Abstract
YKL-40, a chitinase-like protein mainly secreted by macrophages, neutrophils and epithelial cells, is increased in patients with idiopathic interstitial pneumonia and sarcoidosis. We aimed to investigate the role of YKL-40 as a biomarker in hypersensitivity pneumonitis (HP).
72 HP patients, 100 interstitial lung disease (ILD) controls and 60 healthy controls were studied. YKL-40 was measured by ELISA in serum and bronchoalveolar lavage fluid (BALF) at baseline and follow-up. The relationship between YKL-40 levels, clinical variables and disease outcome was evaluated.
Baseline serum YKL-40 levels were significantly higher in HP patients than in healthy controls (p<0.001), but lower than in patients with other ILDs. Baseline BALF YKL-40 levels in HP patients were the highest among ILD patients. In HP patients, serum YKL-40 correlated with the diffusing capacity of the lung for carbon monoxide at baseline (p<0.01) and over time (p<0.001). HP patients whose disease progressed or who died had higher baseline YKL-40 levels than those who remained stable and survived (p<0.001). At a cut-off of 119 ng·mL−1, the baseline serum YKL-40 level predicted disease progression (hazard ratio 6.567; p<0.001), and at a cut-off of 150 ng·mL−1 was associated with mortality (hazard ratio 9.989; p<0.001).
Serum YKL-40 may be a useful prognostic biomarker in HP patients.
Abstract
YKL-40 could help clinicians to assess disease activity and outcome of patients with hypersensitivity pneumonitis http://ow.ly/U4Qw304PTJ5
Introduction
Hypersensitivity pneumonitis (HP), also called extrinsic allergic alveolitis, is a T-cell-driven immunologic lung disease caused by repeated exposure to organic particles in susceptible subjects [1, 2]. Type III and type IV hypersensitivity reactions are involved in the pathogenesis of HP [3]. HP may progress to severe pulmonary fibrosis [1, 4]. Therefore, there is a need to identify noninvasive biomarkers that can predict disease outcome in HP patients.
YKL-40 is a chitinase-like protein [5] mainly secreted by macrophages, neutrophils and epithelial cells, which is involved in the inflammatory response to tissue injury [6, 7]. It is best known as a biomarker for diseases that are characterised by inflammation, fibrosis and tissue remodelling [6, 8–10]. For instance, circulating levels of YKL-40 have been found to be associated with liver fibrosis [9], Crohn's disease [11] and rheumatoid arthritis [12]. Recently, YKL-40 has been proposed as a diagnostic and prognostic biomarker for various forms of interstitial lung disease (ILD), especially idiopathic pulmonary fibrosis (IPF) and sarcoidosis [8, 13, 14]. Whether YKL-40 is also associated with HP is unknown.
The aim of the present study was to determine the clinical utility of YKL-40 as a biomarker in HP patients. We hypothesised that YKL-40 is elevated in HP patients and that YKL-40 level is associated with disease outcome. Some of the results of this study have been previously reported in an abstract [15].
Materials and methods
Study subjects
We retrospectively studied 72 HP patients and 100 ILD controls (45 patients with IPF, 34 with idiopathic nonspecific interstitial pneumonia (iNSIP) and 21 with cryptogenic organising pneumonitis (COP)), who were followed in our institution between January 2007 and December 2013. 60 volunteers served as healthy controls. The diagnosis of HP and ILD was based on clinical/high-resolution computed tomography (HRCT) findings, bronchoalveolar lavage fluid (BALF) characteristics and/or histopathological findings on biopsy [1, 16–19]. Acute, subacute and chronic HP were defined as previously described [1, 16]. The study was approved by the local institutional review board (approval number 15-6486-BO) and the experiments complied with current laws in Germany. Informed consent was obtained from the ILD patients and healthy controls.
Definition of disease progression in HP
Disease progression was defined as deterioration of self-reported symptoms (worsening of dyspnoea, cough, chest pain and weight loss), and/or lung function (decrease in forced vital capacity (FVC) >10% and/or diffusing capacity of the lung for carbon monoxide (DLCO) ≥15%), and/or chest imaging (increase in existing or appearance of new infiltrates compatible with HP) since the last follow-up visit [20, 21]. Otherwise, the patients were defined as stable/improved.
YKL-40 and lactate dehydrogenase laboratory assays
Serum samples were obtained from all patients at the first evaluation and in 39 HP patients also during follow-up at intervals ranging from 6 to 36 months. A total of 59 follow-up measurements were obtained. The samples were stored at −20°C or −80°C until analysis. Bronchoalveolar lavage was performed via a fibreoptic bronchoscope as previously described [22]; the supernatant was collected and stored at −80°C. YKL-40 in serum and BALF was measured by ELISA (Quidel, San Diego, CA, USA) as previously described [23]. Lactate dehydrogenase (LDH) in serum was routinely measured (normal value for LDH <225 U·L−1).
Pulmonary function tests and blood gas analysis
Measurements including FVC, forced expiratory volume in 1 s (FEV1), total lung capacity (TLC), DLCO, arterial oxygen tension (PaO2), arterial carbon dioxide tension (PaCO2), arterial oxygen saturation (SaO2) and alveolar–arterial oxygen tension difference (PA–aO2) were performed at the time of blood sample collection. Values were expressed as percentages of predicted normal values [24].
Statistical analysis
Continuous variables were evaluated for a normal distribution with the Kolmogorov–Smirnov test. Parametric data are presented as mean±sem. Categorical variables are presented as either a percentage of the total or numerically, as appropriate. Multiple comparisons were performed by one-way ANOVA and least significant difference or Dunnett's post hoc tests. Comparison between two groups was done with t-test or Wilcoxon's rank test for continuous variables, and Chi-squared or Fisher's exact test for categorical variables. Spearman's correlation coefficient was obtained for correlations. Receiver operating characteristic (ROC) analysis was used to test the role of baseline YKL-40 in serum and BALF as a diagnostic marker of HP and a predictor of disease outcome. Univariate and multivariate Cox's proportional hazard regression models were used to analyse prognostic factors. The Kaplan–Meier method with log-rank test was used to analyse whether baseline YKL-40 levels were associated with disease outcome. p-values of <0.05 were considered statistically significant. All statistical analyses were performed using SPSS 17.0 (SPSS Inc., Chicago, IL, USA).
Results
Characteristics of study subjects
Demographic and laboratory characteristics of the enrolled subjects are shown in table 1. The mean±sem follow-up time of the HP patients was 38±2 months (range 1–60) from initial blood collection. Of the 72 HP patients, 11 presented with the acute/subacute form and 61 presented with the chronic form of HP. Relevant antigen exposures were fungi/moulds (4%), birds/feathers (74%), both (21%) or unknown (1%). During follow-up, 29 HP patients experienced disease progression (symptom worsening in 100%, lung function deterioration in 90%, chest imaging worsening in 38%). No patient had disease progression on the basis of symptoms alone. 16 HP patients died. The majority of patients in all groups were not treated with corticosteroids at baseline; these were those patients who were enrolled at the time of diagnosis, before treatment was started.
Baseline serum and BALF levels of YKL-40
Serum YKL-40 levels were significantly higher in HP patients (127±9 ng·mL−1) than in healthy controls (39±4 ng·mL−1; p<0.001), but lower than in IPF (214±20 ng·mL−1; p<0.001), iNSIP (184±21 ng·mL−1; p<0.05) and COP (213±33 ng·mL−1; p<0.05) patients (figure 1a). Among all ILD patients, serum YKL-40 levels were higher in males than females (213±14 versus 163±12 ng·mL−1; p<0.01) and in smokers than nonsmokers (214±13 versus 161±12 ng·mL−1; p<0.01). Serum YKL-40 levels were higher in untreated patients than in those treated with corticosteroids (206±11 versus 141±13 ng·mL−1; p<0.001). Among HP patients, serum YKL-40 levels were higher in patients with acute/subacute disease than in those with chronic disease (179±27 versus 117±9 ng·mL−1; p<0.05).
BALF YKL-40 levels were significantly higher in HP patients (21±5 ng·mL−1) than in healthy controls (3±1 ng·mL−1; p<0.001) and in those with IPF (9±2 ng·mL−1; p<0.05) (figure 1b). BALF YKL-40 levels were higher in patients with acute/subacute HP patients than in patients with chronic HP (42±10 versus 15±7 ng·mL−1; p<0.05). No effect of sex, smoking or treatment on BALF levels was observed (data not shown).
Using ROC analysis, at a cut-off of >47 ng·mL−1, serum YKL-40 levels showed the best sensitivity (88%), specificity (77%) and accuracy (88%) to discriminate HP from healthy controls (area under the curve (AUC) 0.904; p<0.001), and at a cut-off of >134 ng·mL−1, serum YKL-40 levels showed the best sensitivity (76%), specificity (61%) and accuracy (71%) to discriminate HP from IPF (AUC 0.727; p<0.001). For BALF YKL-40 levels, no significantly predictive cut-off value was found.
Correlations between serum YKL-40 and other characteristics in HP
Baseline serum YKL-40 levels positively correlated with baseline serum LDH levels (r=0.554, p<0.001), BALF total cell counts (r=0.496, p<0.001) and BALF lymphocyte counts (r=0.451, p<0.001); a weak correlation was seen with age (r=0.264, p<0.05). Serum YKL-40 levels inversely correlated with DLCO (r= −0.310, p<0.01) (figure 2a) and FVC (r= −0.376, p<0.001) at baseline. Moreover, changes in serum YKL-40 levels inversely correlated with changes in DLCO over time (r= −0.352, p<0.001) (figure 2b), but not with changes in FVC (r=0.221, p=0.176).
Serum YKL-40 levels and disease outcome in HP
HP patients who experienced disease progression had higher baseline serum YKL-40 levels than those who remained stable during follow-up (167±12 versus 98±11 ng·mL−1; p<0.001) (figure 3a). HP patients who died (n=16) had higher baseline serum YKL-40 levels than those who survived (n=56) (190±22 versus 107±9 ng·mL−1; p<0.001) (figure 3b).
Predictive value of baseline serum YKL-40 levels for disease progression in HP
Using ROC analysis, at a cut-off level of >119 ng·mL−1, serum YKL-40 levels showed the best sensitivity (81%), specificity (77%) and accuracy (79%) to predict disease progression (AUC 0.797; p<0.001). At a cut-off level of >303 U·L−1, serum LDH showed similar results (sensitivity 78%, specificity 81%, accuracy 79%; AUC 0.811; p<0.001) (figure 4a). The combination of the two cut-offs (serum YKL-40 119 ng·mL−1 and serum LDH 303 U·L−1) yielded better results (sensitivity 81%, specificity 91% and accuracy 80%) to predict disease progression.
According to the best cut-off obtained by ROC analysis, we also divided all HP patients into a high YKL-40 level group (n=34), with baseline concentrations >119 ng·mL−1, and a low YKL-40 level group (n=38) (baseline concentrations ≤119 ng·mL−1). The characteristics of patients stratified according to this cut-off are shown in table 2.
Univariate and multivariate analyses were performed to verify whether the identified cut-off predicts disease progression. In the univariate analysis, baseline serum YKL-40 levels >119 ng·mL−1 were associated with an increased risk of disease progression (hazard ratio (HR) 6.567; p<0.001). In the multivariate analysis, serum YKL-40 levels >119 ng·mL−1 were independently associated with risk of disease progression after adjustment for age, sex, body mass index (BMI), smoking history, steroid therapy, disease type, baseline FVC, DLCO and serum LDH as covariates (HR 5.208; p=0.004) (table 3).
Kaplan–Meier analysis confirmed the predictive value of serum YKL-40 for disease progression in HP (figure 5a). At 5 years, the rate of disease progression of all HP patients was 40%. The patients in the high YKL-40 group had a higher rate of disease progression (71%) than those in the low YKL-40 group (13%) (log-rank test, p<0.001).
Predictive value of baseline serum YKL-40 levels for mortality in HP
Using ROC analysis, at a cut-off level of >150 ng·mL−1, serum YKL-40 levels showed the best sensitivity (78%), specificity (84%) and accuracy (83%) to predict death (AUC 0.787; p<0.001). For serum LDH, no predictive value for mortality was found (p>0.05) (figure 4b).
In the univariate analysis, baseline serum YKL-40 levels >150 ng·mL−1 were associated with an increased risk of death (HR 9.989; p<0.001). In the multivariate analysis, serum YKL-40 levels >150 ng·mL−1 were independently associated with the risk of death after adjustment for age, sex, BMI, smoking history, steroid therapy, disease type, baseline FVC, DLCO and serum LDH as covariates (HR 6.413; p=0.013) (table 4).
Kaplan–Meier analysis showed that higher serum YKL-40 was associated with mortality in HP (figure 5b). The mortality rate of all HP patients after 5 years was 22%. The patients in the high YKL-40 group had a higher mortality rate (57%) than those in the low YKL-40 group (8%) (log-rank test, p<0.001).
Discussion
In this study, we show that serum and BALF YKL-40 levels are elevated in HP patients and other ILD patients compared with healthy controls, and correlate with pulmonary function tests (PFTs) and BALF cell counts. Moreover, serum YKL-40 was found to be an independent predictor of disease progression and mortality in HP. To the best of our knowledge, this is the first investigation of YKL-40 levels in serum and BALF and its potential role as a biomarker in HP.
HP is caused by repeated inhalation of organic particles; thus, antigen presentation and hypersensitivity reactions are involved in the development of this disorder [1, 3]. As previously reported, antigen-presenting cells produce YKL-40 in response to specific stimulation [25, 26]. Thus, it can be speculated that activated antigen-presenting cells in the lung of HP patients are responsible for the elevated levels of YKL-40 in BALF of these patients. In support of this hypothesis, we found a positive correlation between serum YKL-40 levels and BALF lymphocyte counts in HP patients. Although lymphocytes cannot directly produce YKL-40, they can induce the production of interleukin-13, which can stimulate YKL-40 production in macrophages, dendritic cells and epithelial cells [7]. On the other hand, YKL-40 levels were higher in serum than BALF of patients with HP and other ILDs, suggesting a possible spill-over of YKL-40 from the bloodstream to the alveolar space. This has also been reported in a previous study on YKL-40 in IPF [23]. As postulated for asthma [27], the elevated concentration of YKL-40 in serum of HP patients may reflect the activity of antigen-presenting cells and the systemic immune response. In line with this hypothesis, serum and BALF levels of YKL-40 were significantly higher in acute/subacute HP patients than in those with chronic HP.
Serum YKL-40 was inversely correlated with PFTs at baseline; this finding has also been observed in IPF [13] and sarcoidosis [28]. Interestingly, change in YKL-40 levels in serum also correlated with change in DLCO over time. Moreover, the serum level of YKL-40 correlated well with LDH, a nonspecific marker of tissue injury widely used to monitor the course of acute lung disease and chronic fibrotic disease [29].
We found that HP patients with progressive disease had higher levels of YKL-40 than those who remained stable or improved. Serum YKL-40 levels >119 ng·mL−1 predicted disease progression with a sensitivity of 81% and specificity of 77%, similar to serum LDH (cut-off 303 U·L−1, sensitivity 78%, specificity 81%). Moreover, the combined assessment of serum YKL-40 and LDH provided better results (sensitivity 81%, specificity 91%) than either single marker alone. By using the identified cut-off to stratify the patients, those with serum YKL-40 levels >119 ng·mL−1 had worse PFTs and a higher rate of disease progression (71%) than those with serum YKL-40 ≤119 ng·mL−1 (13%). The multivariate analysis confirmed that both a serum YKL-40 level >119 ng·mL−1 and YKL-40 as a continuous variable were independently associated with disease progression, even after adjusting for covariates. The same was not observed for serum LDH.
Moreover, HP patients with serum YKL-40 levels >150 ng·mL−1 had worse PFTs and a worse survival (57% mortality rate) compared with those with ≤150 ng·mL−1 (8% mortality rate). The multivariate analysis confirmed that a serum YKL-40 level >150 ng·mL−1 was associated with mortality, even after adjusting for covariates. This predictive cut-off for survival in our cohort is approximately two-fold higher than that reported in IPF patients (79 ng·mL−1) [23]. This suggests that production of YKL-40 level is enhanced in HP patients with active disease, and correlates with outcome.
The thresholds for serum YKL-40 described in this study might be different in other cohorts. However, this does not impact our main finding that HP patients with high serum YKL-40 levels at baseline may have a higher rate of disease progression (or death) than those with lower levels. In recent years, there has been increasing interest in defining accurate predictors of outcome in HP: a higher fibrosis score on HRCT [30], increasing severity of traction bronchiectasis [31], lower lymphocyte levels in BALF [32] and a usual interstitial pneumonia-like pattern on histology [31, 32] have been found to correlate with disease outcome. Whereas Myazaki et al. [32] previously identified a low lymphocyte level in BALF as a potential risk factor for acute exacerbation of HP, in our cohort a higher BALF lymphocyte count was seen in the HP group with high YKL-40 levels, driven by the higher number of patients with acute/subacute disease in this group. However, univariate and multivariate analyses demonstrated that lymphocyte count was not an independent predictive factor for disease progression or survival.
Besides clinical characteristics, two biomarkers have been studied in the past to evaluate outcome in HP patients: KL-6 [33, 34] and CCL17 [35]. Serum KL-6 has been shown to predict disease progression and survival in HP patients [33, 34], whereas serum CCL17 was shown to be a predictor of disease progression but not of survival [35].
With regard to the factors responsible for YKL-40 variability, we found, in agreement with others [36, 37], that YKL-40 serum levels correlated with age in our entire cohort, but the correlation was weak (data not shown). As reported previously [36, 38], we also found that former and current smokers had higher serum YKL-40 levels than never-smokers. Cigarette smoke increases the permeability of the air–blood membrane [39] and could increase both production and spill-over of YKL-40; it has been shown that YKL-40 expression is increased in alveolar macrophages and epithelial cells in lung tissue of cigarette smokers with chronic obstructive pulmonary disease and cigarette smoke-exposed mice [37]. We believe that smoking habits should be taken into consideration when interpreting YKL-40 serum levels. The relevance of this aspect for HP patients is low, since the majority of HP patients are nonsmokers. Moreover, we observed that corticosteroid-treated patients had significantly lower levels of serum YKL-40 levels than untreated patients, suggesting that corticosteroids may suppress YKL-40 release [40, 41]. This observation has not been reported in other studies of ILD patients and could be of clinical utility in assessing the clinical response to steroid treatment in HP patients.
Despite the novel findings of this study, it has several limitations. First, the healthy controls were younger than the ILD patients, and this difference could affect the respective serum levels of YKL-40. Secondly, only limited data on serum YKL-40 levels during follow-up were available, and these data did not allow us to assess the clinical value of serial YKL-40 measurements. Thirdly, baseline single nucleotide polymorphisms in the chitinase 3-like 1 (CHI3L1) encoding gene, known to influence YKL-40 levels in serum, were not determined [23]. Finally, immunohistochemistry to demonstrate YKL-40 in lung tissue obtained by biopsy was not performed, and therefore we can only speculate on the source and localisation of this protein in HP and other ILDs.
In conclusion, this study indicates that serum YKL-40 can be a prognostic biomarker in HP, suggesting a clinical role in predicting disease progression and survival, but its serum levels should be carefully interpreted in light of the confounding factors. Although the data are promising, a multicentre validation study is necessary to determine whether serum YKL-40 measurements should be routinely used in HP.
Acknowledgements
Authors’ contributions: X. Long and F. Bonella contributed to the conception and design of the study, sample collection, measurement of biomarkers, analysis and interpretation of the data, and drafting and finalisation of the manuscript. X. He collected samples and performed biomarker measurement. S. Ohshimo and M. Griese contributed to analysis and interpretation of the data. R. Sarria and J. Guzman contributed to the conception and design of the study, and drafting of the manuscript. U. Costabel contributed to the conception and design of the study, and finalisation of the manuscript. All authors have read and approved the final manuscript.
Footnotes
Support statement: This study was supported by Arbeitsgemeinschaft zur Förderung der Pneumologie an der Ruhrlandklinik (AFPR). Funding information for this article has been deposited with the Open Funder Registry.
Conflict of interest: None declared.
- Received November 18, 2015.
- Accepted September 29, 2016.
- Copyright ©ERS 2017