Identification of the lipid biomarkers from plasma in idiopathic pulmonary fibrosis by Lipidomics

Idiopathic pulmonary fibrosis (IPF) is a disease featured as chronic, progressive, irreversible interstitial pneumonia with a poor prognosis of unknown etiology [1] and a median survival of 3 to 5 years after initial diagnosis [2, 3]. It more commonly occurs in patients of 50 to 70 years of age. This disease is characterized by the histological pattern of usual interstitial pneumonia [4, 5] with parenchymal fibrosis and excess collagen deposition [6]. It is well accepted that the pathology of this disease includes fibroblast/myofibroblast proliferation, activation of alveolar epithelial cells, with exacerbated deposit of extracellular matrix leading to the gradual destruction of the lung tissue [7]. Although there have been increasing number of studies investigating the pathogenesis of idiopathic pulmonary fibrosis (IPF), the lacking of effective treatments and early diagnostic tools indicate the urgent needs for reliable biomarkers in both diagnosing at early stage and monitoring the progression of this disease. In the recent publications, a few potentially useful blood cellular and molecular biomarkers have been identified, including chemokines, proteases and growth factors [8]. Despite the recent progress, there has not been a single biomarker proved to be useful in diagnosing and monitoring the progress of IPF. The diagnosis of IPF has not changed much over the last few years, which means that a multi-disciplinary approach is probably required for a breakthrough. With the new emergent technology of lipidomics, to identify novel lipid biomarkers for diagnostic and monitor purposes of IPF becomes possible.

Lipids comprise diverse classes of molecules that play a critical role in cellular energy storage, structure, and signaling [9‐11]. Previously, lipids are only considered as the components of membranes and source. Now lipids are known to act as a indispensable factor in the immune response by organizing signaling complexes in cellular membrane, such as lipid rafts [12] or affecting the immune reaction by release of lipid-derived mediators [13, 14]. The etiologies of certain diseases have been proven to be associated with individual lipid molecules and many studies have indicated that certain lipid metabolic disorders or abnormalities can lead to a variety of human diseases [10, 15‐19].

The role of lipids in lung and respiratory diseases has attracted more attention in recent years including cystic fibrosis, asthma and COPD which are all associated with abnormal metabolism. For example, the epithelium lipid metabolism has been proved to be changed in asthmatic patients [20, 21]. The amount of ceramides in the airway epithelium of a guinea pig model was found to be increased in response to the induction of experimental allergic asthma [22]. In another study, the increased amount of ceramides was detected in the airway epithelium and has been linked to cell death, infection susceptibility and immune inflammation in cystic fibrosis [23]. When it comes to IPF however, the human plasma lipid profiles of IPF is so far poorly understood and to identify reliable and unique lipid molecular biomarkers will be very beneficial in IPF diagnosis and management [24]. Based on the above reasons, to analyze and identify potential IPF-specific lipid biomarkers will contribute significantly to the diagnosis and management in IPF patients.

With the development of omicis [25] and the advanced mass spectrometry have made it feasible to identify and quantify a variety of lipids species in human samples such as the tandem mass spectrometer utilized (MS/MS) [26, 27],direct infusion MS (DIMS) [28], and liquid chromatography-mass spectrometry (LC-MS) [29‐32]. Among different LC-MS platforms, ultraperformance liquid chromatography coupled with quadrupole time of flight mass spectrometry (UPLC-QTOF/MS) is widely adapted to lipidomics due to its enhanced reproducibility of retention time [33‐35]. In this study, a global lipid profiling was performed containing measurement of 20 kinds of fatty acid, 159 kinds of glycerolipids, 221 kinds of glycerophospholipids, 47 kinds of sphingolipids, 46 kinds of sterol lipids, 7 kinds of prenol lipids, 3 kinds of saccharolipids, and 4 kinds of polyketides. Subsequently, the correlation analysis, receiver operating characteristic (ROC) analysis and orthogonal partial least squares discriminant analysis (OPLS-DA) were performed to evaluate the variations in lipid metabolites. The potential influence of gender, smoking history, and disease stages on the lipid metabolites was also looked at between IPF object and/or controls. Our results provided vital information regarding lipid metabolism in IPF patients and more importantly, a few potentially promising biomarkers were firstly identified which may have a predictive role in monitoring and diagnosing IPF disease.

IPF is a kind of chronic and progressive disease with low survival rate, and remains to be a clinical challenge. It is regarded as a fetal disease due to its poor prognosis and a low median survival of only 3~5 years after diagnosis. At the moment, only clinical data are available for diagnosis to researchers and clinicians, which is of limited value as they do not reflect the precise pathological mechanisms underlying IPF. To better elucidate disease mechanism and make early diagnosis of IPF, identification of molecular biomarkers with high diagnostic value is paramount. In recent publications, a number of biomarkers have been hypothesized to be present in serum used as potential prognostic or diagnostic tools. The main protein biomarkers associated with cell dysfunction are the Krebs con den lungen-6 (KL-6) antigen and the surfactant protein A and protein D (SP-A and SP-D) [39]. Biomarkers found in IPF involved in fibrogenesis are matrix metalloproteinases-1 and -7 (MMP-1 and MMP-7), which play a role in the breakdown and remodeling of extracellular matrix components [40]. However, no lipid biomarkers have yet been studied for IPF. This work is the first comprehensive investigation into the potential lipid biomarkers for IPF using UPSL-QTOF-MS/MS.

Lipids is the fundamental component of cellular membranes, also exert several essential and critical roles in cellular functions including energy storage, signal transduction, formation of membrane bilayer and cellular barriers. The metabolism of lipid is indicated in numerous human diseases, such as Alzheimer’s disease, diabetes, obesity, atherosclerosis and several types of respiratory diseases. There are eight major categories of lipid types based on their structures: fatty acid, glycerolipid, saccharolipid, polyketide, sphingolipid, sterol lipids, prenol lipid and glycerophospholipid. Although abnormal lipid metabolism has been shown to result in cystic fibrosis and lung injury published by Ollero’s and Goss’ research groups, respectively [41, 42], its potential role in pathology of IPF remains unclear.

In this study, analysis of lipid profile from 22 IPF patients and 18 control subjects revealed the characterization of lipid composition. The glycerophospholipids (GPs) appears to be the important biological molecules for the backbone of cellular membranes. Besides an integral component of biomembranes, GPs also seem to be a reservoir of a large amount of many bioactive mediators [43], which are produced by the reaction of phospholipases on GPs. GPs can be further divided into different categories including glycerophosphoglycerols (PGs), glycerophosphatidic acids (PAs), glycerophosphoinositols (PIs), glycerophosphoserines (PSs), glycerophosphoethanolamines (PEs) and cardiolipin (CLs).

In our study, 2 CLs, 2 PAs, 13 PCs, 2 PEs, 1 PIs, and 4 PSs were identified as unique lipids of IPF patients based on VIP scores (Table 2). The levels of all screened GPs was decreased in IPF patients compared to the control subjects. This phenomenon may be due to the fact that PG is a precursor of CL biosynthesis and PA is the critic substrate for biosynthesis of PI, PG, PE, and PC. A previous TLC-MALDI/TOF-based metabonomics study displayed significantly decreased plasma PCs in cystic fibrosis patients, including PC(P-40:1), PC(38:6), PC(38:5), PC(38:4), PC(38:2), PC(38:0) and PC(36:5) [44]. However, using electrospray ionization tandem mass spectrometry (ESI-MS/MS)-based metabonomics, Marien et al. showed an increase in non-small cell lung cancer of PI(38:3), PI(40:3), PI(38:2), PS(32:0), PS(36:4), PS(36:1), PS(40:2), PS(38:1), PS(34:0), PS(40:1), PS(38:5), PS(34:2) and PS(38:4), compared to normal lung tissues [45]. Increasing evidence implicates that GPs biomarkers were several in agreement with abnormal GPs metabolism found in above mentioned lung disease and IPF model. This result indicated that GPs with top scores could be used as potential biomarkers for IPF.

Our study identified 159 glycerolipids in IPF patients and 30 out of these 159 have the capacity to differentiate IPF patients from control subjects (VIP > 1, Table 2). Glycerolipids possess long chain fatty acids in ester linkage to the glycerol backbone and were classified into two major groups: diacylglycerol (DG) and triacylglycerol (TG), which are the most abundant lipids found in circulating plasma [46]. For example, all six DGs out of 30 glycerolipids showed similar tendencies to increase in the plasma from IPF patients with variable magnitudes. Interestingly, 9 out of 24 TGs displayed increased levels in IPF patients compared to control subjects. A previous study demonstrated that glycerolipids were significantly reduced in lung tissues of mice, which lacked the p53 oncogene [47]. This result indicates that the abnormal function of glycerolipid was associated with lung disease. Further investigation of the exact roles of glycerolipids in IPF patients would be beneficial.

We detected 47 sphingolipids (SLs) by UPSL-QTOF-MS/MS in plasma samples and found that 4 out of them can distinguish the IPF patients from control subjects. 4 types of SLs (GalCerd18:1/26:1, d18:1/20:0, d16:1/18:1, d17:1/24:0) significantly decreased in plasma of patients with IPF. SLs are ubiquitous cellular membrane components that are implicated in multiple cellular processes including autophagy, apoptosis, differentiation and cell division [48]. Ceramide is generated from either sohingomyelin or de novo sphingolipids synthesis [49] and its upregulation has been found in chronic-obstructive pulmonary disease [50]. In contrast, our results showed decreased SLs serum level in IPF patient, which could be due to the differences between different respiratory diseases. In the future, we will compare SLs levels in serum and bronchoalveolar lavage fluids between respiratory diseases, such as COPD and pneumonia. Although it remains to be equivocal whether differences exist in SLs levels between IPF and other respiratory diseases, our data have provided a direction for future investigation into implications of these SLs in IPF patients.

Moreover, this study identified 46 kinds of serol lipids and 20 kinds of fatty acids. 3 kinds of 46 serol lipids and 1 kind of 20 fatty acids were considered as potential biomarkers. Fatty acids are the most important class of lipids and function as precursors of various bioactive lipid molecules. In our study, fatty acid (E,E)-3,7,11-Trimethyl-2,6,10-dodecatrienyl dodecanoate was shown to be positively correlated with IPF and may distinguish IPF patients from control subjects. Three (3-Deoxyvitamin D3,16:1 Stigmasteryl ester, and 20:1-Glc-Sitosterol) out of 46 sterol lipids identified have been shown to correlate with the IPF. Other studies have shown that plasma cholesterol level was significantly increased in diet-induced hyperlipidomic rats [51], and cholesteryl ester apparent lipid molecular species have high sensitivity, specificity and accuracy in the diagnosis of prostate cancer [10]. Further validation of the specificity of these lipid molecules in IPF patients among respiratory disease would be desirable.

Furthermore, we investigated lipid molecules of 12 metabolites possessing close correlations with IPF disease. 6 of them showed higher sensitivity and specificity for identifying IPF patients from control subjects ROC analysis (Fig. 3). Previous study has demonstrated that age, gender, and smoking status can affect plasma lipid metabolite levels in healthy adults [52, 53]. The impact of gender and smoking on 6 promising biomarker levels were determined in this study and no correlation was found (Table 3). Further validations on whether the 6 identified promising biomarkers has the increased ability to discriminate IPF objects from healthy controls or other respiratory disease are highly suggested.

For the analysis of IPF patients and biomarkers, the potential defect in this study was the small sample size. Another limitation is the lack of a longitudinal study, which made it impossible to observe the clinical impact of the present discovery as no treatment responses can be assessed. Although we have identified promising lipid biomarkers in this study, further validations are necessary to evaluate the specificity of the identified biomarkers for IPF diagnosis. Then, further longitudinal multicenter studies would contribute more to evaluate the real value of lipid biomarkers as diagnostic and prognostic tools. In addition, the lipid biomarkers identified in this study should be compared with known diagnostic and prognostic biomarkers in the future, such as KL-6, SP-A, SP-D. This will provide an insight into a better understanding of the diagnostic and prognostic utility of identified lipid biomarkers. Lastly, more specific lipidomics analyses of these lipid changes in IPF will probably help to better understand the IPF pathology and may contribute to future development of novel therapeutic targets.

In conclusions, our study has yielded important information regarding lipid metabolism in IPF patients and presents the first identification of promising potential biomarkers for the diagnosis of IPF. Our results demonstrate that individual lipid molecules have the ability to differentiate IPF from controls. Implications for future studies include validation of the accuracy of biomarkers to diagnose IPF and investigation in their IPF-specificity compared to other respiratory diseases, such as asthma, COPD, and infective pneumonia.