A pilot data analysis of a metabolomic HPLC-MS / MS study of patients with COPD

Material and methods. Participants were recruited at the University Hospital, Hradec Králové, Czech Republic, with the approval of the ethics committee. The analysis of blood samples was performed at Health Sciences Center (HSC) in Kuwait. The blood samples were analyzed for concentrations of acylcarnitines and amino acids by high performance liquid chromatography (Waters 2690 HPLC; Waters, Milford, USA) and a triple-quadruple tandem mass spectrometer (Quattro LC, Micromass, Manchester, United Kingdom).


Introduction
Chronic obstructive pulmonary disease (COPD) is a term referring to a heterogeneous group of chronic lung diseases marked by restricted airflow.Lung damage in COPD is nonreversible and progressive, but may be halted through proper treatment.2][3] COPD has a severe impact on patients' health and quality of life. 4,5etabolomic studies represent a modern trend in biological and clinical scientific work. 6Metabolomic profiling studies small molecules (molecular weight <900 Da) present in biological samples.Originally, mainly nuclear magnetic resonance (NMR) was applied for this purpose.Presently, the combination of chromatographic methods with mass spectrometry (combinations of gas or liquid chromatography coupled to mass spectrometry -GC-MS or LC-MS) has become a widely applied analytical method for global profiling of metabolites. 7nly a limited number of metabolomic profiling studies was performed in patients with COPD.[10][11] A metabolomic study based on the use of NMR spectroscopy was able to pinpoint the differences between serum of healthy individuals and COPD patients; particularly, decreased concentrations of lipoproteins and amino acids and increased concentrations of glycerolphosphocholine, acetate, ketone bodies, carnosine, m-hydroxyphenylacetate, phenylacetyglycine, pyruvate, and α-ketoglutarate. 12Another NMR metabolomic study demonstrated the method by investigating plasma and urine samples from 197 COPD patients and 195 adults without COPD. 13The findings were not conclusive and experiments need to be repeated and validated.On the other hand, the authors found this approach suitable for the identification of novel biomarkers, useful for determining COPD therapeutic outcomes. 13A metabolomic approach for studying COPD patients was applied to a defined cohort from the Evaluation of COPD Longitudinally to Identify Predictive Surrogate End-points (ECLIPSE) study. 14Serum from healthy subjects and Global Initiative for Chronic Obstructive Lung Disease (GOLD) stage II, III and IV patients with COPD was investigated using NMR spectrometry.The findings were confirmed by liquid chromatography with tandem mass spectrometry (LC-MS/MS) and correlated with other data.NMR spectrometry helped in identifying decreased concentrations of lipoproteins and N,N-dimethylglycine.On the other hand, increased concentrations of glutamine, phenylalanine, 3-methylhistidine, and ketone bodies were shown in GOLD stage IV COPD patients who also had decreased branched-chain amino acids (BCAAs) levels.There were found negative correlations between BCAAs, their degradation products, 3-methylhistidine, ketone bodies, and triglycerides on one side and cachexia on the other side.These concentrations positively correlated with systemic inflammation. 14n the ECLIPSE study, the concentrations of amino acids and dipeptides were measured by LC-MS/MS in groups of patients divided according to the GOLD characteristics.Differences in the levels of several amino acids and other small molecules were confirmed in the groups of COPD patients. 15The study showed differences among various groups of COPD patients (emphysema and no emphysema; cachectic and non-cachectic).Additionally, COPD patients were distinguished from 'healthy' smokers used as a control group.These results are in agreement with the findings of another study that concentrated on the differentiation of subjects with or without emphysema. 16igh performance liquid chromatography combined with tandem mass spectrometry (HPLC-MS/MS) may be used for the identification of subtle differences in biomarkers between COPD patients, healthy controls, smokers and nonsmokers, etc.8][19] Some proteins present at very low concentrations show different expression in various COPD patients, reflecting the possible pathogenic mechanisms and the extent and severity of lung remodeling.This has a potential in COPD diagnosis and disease activity tracking by biomarkers for the benefit of patients by improving the disease management.
The goal of our work was to perform a HPLC-MS/MS metabolomic study to compare the abundance of several small molecules in COPD patients and healthy volunteers.

Diagnostic criteria for chronic obstructive pulmonary disease
Chronic obstructive pulmonary disease was diagnosed on the basis of an evaluation of patients' lung functions, symptoms and history of exacerbations.All patients in this study performed a spirometry test and met the COPD diagnostic criteria of post-bronchodilator values of forced expiratory volume in 1 s/forced vital capacity (FEV1/FVC) <0.70. 1,2

Patients
Blood samples from COPD patients for this study were obtained from the participants registered in the Czech Multicenter Research Database of Chronic Obstructive Pulmonary Disease (ClinicalTrials.govidentifier NCT01923051).Blood samples from patients with coronary artery disease and healthy controls were obtained from the Department of Cardioangiology and the Transfusion Department of the University Hospital in Hradec Králové (Czech Republic), respectively.The ethics committee of the hospital approved the study.All participants signed an informed consent form.The analysis of blood samples was performed at the Department of Pharmaceutical Chemistry, Faculty of Pharmacy, Kuwait University (Kuwait).
The inclusion criteria for this study were: non-smokers or ex-smokers of at least 6 months, patients without acute exposition to carbon monoxide (CO) due to smoke or pollution inhalation (measured by smokerlyzer) and COPD patients with post-bronchodilator values of FEV 1 < 60%.
Patients were excluded upon the following criteria: current smoker or ex-smokers of less than 6 months (CO levels > 10 ppm), a known metabolic disease including diabetes type I and II, kidney disease with moderate to severe creatinine clearance, and the presence of coronary artery disease.
Venous blood was collected by standard venipuncture into BD Vacutainer ® Heparin Tubes (Becton Dickinson, Mississauga, Canada).Heparinized blood was spotted onto a filter paper cards and left to dry at room temperature for at least 24 h.
We analyzed 2 groups of subjects (Table 1).They included 10 subjects with COPD and 10 healthy controls.The mean age of the COPD group was 61.5 years with a male to female ratio (m:f) of 1.The mean age of the control group was 55 years with m:f of 1.Although the healthy control group was younger, this difference was not significant (p = 0.052).The mean body mass index (BMI) was matched in the COPD group (25.3 kg/m 2 ) and in the control group (27.10 kg/m 2 ), with p = 0.496.The COPD patients had an average fat free mass index (FFMI) of 18.79 kg/m 2 ; FEV 1 of 33 %; and FEV 1 /FVC of 0.52.Spirometric data and FFMI were not available for the control group.

HPLC-MS/MS instrumentation
The blood samples were analyzed for acylcarnitines and amino acids by a triple-quadruple tandem mass spectrometer (Quattro LC, Micromass, Manchester, UK) with a positive electrospray ionization probe.High performance liquid chromatography (Waters 2690 HPLC) with an autosampler (Waters, Milford, USA) was used for automatic injection.A designed tandem mass spectrometry (MS/ MS) program for automatic profiling of acylcarnitines and amino acids was used.The NeoLynx program (Quattro LC, Micromass, Manchester, United Kingdom) was used to determine the concentration of the diagnostic metabolites.

HPLC-MS/MS method
Using a BSD 400 Puncher (BSD Biosample punchers, Brendale, Australia), 2-3-mm disks were punched from the center of dried blood spots.The disks were placed in a capped glass for extraction of the blood with 200 µL

Statistical methods
Standard descriptive statistics was applied for the analysis: absolute and relative frequencies for categorical variables, and median supplemented by 5 th -95 th percentile range for continuous variables.Statistical significance of differences was tested using Fisher's exact test for categorical variables and the Mann-Whitney test for continuous variables.Spearman's correlation coefficient was used for the analysis of the relationship between continuous variables.Statistical analysis was computed using SPSS 22 (IBM Corporation, New York, USA; 2013).

Results
In the analysis of the concentrations of free carnitine and acylcarnitines, small variations between the 2 groups were noted.Most notably, the free carnitine to acylcarnitine ratio (C0/AC ratio) was significantly lower in the COPD group compared to the control group -COPD 0.58 µM/L (0.53; 0.68) and control group 0.73 µM/L (0.57; 0.86) with p = 0.002.The mean C8/C2 ratio in the COPD group was significantly higher, 0.03 µM/L (0; 0.50), whilst it was 0 µM/L in the control group (p = 0.03).
Gender had no influence on the metabolomic profile of COPD patients, and only a small effect in the control group (Table 2).In the control group, males had significantly higher levels of C5OH (0.11 µM/L) compared to females (0 µM/L); p = 0.034.Also, the levels of C16 were significantly higher in males (0.81 µM/L) than in females (0 µM/L); p = 0.034.

Discussion
The severity of COPD in patients is normally graded based on the degree of obstruction as measured by FEV 1 .Today, it is accepted that COPD is a complex disease with systemic effects.Transcriptomics, proteomic and metabolomic investigation have the potential to contribute to a better understanding of COPD and its natural course.This is paramount, as COPD has been recently recognized as a condition affecting the whole organism.For example, it was shown that COPD patients are very likely to suffer from metabolic syndrome with increased pro-inflammatory markers from the lungs and adipose tissue. 20he development of modern instrumental analytical techniques contributes to the investigation of many pathological states and helps in the elucidation of metabolic and other processes inside the body of a patient.In our work, we investigated the concentrations of various small molecules, mainly various amino acids and free and acylated carnitines with a different length of the alkyl chain, using the HPLC-MS/MS method.
Carnitine is a quaternary ammonium compound, synthesized by the kidneys and liver from the amino acids lysine and methionine.It is also obtained through a diet, particularly through red meat and nuts.Carnitine has a major role in the citric acid cycle by binding and transporting various acyl groups to be metabolized through β-oxidation.Consequently, the disturbance of these processes leads to various pathological states. 21,22Genetic disorders, such as primary carnitine deficiency, carnitine palmitoyltransferase deficiency type I and II among others, affect different steps in carnitine biotransformations. 23he significantly lower ratio of free carnitine/acylated carnitine (C0/AC) in COPD patients, despite the lack of distinct differences in the concentration of free carnitine, and the total and individual acylcarnitine levels, possibly indicates a predisposition to atherosclerosis as a result of inadequate β-oxidation of fatty acids, as well as of being at risk for carnitine deficiency. 24,25vidence suggests that impaired fatty acid oxidation has some effect on the development of type-2 diabetes. 26,279][30] Acylcarnitines have a potential to activate proinflammatory signaling pathways as part of the shared mechanism of COPD and cardiovascular disease. 27,31Metabolic heritability may be implicated in aberrant levels of some acylcarnitines, as it was shown that some metabolite levels are inheritable at birth and that some genes are associated with acylcarnitines. 32he role of inherited fatty acid oxidation deficiency and genetics in COPD may be further implicated by the higher levels of the acylated C8/C2 carnitines ratio in COPD patients.The C8/C2 ratio and the C8 levels are the most accurate markers of medium-chain acyl-CoA dehydrogenase deficiency (MCADD), the most common inherited defect in the fatty acid oxidation pathway marked by an inability to break down medium-chain fatty acids during periods of fasting. 33dditional findings relate to alanine, a nonessential amino acid found in a variety of foods.It may be synthesized by the body itself from pyruvate and a branched-chain amino acid.Dietary sources include meat, seafood, dairy, eggs, beans, and nuts.It is an important source of muscle energy.Alanine is an amino acid linked to energy metabolismrelated pathways of glycolysis, gluconeogenesis, the alanine cycle, and the citric acid cycle. 34Decreased levels of alanine were reported earlier in relation to COPD and muscle wasting because of disturbed metabolism of proteins. 35Our finding of decreased levels of alanine in patients with COPD compared to the control group may indicate a risk of protein malnutrition as do significantly lower levels of phenylalanine in COPD patients compared to the control group.Higher levels of pyroglutamate in COPD patients are of unknown significance and should be researched further.
These are important findings, as natural physiological concentrations of substrates related to energy metabolism (alanine, phenylalanine and carnitines), and consequently to malnutrition, are disturbed in COPD patients, while the concentrations of substrates not linked directly to energy metabolism (other amino acids) are not significantly compromised.Chronic malnutrition is a common problem in a myriad of chronic diseases, COPD and heart disease being amongst them. 36,37It is connected to complications during hospitalizations, a higher rate of readmissions, increased costs of care, and the loss of quality of life. 36,38Malnutrition in COPD is a negative predictive factor, being strongly associated with in-hospital mortality and readmittance 30 days after discharge. 39,40Skeletal muscles are affected by inactivity, unbalanced nutrition, age, and inflammation, leading to sarcopenia, which involves both muscle loss and muscle dysfunction with contractile impairment and metabolic anomalies.These responses are poorly understood (and are beyond the scope of this article), but 2 groups of patients may be identified: cachectic and with normal or increased BMI.Malnutrition may be present in cachectic patients as well as in patients with normal or increased BMI, i.e., sarcopenia and sarcopenic obesity. 41utritional intervention is beneficial for malnourished patients, combined with exercise or a rehabilitation program, which, in the long term, may be associated with lower health care costs. 36,38,42This is already being echoed by some national guidelines. 2 Metabolomic profiling using the HPLC-MS/MS method is a new application in the investigation of COPD, and as such it serves as a source of essential information on processes that occur in COPD patients' bodies.Our results were obtained in the limited number of patients.We intend to continue our research in this direction, as it is highly desirable that future studies involve a higher number of subjects to investigate a varying role of individual comorbidities and physiological COPD parameters in the long term to elucidate further COPD pathology.

Limitations
This was a pilot study conducted on a small number of patients.The control group was younger than the COPD test subjects, although not statistically significantly.This was due to the strict exclusion criteria, namely smokers or exsmokers of less than 6 months, and the presence of metabolic diseases including diabetes mellitus types 1 and 2. There is a high prevalence of diabetes mellitus in the Czech Republic, which is 8%, compared to the estimated worldwide prevalence of 2.8-4.4%. 43,44As the incidence of diabetes mellitus increases with age, the set criteria excluded most of our older control group candidates from the study. 44Further studies on significantly larger cohorts are needed.
Patients with chronic obstructive pulmonary disease are commonly afflicted with comorbidities, particularly coronary artery disease.As there is no data available on the metabolomic effects of cardiovascular diseases on the metabolism of COPD patients, future investigation on whether a cardiovascular disease concomitant with COPD further contributes to metabolomic changes may be beneficial.
As this pilot study already shows, there are metabolic differences between COPD patients and the control group, which makes future investigations of this kind rational.

Conclusions
Our study shows that the HPLC-MS/MS method is able to show minute changes in the metabolism of patients shown on a molecular level.The significance of many of these differences is not yet understood.The carnitine and acylcarnitine levels in COPD subjects in this study possibly indicate a predisposition to atherosclerosis as a result of inadequate β-oxidation of fatty acids and show the presence of oxidative stress.Furthermore, the high sensitivity to changes in circulating amino acid levels may allow us to detect subclinical malnutrition and take early preventative interventions such as nutritional supplementation and patient education.

Table 1 .
Descriptive statistics BMI -body mass index; FFMI -fat free mass index; FEV 1 -forced expiratory volume in 1 s; FVC -forced vital capacity; C0 -free or nonacylated carnitine; C2-C18 -carnitine acylated with an acyl containing the indicated number of carbon atoms; -OH -a hydroxyl group on an acyl; AC -acylcarnitines; Pyg -pyroglutamic acid; Met -methionine; Phe -phenylalanine; Tyr -tyrosine; Cit -citrulline; Glyc -glycine; Leu -leucine; Ile -isoleucine.The concentrations of metabolites are in units µM/L; categorical variable is described by absolute (relative) frequency and tested by Fisher's exact test; continuous variable is described by median (5 th and 95 th percentiles) and tested by the Mann-Whitney test.

Table 2 .
Influence of sex on the results methanol, containing known concentrations of a mixture of isotopically labeled internal standards of diagnostic acylcarnitines and amino acids.After 20 min of shaking, the solvent was gently evaporated, and the residue was mixed with 80 µL of butanolic HCl and heated at 65°C for 15 min in a capped glass tube.The solvent was evaporated again and the residue was finally reconstituted in 80 µL of the mobile phase acetonitrile/water (80/20 v/v).A 20 µL aliquot of each sample was directly injected into the mobile phase flowing to the ionization probe of the tandem mass spectrometer at a flow rate of 0.1 mL/min.A Waters 2690 HPLC with an autosampler was used for automatic injection.The run cycle time for each sample was 2-3 min from injection to injection.A designed MS/MS program for automatic profiling of acylcarnitines and amino acids was used.The NeoLynx program for Neonatal Screening (Quattro LC, Micromass, Manchester, United Kingdom) was used for automatic detection of abnormal samples.The values obtained for COPD patients and patients with coronary heart disease were compared to those of healthy controls.

Table 3 .
Influence of age on the results

Table 4 .
Influence of BMI on the results

Table 5 .
Influence of FFMI on the results

Table 6 .
Influence of FEV 1 on the results

Table 7 .
Influence of FEV 1 /FVC on the results