The Role of the IL23/IL17 Axis in Bronchiolitis Obliterans Syndrome After Lung Transplantation


* Corresponding author: Geert M. Verleden,


Bronchiolitis obliterans syndrome (BOS) is the leading cause of death after lung transplantation. Treatment is challenging, as the precise pathophysiology remains unclear.

We hypothesize that TH17 lineage plays a key role in the pathophysiology of BOS by linking T-cell activation to neutrophil influx and chronic inflammation.

In a cross-sectional study, bronchoalveolar lavage (BAL) samples of 132 lung transplant recipients were analyzed. Patients were divided in four groups: stable or suffering from infection (INF), acute rejection (AR) or BOS. The upstream TH17 skewing (TGF-β/IL1β/IL6/IL23), TH17 counteracting (IL2), TH17 effector cytokine (IL17) and the principal neutrophil-attracting chemokine (IL8), were quantified at the mRNA or protein level in combination with the cell profiles.

The BOS group (n = 36) showed an increase in IL1β protein (×1.5), IL6 protein (×3), transforming growth factor-beta (TGF-β) mRNA (×3), IL17 mRNA (×20), IL23 mRNA (×10), IL8 protein (×2), IL8 mRNA (×3) and a decrease in IL2 protein (×0.8). The infection group (n = 11) demonstrated an increase in IL1β protein (×5), IL6 protein (×20), TGF-β mRNA (×10), IL17 mRNA (×300), IL23 mRNA (×200) and IL8 protein (×6). The acute rejection group (n = 43) only revealed an increase in IL6 protein (×6) and IL8 protein (×2) and a decrease in IL2 protein (×0.7). Lymphocytes andneutrophils were increased in all groups compared to the stable (n = 42).

Our findings demonstrate the IL23/IL17 axis to be involved in the pathophysiology of BOS potentially triggering the IL8-mediated neutrophilia. IL6, IL1β and IL23 seem to be skewing cytokines and IL2 a counteracting cytokine for TH17 alignment. The involvement of TGF-β could not be confirmed, either as TH17 steering or as counteracting cytokine.


Bronchiolitis obliterans syndrome (BOS) after lung transplantation not only leads to morbidity, but is also the single most important cause of late mortality after lung transplantation (1). BOS is diagnosed as an irreversible loss of lung function and threatens over 50% of long-term lung transplantation (LTx) recipients (2). Until 2003, treatment for BOS was designed to suppress lymphocyte responses (adaptive immunity) and aimed to prevent or stop/slow-down the decline in forced expiratory volume in 1 s (FEV1) (3). Recently, pharmacological approaches to inhibit the innate immune system activation, like azithromycin, are claiming their place in clinical practice. Azithromycin treatment not only stops the decline in FEV1, it can even improve FEV1 in some patients with BOS. We recently attributed this improvement in FEV1 to an effect on the IL8-neutrophil-mediated airway inflammation, an important axis in the innate immune defense (4,5). Thus, elements of both the innate and adaptive immune response are targets for the therapeutic strategies utilized in BOS. To define more specific targets in the early immunological events leading to BOS, research efforts should focus especially on the early events in the pathogenesis of BOS.

BOS is accepted to be both an alloantigen-dependent and an alloantigen-independent process leading to an early bronchiolar infiltration of neutrophils, inducing tissue damage, excessive repair and finally fibrosis and obliteration of the airways (1,6). Yet, the detailed mechanism is much less clear.

In the adaptive arm of the immune response, several T helper cell subsets are being distinguished. After antigen presentation CD4+ precursor TH cells (THp) can differentiate into type 1 T helper cells (TH1) or type 2 T helper cells (TH2), depending upon the cytokine environment. IL12 is the main TH1-inducing cytokine leading to production of IFNγ, IL2 and TNF-α. TH1 cells stimulate phagocyte activation as well as delayed-type hypersensitivity and are essential for killing of intracellular pathogens. IL4 is essential for TH2 differentiation and induces TH2 cells to produce IL4, IL5 and IL13. The differentiation of TH2 cells occurs in response to helminthic parasites and environmental allergens, which causes chronic T-cell stimulation, with little innate immune response activation (7,8). Harrington et al. demonstrated that the pathway leading to generation of IL17-producing CD4+ cells is distinct from both TH1 and TH2 lineage development. The development of TH17 cells is antagonized by both TH1 and TH2 cells through IFNγ and IL4, respectively (9,10). In mice it was first described that IL23, a heterodimeric cytokine sharing a subunit with IL12, was the key cytokine in TH17 differentiation. Later on, Harrington et al. showed that IL23 failed to induce TH17 fate, but that it is essential to expand TH17 cells and to maintain TH17 commitment in mice (11). Nowadays, it is believed that IL6 acts in combination with transforming growth factor-beta (TGF-β) (12–14) to induce TH17 commitment, leading to IL17A, IL17F and IL21 expression (15,16). There is evidence that RORγT (retinoid-related orphan receptor), STAT3 (signal transducers and activators of transcription-3) and IRF4 (interferon-regulatory factor-4) act as master regulators of TH17 differentiation (17,18).

These data on TH17 development have all been generated in mice. Yet, the human situation proved not to be entirely identical. Acosta-Rodriquez et al. and Wilson et al. showed that the combination of IL1β and IL6 is primordial for TH17 alignment. IL23 has regained importance as an essential TH17 differentiating in combination with IL1β (19,20). As for TGF-β both authors demonstrated that this cytokine is not essential for IL17 production in human T cells. Acosta-Rodriquez et al. and Wilson et al. add to the controversy by suggesting that TGF-β inhibits IL17 production (19,20). The master regulators of human TH17 fate seem to be RORγT, RORα and STAT3 (21–23).

Although the precise role of the IL23/IL17 connection in host defense remains to be established, its role in recruiting neutrophils suggests that it is important for defense against extracellular microorganisms (24). On the other hand, excessive activation of the TH17 cells plays a role in several diseases characterized by immune dysregulation (psoriasis, multiple sclerosis, inflammatory bowel disease, rheumatoid arthritis…)(25), yet its role in transplantation immunology is unclear. In the context of lung transplantation, it was postulated that TH1-type cytokine profiles are found in situations of allograft rejection and TH2-type cytokine profiles in immune tolerance. In reality however, the TH1/TH2 model could not solve the puzzle of allograft rejection (26,27).

We have previously demonstrated the presence of IL17 in the airways of patients with acute rejection (AR) (28), the most important risk factor for BOS. We hypothesized that TH17 cells play a key role in the pathophysiology of BOS, especially in view of the predominant presence of neutrophils in the inflamed BOS airway. Here we present the data of a study of TH17-pathway-involved factors (IL1β/IL2/IL6/TGF-β/IL23/IL17) together with the IL8-mediated neutrophilic inflammation in a cross-sectional analysis of bronchoalveolar lavage (BAL) samples from lung transplant recipients experiencing either an AR, infection, BOS or with stable parameters.

Material and Methods

Patients/sample collection protocol

Since October 2001, all lung transplant patients in our center are enrolled in a routine prospective bronchoscopy follow-up study. FEV1, radiological examination, physical examination and bronchoscopy with BAL for determination of cell and cytokine profiles are performed around fixed time points: 21, 90, 180, 360, 540, 720, 900… days after transplantation, or in case AR, infection (INF) or BOS are suspected. Transbronchial biopsy (TBB) was performed routinely at day 21 and 90 posttransplant or in case of suspected infection or AR based on clinical, radiological or pulmonary function criteria. Sample inclusion for this study was stopped in January 2006. Out of all samples, four groups were created: a stable group, an AR group (without infection or BOS), an infection group (without AR or BOS) and a BOS group (without infection or AR). Infection was defined as a positive (myco)-bacterial or fungal culture or viral culture, immunoassay or molecular test on sputum, BAL or blood, combined with clinical findings (fever, need for antibiotics, antifungal or antiviral treatment), measurement of C-reactive protein (CRP, > 5 mg/dL), radiological examination (new radiological infiltrates) and, if available, histological examination on TBB. Colonization was defined as a positive culture of the investigated BAL samples without clinical findings, no increased CRP (<5 mg/dL), and, if available, no new radiological infiltrates or histological findings of infection on TBB. Out of the pool of samples collected during this period, strict selection criteria were used to create the aforementioned groups: Group 1: patients with confounding problems (posttransplant lymphoproliferative disorder (PTLD), suture problems or airway stenting, lung carcinoma, early death, etc) were eliminated; Group 2: incomplete samples (no FEV1, cell count or others) were eliminated; Group 3: samples with a combination of complications of AR, infection or BOS were eliminated; Group 4: only one sample per patient was retained only one time in the study (if several samples were eligible according to the previous three criteria, the most recent sample was included). The study was approved by the local hospital's Ethics Committee. FEV1 was measured according to American Thoracic Society (ATS) criteria as previously described (28,29).


Bronchoscopy was performed as previously described (30,31). Briefly, BAL was performed in a subsegmental branch of the right middle lobe or the lingula and 2 fractions of 50 mL of saline were instilled. The two recovered fractions were pooled. Part of the recovered fluid was cultured; the other part was used for total/differential cell count (by cytospin and May–Grünwald–Giemsa staining) and cytokine analysis. For the cytokine analysis, a fraction was centrifuged (500 g; 10 min; 4°C). The cell pellet was used for mRNA measurement by qPCR, the supernatant for protein measurement by ELISA. TBB specimens were examined for infection or AR. AR were classified according to International Society of Heart and Lung Transplantation (ISHLT) guidelines (32).

mRNA measurement of IL8, IL17, IL23, TGF-β and β-actin in the cell pellet of BAL

Classical and commercially available methodology was used for extraction of total cellular RNA (Invisorb Spin Cell RNA Mini kit, Invitek, Berlin, Germany), RT-PCR (SuperScript III first-strand Synthesis system, Invitrogen SA, Merelbeke, Belgium) and real-time amplification (Platinum Quantitative PCR SuperMix-UDG, Invitrogen SA) as already described in previous studies of our group (5,28). Real-time quantitative polymerase chain reaction (PCR) amplification was performed with the ABI prism 7700 Sequence detector (Applied Biosystems, Lennik, Belgium) using a plasmid-based internal calibration standard.

All primers and probes were designed based on published sequences and supplied by Eurogentec (Seraing, Belgium). The primer and probe sequences for IL8, IL17, TGF-β and β-actin cDNA have been previously published (28,33). Primer and probe sequences (in 5′-3′ direction) for IL23 were designed with Primer Express (Applied Biosystems): forward primer: TCAGTGCCAGCAGCTTTCAC; reverse primer: TCTCTTAGATCCATGTGTCCCAC and probe: FAM-CTCTGCACACTGGCCTGGAGTGCA-TAMRA. Each mRNA was measured with 45 amplification cycles consisting of a denaturation step (94°C for 15 s) and an annealing/extension step (60°C—IL8, IL17, IL23, TGF-β—or 67°C—β-actin—for 60 s). The mRNA levels are presented as ratio of cytokine mRNA over β-actin mRNA.

Protein measurement of IL1β, IL2, IL6, IL8, IL17, IL23, and TGF-β in the supernatant of BAL

Protein concentration was measured in 100 μL undiluted BAL by means of a standard sandwich ELISA purchased from Invitrogen SA (IL6, IL8, IL17 and TGF-β)/AMDS Benelux, Antwerpen, Belgium (IL23) or by a multiplex bead assay purchased from Becton Dickinson (Erembodegem, Belgium)(IL1β and IL2) with sensitivities of 0.2, 0.5, 0.5, 10; 5, 2 and 2 pg/mL, respectively. If concentrations were under the detection limit, a value of 50% of the detection limit was accorded.

Statistical analysis

In tables results are given as median (IQR), when appropriate, and in figures results are presented as scatter dot plot with the median values. Significances between groups were tested by using Kruskal–Wallis one-way analysis of variance (ANOVA) in combination with Mann–Whitney post hoc test. Correlation analysis was performed by Spearman's rank test. A p-value of < 0.05 was considered significant.


Study population

From all samples/patients enrolled in the follow-up procedure during the study period, 132 lung transplant patients were retained according to the selection/exclusion criteria. Out of the 132 patients, 36 were diagnosed with BOS, 11 with infection, 43 demonstrated an AR and 42 were stable at the time of bronchoscopy. Of all samples, colonization was present in 41 patients and absent in 91 patients (Table 1). Nine of the 41 colonized patients had a multiple colonization; in the stable group: Pseudomonas aeruginosa+Geotrichium species, Stenotrophomonas maltophila+ MRSA and CNS +Aspergillus fumigatus; for the BOS group: MRSA +P. aeruginosa; for the AR group: CNS +P. aeruginosa and Pasteurella multocida+A. fumigatus. In the infection group A. hollandicus+ CNS +S. maltophila, P. aeruginosa+ MRSA and Serratia marcescens+P. aeruginosa were cultured. The detailed patient characteristics at time of bronchoscopy are summarized in Table 1.

Table 1.  Patient characteristics
  StableINFARBOSChi-square or Kruskal–Wallis
  1. BOS = bronchiolitis obliterans syndrome; INF = infection; AR = acute rejection; SSLTx = double lung transplantation; SLTx = single lung transplantation; HLTx = heart-lung transplantation; α1-ATD =α1-anti-trypsin deficiency; PPH = primary pulmonary hypertension; OB = obliterative bronchiolitis; FEV1= forced expiratory volume in 1 s; POD = postoperative day; CsA = cyclosporin A; FK506 = tacrolimus; AZA = azathioprine; MMF = mycofenolate mofetil; AZI = azithromycine; RAD = rapamycine; Miscellaneous = bronchiectasis (4), retransplantation for BOS (2), lymfangioleiomyomatosis (1), histiocytosis X (1), obliterative bronchiolitis (1), asbestosis (1), asthma (1), end-stage obstructive lung (1).

N value 42114336/
Age 49 (40–59)52 (41–47)46 (28–58)44 (29–57)0.56
Gender 28M/14F6M/5F23M/20F14M/22F0.11
DiagnosisEmphysema18 313 10 0.12
Pulmonary fibrosis9275 
Cystic fibrosis6111 5 
Eisenmenger's syndrome0026 
FEV1,% pred 84 (70–110)60 (50–75)72 (40–79)56 (45–72)/
POD (days) 528 (192–903)175 (88–606)35 (24–96)1472 (710–2613)<0.0001
TreatmentCsA/FK50616/261/109/342/34 0.0038
AZI54523 <0.0001
RAD0005 0.0031
ColonizationAbsence of colonization30 532 24  0.082
Gr −6229 
Gr +1020 

Early postoperative conventional immunosuppressive regimen includes methylprednisolone (125 mg; 24 h), a purine synthesis inhibitor: azathioprine (1.5–2.5 mg/kg) or mycophenolate mofetil (2 g) and a calcineurin inhibitor: cyclosporine A (2 mg/kg/day) or tacrolimus (0.01–0.015 mg/kg/day), besides rATG (3 mg/kg; 3 days). Before discharge, immunosuppressives were tapered to methylprednisolone (0.2–0.4 mg/kg), azathioprine (50–150 mg) or mycophenolate mofetil (2 g) and cyclosporine A (3–6 mg/kg) or tacrolimus (0.1–0.3 mg/kg) according to trough levels (250–300 and 12–15 ng/mL, respectively). Standard immunosuppressive regimen was not altered in the presence of colonization. All patients received a prophylactic H2-receptor antagonist or a low-dose proton pump inhibitor if pathological gastroesophageal reflux had been diagnosed before transplantation.

Postoperative AR was treated with methylprednisolone (0.5–1 g; IV for 3 days), as well as conversion from cyclosporine A to tacrolimus for recurrent AR and rATG (3 mg/kg; 10 days) for persistent AR. Azithromycin treatment for BOS consisted of 250–500 mg for 5 days followed by 250–500 mg 3 times a week). Progression of BOS was treated with rapamycin (dose according to trough levels of 4–8 ng/mL), total lymph node irradiation or retransplantation. Antibiotic treatment for bacterial infection after LTx was guided using bacteriologic cultures; airway colonization, however, was not treated by antibiotics, except in those patients with recurrent infections due to multiresistant P. aeruginosa, in whom maintenance therapy with inhaled colistin was started. Detailed patient characteristics on variation in cyclosporin A/tacrolimus (CsA/FK506) and azathioprine/mycofenolate mofetil (AZA/MMF) and azithromycin and rapamycin are summarized in Table 1.

Cell count and differentiation in BAL

Kruskal–Wallis one-way ANOVA test demonstrated differences in total cell counts, number of neutrophils, macrophages, lymphocytes and in the percentage of macrophages and neutrophils in the BAL. Yet, percentages of lymphocytes in BAL did not differ significantly. The cell profiles of the BAL samples are shown in Table 2 and Figure 1.

Table 2.  Cell and cytokine profile
 StableINFARBOSKruskal–Wallis ANOVA
  1. Results are expressed as median (IQR).

  2. The mRNA levels are presented as ratio of cytokine mRNA over β-actin mRNA.

  3. The variation between all groups is calculated with one-way ANOVA and the Mann–Whitney test is used as post hoc test for significances of the BOS, infection and AR groups versus the stable group.

  4. Significance *p < 0.05; ** p < 0.01; *** p < 0.001.

Total cells103/mL 79 (38–137)488 (294–800)***177 (85–350)***80 (44–146)<0.0001
Neutrophils%2.0 (0.8–7.3)46.5 (19.4–55.8)***6.8 (1.8–19.2)**10.8 (2.0–45.0)***<0.0001
103/mL1.3 (0.6–4.6)134.1 (40.0–380.4)***7.4 (3.1–22.1)***6.9 (1.2–56.7)**<0.0001
Macrophages%90 (82–96) 45 (31–54)***87 (65–93)74 (50–87)*** 0.0002
103/mL 66 (28–122)249 (76–347)**115 (63–237)**52 (36–82)<0.0001
Lymphocytes%4.1 (1.8–7.5)5.8 (2.5–17.4)5.3 (2.8–9.5)6.0 (2.4–17.6)0.36 
103/mL2.7 (1.2–6.7)25.9 (15.7–52.2)***10.3 (3.6–26.1)***4.6 (2.5–11.6)*<0.0001
Eosinophils%0.0 (0.0–0.2)0.2 (0.0–3.0)*0.0 (0.0–0.5)0.0 (0.0–0.2)0.037
103/mL0.0 (0.0–0.2)5.3 (0.0–10.4)*0.0 (0.0–0.5)0.0 (0.0–0.0) 0.0043
IL1βProtein4.1 (2.6–5.7)21.2 (6.0–174.0)**5.1 (1.4–6.5)6.3 (4.3–13.9)** 0.0006
IL6Protein0.1 (0.1–0.5)2.1 (0.2–18.1)***0.6 (0.1–1.9)***0.3 (0.1–1.1)* 0.0001
IL2Protein 9.4 (8.3–10.2)7.0 (4.4–9.8)7.0 (1.7–9.1)***8.0 (5.4–9.7)*0.021
TGF-βProtein 13.8 (5.0–110.7)58.8 (29.4–121.5)59.0 (5.0–139.1)48.0 (5.0—144.3)0.41 
mRNA 0.02 (0.01–0.05)0.25 (0.01–0.68)*0.03 (0.01–0.10)0.06 (0.01–0.66)*0.040
IL23ProteinUnder the detection limit/
mRNA21.2 (0.2–73.6)2613 (185–1.2 × 108)***8.1 (0.1–195.1)216.5 (1.4–1250.0)* 0.0004
IL17ProteinUnder the detection limit/
mRNA0.1 (0.0–3.7)32.2 (1.8–600 443.0)**0.6 (0.0–239.6)2.0 (0.1–189.6)*0.021
IL8Protein26 (11–51) 141 (89–392)**47 (16–149)*40 (24–93)* 0.0032
mRNA 93 (14–389)173 (13—315 600)42 (10–320)272 (62–7474)*0.015
Figure 1.

BAL cell profile in lung transplant patients. The cellular profile of bronchoalveolar lavage (BAL) samples in the stable (n = 42), bronchiolitis obliterans syndrome (BOS, n = 36), infection (INF, n = 11) and acute rejection (AR, n = 43) group.

Total cells in BAL were significantly higher in infection (p < 0.0001) and in AR (p = 0.0003) groups but not for the BOS group (p = 0.63) (all compared with the stable group). Neutrophil numbers as well as percentages were significantly higher in BOS group (p = 0.0032 and p = 0.0003, respectively), infection group (p < 0.0001 and p < 0.0001, respectively) and AR group (p = 0.0030 and p < 0.0001, respectively) (compared with the stable group). Lymphocyte percentages did not differ between the groups. On the contrary, the number of lymphocytes was significantly increased in the BOS group (p = 0.032), infection (p < 0.0001) and AR group (p = 0.0005). Macrophage percentages were significantly decreased in the BAL samples of the BOS group (p = 0.0005) and infection group (p = 0.0004) compared to stable patients. No significant difference between the AR group and stable group could be observed (p = 0.099). Macrophage numbers were significantly increased in the BAL samples of the infection group (p = 0.0021) and AR group (p = 0.0045) compared to stable but there was no significant difference for the BOS group (p = 0.30).

IL1β, IL6, TGF-β and IL2 measurement in BAL

Kruskal–Wallis one-way ANOVA test demonstrated differences for IL1β protein, IL2 protein, IL6 protein and TGF-β mRNA levels, but not for TGF-β protein levels. The levels of TGF-β and IL6 in BAL are shown in Table 2 and Figure 2.

Figure 2.

The profile of TH17 skewing cytokines. IL1β protein, IL2 protein, IL6 protein, TGF-β protein and mRNA and measured with PCR and ELISA in bronchoalveolar lavage (BAL) from stable, BOS, infected and AR patients. The mRNA levels are expressed as a ratio of TGF-β/IL6 over β-actin and the protein levels are expressed in pg/mL.

IL1β protein was significantly increased in the infection group (p = 0.0034) and BOS (p = 0.0019) group compared to the stable group, but not in the AR group (p = 0.55). IL6 protein was significantly increased in the three different groups compared to stable group (INF: p = 0.0006; AR: p = 0.0005 BOS; p = 0.039). TGF-β protein levels did not differ significantly between the study groups. However, we did observe an increase in TGF-β mRNA levels in both BOS and infection samples (p = 0.029 and p = 0.036, respectively) compared to stable samples, but not in the AR samples (p = 0.25). IL2 protein was significantly decreased in the AR and BOS group compared to stable group (AR: p = 0.0002; BOS: p = 0.033) but not in the infection group (p = 0.20).

IL23 and IL17 measurement in BAL

IL23 and IL17 protein levels could not be measured; more than 50% of the samples were under the detection limit in all groups. Kruskal–Wallis one-way ANOVA test was significant for IL23 and IL17 mRNA levels. The levels of IL23 and IL17 in BAL are shown in Table 2 and Figure 3.

Figure 3.

The profile of TH17 stabilizing and effector cytokines and IL8. Levels of IL23 mRNA, IL17 mRNA, IL8 mRNA and IL8 protein measured in bronchoalveolar lavage (BAL) from stable, BOS, infected and AR patients after lung transplantation. The mRNA levels are expressed as a ratio of IL23/IL17 over β-actin and the protein levels are expressed in pg/mL.

IL23 mRNA levels were significantly increased in the BOS group (p = 0.010) and in the infection group (p = 0.0003), yet not in the AR group (p = 0.78). The same was true for IL17 mRNA, with a significant increase in the BOS group (p = 0.035) as well as in the infection group (p = 0.0066), but not in the AR group (p = 0.12).

IL8 measurement in BAL

Kruskal–Wallis test showed significant differences in both the protein as well as the mRNA levels, as shown in Table 2 and Figure 3.

Compared to the stable group, we could observe a significant increase in IL8 mRNA in BAL from BOS patients (p = 0.017); this was not the case in patients with infection or AR (p = 0.19 and p = 0.51, respectively). At the IL8 protein level, a significant increase in BOS (p = 0.020), infection (p = 0.0014) and AR patients (p = 0.033) was detected compared to stable patients.


In this study we showed an increase in the TH17 signature cytokine IL17 in BAL obtained from lung transplant recipients with BOS. Moreover, we demonstrated that these patients have higher levels of TH17 differentiating/stimulating (IL1β, IL6 and IL23) cytokines in combination with lower levels of TH17 counteracting cytokine (IL2) in the BAL. We could not clarify the involvement of TGF-β as the increase in mRNA levels was not confirmed at the protein level. The increase in both IL17 as well as in IL1β, IL6 and IL23 indicate a key role for the TH17 effector cell subset in BOS. The increase in IL17 is in line with the finding in BOS of increased numbers of neutrophils and of IL8, a potent neutrophil-attracting chemokine that is potently induced by IL17. The IL23/IL17 axis also appeared to be activated in the infection group, yet not in the AR group. In the latter, other mechanisms are likely to induce IL8 expression and the ensuing neutrophilic inflammation.

From hindsight, data published before the description of the novel TH17 subset already point to a role for the IL23/IL17 pathway in BOS. Several papers reported increased levels of cytokines, now known to induce TH17 cells, in lung transplant patients with BOS or who suffer from conditions predisposing to BOS, that is, primary graft failure, AR and infection (34–37). Belperio et al. focused on IL1β but did not show an increase in IL1β. The differences in BAL sampling and in the method of protein measurement may account for this discrepancy (38). Moreover, IL6, the hinge cytokine that directs T-cell development toward TH17 and not regulatory T (Treg) cells in the presence of TGF-β was found to be increased not only in BOS, but also in conditions predisposing to BOS (36,37).

Still, as expected, IL6 in combination with IL1β or IL23 are not the sole dictators of TH17/Treg development. IL2, the principal T-cell growth factor can actively inhibit the development of TH17 cells, possibly through a negative action of STAT5 on RORγt and RORα (39,40). Indeed, Laurence et al. showed that blockade of IL2 promotes TH17 cell differentiation. This situation can be translated to the setting of lung transplantation as IL2 is the main target for immunosuppressive therapies (41), yet this blockade of IL2 signaling is unable to prevent rejection in which TH17 cells are likely to play a crucial role. This corroborates with the findings of Laurence et al. This way, Meloni et al. demonstrated decreased numbers of Treg cells in patients with established BOS. (42). With this respect, one could also ask whether the transplant society should re-evaluate the common clinical practice of increasing immunosuppressive, IL2 downregulating therapy in case of rejection. Therefore, alternative therapies like extra corporal photopheresis (ECP) (43) and total lymphoid irradiation (TLI) (44) merit further assessment, to dampen alloreactive T cells by resetting the immunological system to tolerance. Also, all-transretinoic acid therapy could be a possible alternative therapy both in BOS and AR, as it has recently been shown that this agent inhibits TH17 polarization and enhances FoxP3 expression (45). It is clear that both IL2 and IL6 are crucial elements or ‘switch’ cytokines in the balance between tolerance and autoimmunity/rejection.

However, the relationship between TH17 cells and Tregs does not appear to be a simple reciprocal one. Unlike in mice, in the human situation there seems to be some sort of plasticity (46). A recent study showed that in the presence of IL6 or activated dendritic cells, Tregs are able to differentiate into IL17-producing TH17 cells. In a T-cell-mediated autoimmune disease model, adoptively transferred Tregs increased IL17 production by effector T cells and failed to cure established disease. Korn et al. showed that in experimental autoimmune encephalo-myelitis (EAE), Tregs were unable to abrogate IL17 production by encephalitogenic T cells, which produced high amounts of IL6 and tumor necrosis factor (TNF) (47). The authors suggested that the IL6 production in an inflammatory milieu might hamper regulation by Tregs by tipping the balance in favor of TH17 induction. This finding has significant implications for the possible use of transfer of ex vivo generated Tregs in the induction of transplant tolerance. To come back to our findings of upregulated IL6 expression in infection and AR in transplant recipients, one could speculate that the increase in IL6 and the downregulation of IL2 creates the optimal cytokine environment for TH17 and not Treg development, thus putting the patient at risk for developing BOS. This way, meticulous prevention of infection and AR would be one of the most important actions to prevent a patient from developing BOS.

Our results on TGF-β are less clear-cut, in concordance with the yet incompletely unraveled ambiguity of this cytokine in the development of TH17 and Treg cells in mice and humans (48). We show increased mRNA expression in the absence of differences in protein levels. At the start of the study, TGF-β was chosen as possible skewing cytokine for the TH17 alignment. The results cannot confirm this involvement, although other groups previously have demonstrated TGF-β to be increased in BOS (but not in AR or infection) (36,49). Moreover, these results are conflicting with the currently new postulation of Acosta-Rodriquez et al. and Wilson et al. on a possible antagonistic effect of TGF-β on TH17 fate (19,20).

Certainly, this study has some limitations. Previously, we have demonstrated that IL17 was increased in BAL samples from patients with AR (28). In this study, we were unable to reproduce this finding. One reason may be that the AR group has become larger and thus inevitably more heterogeneous. Also, in the previous study, the percentage of AR grade 2 and 3 was higher (78% vs. 47%). In this study, the AR grade 2 and 3 subgroup tended to have higher IL17 mRNA levels versus the control group (2.2 (0.1–12 685) IL17/β-actin mRNA, n = 20; p = 0.057). Also, IL17 and IL23 protein were under the detection limit in the BAL—we attribute this to the dilution of the BAL sample. Snell et al. reported that IL17 was not associated with BOS after lung transplantation (50). We believe that underdiagnosis of BOS by the use of central bronchial biopsies may account for the discrepancy with our findings. Indeed, BOS lesions are predominantly located in smaller airways rather than in the larger airways, which were biopsied in the study by Snell et al., and therefore deep endobronchial biopsies (EBB) are probably the best tools to investigate BOS lesions.

Another limitation of the study is the absence of TBB for some samples; TBB was only performed routinely around day 21 and 90 after transplantation or when indicated to confirm AR, infection or BOS (exclude AR or infection as explanation for the FEV1 decline for the diagnosis of BOS). One could argue that the findings in the infection and BOS group could therefore to some extend be explained by the presence of subclinical AR. We believe that the presence of AR in the infection and BOS group is improbable. For the infection group TBB was performed in 9 out the 11 samples and AR was absent. The two remaining patients recovered after antibiotic therapy without AR (steroid) therapy. For the BOS group 17 out of the 36 samples had a (negative) TBB. TBB was taken to confirm the diagnosis of BOS (irreversible decline in FEV1 where other complications are excluded). The other patients were already diagnosed with BOS previously and did not have a TBB performed at the time of the BAL sample as there were no clinical indication. Although we cannot exclude a subclinical AR in the absence of TBB, its presence late after transplantation (inherent to BOS) is uncommon and therefore can be put aside.

The single dimension of the analysis could also be seen as a shortcoming. Yet, multivariate analysis is not possible as most of the inflammatory parameters are related to each other. Of all possible confounding factors, therapy (azithromycin, shift toward FK506 and rapamycin) and time of sampling as expected are different between the groups. Another possible confounding factor is colonization. In the two previous studies, we already demonstrated that colonization did not affect the inflammation profile (28,30). Yet we also published the contrary (31). The discrepancy can be explained by the presentation of the colonization. If, on the one hand, there is a heterogeneous pool of colonization within a group, with an equal distribution over the investigated groups, it would not bias the findings, as for this study. If on the other hand one single organism is present and this predominantly in one group (preferentially Pseudomonads) this will have an effect on the inflammatory profile.

Subanalysis of this study demonstrated no difference in colonized versus noncolonized samples in the stable and AR group for both the heterogeneous group of colonizations and specific pool of Pseudomonads colonized samples. In the BOS group, out of the 34 samples 12 were colonized of which 10 had positive Pseudomonas cultures. Subanalysis was difficult as colonization was predominantly present in the higher grade of BOS. Our results are however unlikely to be explained by the presence of pseudomonads as subanalysis of the noncolonized BOS versus the noncolonized stable samples confirmed the present results. Only the significances were less pronounced but this can be due to the lower samples size and elimination of higher BOS graded samples.

In conclusion, in this study, we show for the first time in patients experiencing BOS higher levels of TH17 differentiating (IL1β, IL6 and IL23), effector (IL17) cytokine and IL8 (an important IL17-derived neutrophil-attracting) chemokine. In BOS also, increased neutrophil numbers are demonstrated. These findings clearly highlight the involvement of the IL23/IL17 pathway in lung transplantation and hence the crucial role of the balance between TH17 effector cells and Treg cells in BOS. The role of TGF-β is unclear: a possible role as skewing cytokine could not be confirmed, yet a possible counteracting role seems unlikely. Further studies are needed to confirm these findings and to look into the role of TH17 cells in BOS, as this may have important implications for future therapeutic strategies for BOS, a condition for which there is up till now no satisfying therapy available.


B. M. Vanaudenaerde and S. I. De Vleeschauwer contributed equally to this work. G. M. Verleden is holder of the GSK (Belgium) chair in respiratory pharmacology at the KULeuven, and is supported by the Research Foundation Flanders (FWO): G.0493.04, G.0518.06 and G.0643.08. B. M. Vanaudenaerde and L. J. Dupont are senior research fellows and R. Vos is a Ph.D. fellow of the Research Foundation Flanders (FWO). None of the authors have a financial relationship with a commercial entity that has an interest in the subject of the presented manuscript.