Pgp⁺CD8⁺ T cells correlated with BOS grade, indicating Pgp to be a potentially important therapeutic target to potentially improve graft failure in lung transplant patients undergoing chronic graft rejection. We have previously shown an increase in proinflammatory cytokines in peripheral blood CD8⁺ T cells in lung transplant recipients preceding chronic graft rejection, and it will be interesting to follow stable transplant patients to determine if Pgp expression is increased in these patients before a fall in lung function and subsequent diagnosis of BOS 6.

Immunosuppressive agents such as cyclosporin, tacrolimus, sirolimus and corticosteroids are substrates for Pgp, with the potential to induce up‐regulation of Pgp activity resulting in increased resistance to immunosuppressive drugs 15. A previous study showed an increase in Pgp activity in CD8⁺ T cells from solid organ transplant patients as measured by increased uptake of fluorescent Pgp substrate Rhodamine 123 15. In addition, we have shown a direct correlation between uptake of efflux dye Calcein‐AM and Pgp expression in steroid‐resistant lymphocyte subsets in patients with chronic obstructive pulmonary disease 11. The reason for the increase in the percentage of Pgp⁺CD8⁺ T cells in patients with BOS compared with stable transplant patients is unclear, as we found no correlation between dose of tacrolimus or prednisolone and Pgp expression in these cells 15. However, the correlation between Pgp expression in CD8⁺ T cells in BOS patients, accompanied by up‐regulation of both IFN‐γ and TNF‐α and failure of standard‐dose prednisolone to inhibit these proinflammatory cytokines in vitro, suggests strongly that Pgp expression may be one cause of steroid resistance in these cells. In this regard, our findings of decreased IFN‐γ and TNF‐α synthesis in Pgp⁺CD8⁺ T ells in the presence of Pgp inhibitor CsA is consistent with this idea. There have been several reports of other agents that inhibit Pgp expression in various cells, and these drugs may prove beneficial in inhibiting Pgp in proinflammatory CD8⁺ T cells in patients with BOS 17, 18. Our results are consistent with reports of CD8⁺ effector T cells producing IFN‐γ and TNF‐α trafficking to the lung in murine obliterative airway disease 19. Furthermore, successful treatment of BOS has been documented in a bone marrow transplant patient treated with TNF‐α blockade 20.

Our current study showed that there were no differences in GCR expression in Pgp⁺IFN‐γ/TNF‐α^high+CD8⁺ T cells between patients with BOS and stable transplant patients, suggesting that increased Pgp expression alone may be the cause of steroid resistance in the Pgp⁺IFN‐γ/TNF‐α^high+ lymphocyte subset. Importantly, both the Pgp⁺IFN‐γ/TNF‐α⁺ and the Pgp^– IFN‐γ/TNF‐α^high+ subsets of CD8⁺ T cells produced increased proinflammatory cytokines in BOS patients compared with stable patients and control subjects, suggesting a failure of immunosuppression protocols to both reduce Pgp and up‐regulate GCR expression in proinflammatory CD8⁺ T cells in these patients.

We have shown previously that a loss of GCR was associated with an increased percentage of T and NK T‐like cells producing IFN‐γ and TNF‐α in stable lung transplant patients compared with healthy control subjects, consistent with our current findings 9. We now show that patients with BOS have a loss in GCR expression and, importantly, synthesis of IFN‐γ and TNF‐α was inhibited significantly in the presence of GCR activator, Compound A and standard‐dose prednisolone. We further show a correlation between the loss of GCR and the previous dose of prednisolone received in stable transplant patients 9. However, there was no correlation between GCR expression and prednisolone dose in patients with BOS, suggesting that factors other than GCR expression are involved in steroid resistance in these patients. Measurement of prednisolone levels in these steroid resistance cells would be worthwhile to identify whether intracellular steroid levels are indeed low (and hence due possibly to the increase in Pgp expression). Our recent findings of reduced levels of histone deacetylase 2 in CD8⁺ T and NK T‐like cells following lung transplant may be a further factor contributing to steroid resistance in these cells, and would be another worthwhile investigation 21. Nine of the 10 patients with BOS were being treated with a single calcineurin inhibitor, tacrolimus, and it would be worthwhile studying patients being treated with other calcineurin inhibitors such as CsA to determine if Pgp and GCR expression in CD8⁺ T cells are altered compared with those treated with tacrolimus. In this regard, one study showed that treatment with CsA was superior to tacrolimus at inhibiting T cell function, suggesting that CsA may also be superior at inhibiting Pgp activity 15.

Our findings of increased Pgp⁺GCR^– IFN‐γ⁺ CD8⁺ T cells in the airways of patients with BOS compared with control subjects indicates that these steroid‐resistant cells are also in the lungs of these patients, although further numbers are required to confirm this statistically. Another important investigation would be the characterization and comparison of Pgp and GCR in lymphocyte subsets from the airways and intraepithelial compartment 22. Furthermore, measurement of Pgp/GCR levels in intraepithelial T cells in the grafts of patients compared with those from native lung may also improve our understanding of the cell biology associated with acute rejection 23. We have shown recently that BOS is associated with increased cytotoxic proinflammatory CD8⁺ T cells in the small airways. Importantly, we showed that there was no difference in the percentage of these lymphocytes between small airways and blood samples, suggesting that measurement of proinflammatory cytokines in CD8⁺ T cells in blood is a less invasive measure of proinflammatory CD8⁺ T cell status in the lungs 22.

Despite the historical role of GC in transplantation, many transplant clinicians are now searching for steroid‐sparing regimens 24, and there has been a previous report of successful steroid withdrawal following lung transplant 24. A targeted approach would be to monitor GCR and Pgp levels in individual patients following steroid withdrawal. If GCR was up‐regulated and/or Pgp down‐regulated in patients showing signs of rejection, it would be logical to readminister steroid in the short term and monitor longitudinal GCR and/or Pgp levels in IFN‐γ‐ and TNF‐α‐producing T and NK T‐like cells.

There are several important limitations in the current study. The number of patients and control subjects was small, particularly the number of patients with BOS and there heterogeneous nature of BOS patients with single and double lung transplants. Further studies with larger numbers of well categorized patients with CLAD are warranted.

In conclusion, BOS is associated with increased Pgp expression and loss of GCR from steroid‐resistant proinflammatory CD8⁺ T cells. Treatments that inhibit Pgp and up‐regulate GCR in CD8⁺ T cells may improve graft survival.

Author contributions

G. H. performed the concept and design of experiments, analysis and interpretation of data and manuscript preparation; S. H. helped with study design, statistical analysis and helped draft the manuscript; P. T. N. supplied and characterized patient specimens and helped draft the manuscript; A. Y. supplied and characterized patient specimens and helped draft the manuscript; P. S. supplied and characterized patient specimens and helped draft the manuscript; A. B. supplied and characterized patient specimens and helped draft the manuscript; C. H. L. supplied and characterized patient specimens and helped draft the manuscript; P. N. R. supplied and characterized patient specimens and helped draft the manuscript; M. H. supplied and characterized patient specimens and helped draft the manuscript. All authors read and approved the final manuscript.

Disclosure

Authors have no disclosures.