Comprehensive Metabolomics Profiling Towards Cystic Fibrosis Biomarker Discovery

0 WORDS DUE IN 10 HOURS FROM NOW Only for nursing tutor

Comprehensive Metabolomics Profiling Towards Cystic Fibrosis Biomarker Discovery

Abstract

Cystic fibrosis (CF) is a common genetic disease caused majorly by autosomal recessive mutations in the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) gene. In newborn, the disease can be noticed because it appears with meconium ileus. In males, the disease causes CF diabetes mellitus (CFRD), infertility, Progressive Pulmonary disease, increased chloride, and pancreatic exocrine insufficiency. Delay and/or prevention of morbidity and/or mortality are direct and positive outcomes of newborn screening for remediable genetic conditions as it permits early identification and intervention. In the Kingdom of Saudi Arabia (KSA) and the Arabian Gulf region, cystic fibrosis (CF) is thought to be more common than previously reported. Currently, measurement of Immuno Reactive Trypsinogen (IRT) in dry blood spots (DBS) is the gold standard method for initial newborn screening for CF, followed by targeted CFTR mutation analysis. However, while IRT is a known sensitive marker for CF detection in newborns only, it is also known to be not specific to Cystic Fibrosis. As such, our study, uses high resolution untargeted mass spectrometry-based metabolomics and lipidomics profiling to identify more reliable, sensitive, and specific biomarker(s) for CF in young and adult patients with CF. The identified biomarkers will then be characterized using high-resolution tandem mass spectrometry. A prototype essay will then be developed towards the gold standard LC-MS/MS based routine newborn screening. The biomarker sensitivity, specificity, and predictability will be validated in serum and dried blood spots (DBS). This newly discovered biomarker(s) may serve as an alternative (more reliable) diagnostic biomarker with additional prognostic and therapeutic value in patients with CF.

Contents

2Abstract

3Contents

5List of Figures

6List of Tables

7Chapter 1: Introduction

111.1 Techniques of Cystic Fibrosis diagnosis

131.2 Techniques of metabolomics

16Chapter 2: Literature Review

182.1 Biomarkers of inflammation

192.2 Biomarkers of infection

212.3 Biomarkers for the detection and prediction of exacerbations

212.4 Biomarkers for CF-related diabetes mellitus (CF-RDM)

23Chapter 3: Methodology

233.1 Study Design and Patient Samples

233.2 Chemicals and materials:

243.3 Human Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) ELISA Kit

263.4 DELFLA Neonatal IRT Kit (Time -resolved fluoroimmunoassay):

273.5 Metabolomics

313.6 Statistical Analysis

32Chapter 4: Results

4.132Clinical Demographics

4.2 35Human Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) ELISA:

4.336Neonatal IRT Screening ELISA:

4.4 37GC/MS metabolomics

4.5 48LCMS/MS metabolomics

63Chapter 5: Discussion

64Chapter 6: Conclusion and Future Directions

65References

 

TOC \h \z \c “Figure”

List of Figures

8Figure 1 Pathophysiological cascade of respiratory disorder in cystic fibrosis (25,26)

12 Figure 2 Flowchart of patient recruitment and study participation. PAP, pancreatitis-associated protein; PSP, pancreatic stone protein (‎28)

13 Figure 3 Dried blood spot testing (DBS) (6)

25 Figure 4 Sandwich ELISA Platform Overview

27 Figure 5 Neonatal IRT Sandwich type assay principle (©2005-2014, PerkinElmer, Inc.)

29 Figure 6 The flow work of Gas Chromatography/Mass Spectrometry (GC/MS).

30 Figure 7 The flow work of liquid chromatography /mass spectrometry (MS/MS).

34 Figure 8 Patients genotypic classification

35 Figure 9 Typical Standard Curve for CFTR, Human ELISA

36 Figure 10 IRT mean values according to age of the CF patients. A positive IRT test (> 60 ng/mL) values are high in adults CF Patients.

38 Figure 11 The flow work of human metabolome analysis using DBS sampling protocols.

39 Figure 12 Box plots and kernel density plots before and after normalization.

41 Figure 14 Scores plot between the selected PCs.

42 Figure 15 Scores plot between the selected components by PLS-DA.

43 Figure 16 Score plots of variable importance in projection (VIP).

44 Figure 17 Scores plot and permutation test between the selected components by O-PLS-DA.

46 Figure 18 Receiver Operating Characteristic curve (ROC) plot all the six classifiers.

47 Figure 19 Receiver Operating Characteristic curve (ROC) plot.

49 Figure 20 Box plots and kernel density plots before and after normalization.

50 Figure 21 Important features selected by volcano plot.

52 Figure 22 Scores plot between the selected PCs.

54 Figure 23 Scores plot between the selected components by PLS-DA.

55 Figure 24 Score plots of variable importance in projection (VIP).

57 Figure 25 Scores plot and permutation test between the selected components by O-PLS-DA.

58 Figure 26 Receiver Operating Characteristic curve (ROC) plot all the six classifiers.

59 Figure 27 Receiver Operating Characteristic curve (ROC) plot.

60 Figure 28 A heat map of significant top features from t-test.

61 Figure 29 of Pathway Analysis.

 

 

List of Tables

 

33 Table 1 Genotype characteristic

34 Table 2 phenotype profile

62 Table 3: Detailed results from the pathway analysis

 

Chapter 1: Introduction

Cystic fibrosis (CF) is an autosomal recessive disorder that is mostly found in the Caucasian populations‎ (1). In every 4100 live births, there is at least one incidence of CF in the United States of America, and in one of every 4243 live births in Saudi Arabia. The CFTR gene frequency is estimated at 3%. . Nevertheless, the association between CF and other congenital anomalies or inherited diseases has been rare since no much research has been done on the topic.

Cystic fibrosis causes chronic lungs infection and affects multiple organs. More so, it progressively results in a decline in the respiratory function and may cause eventually premature death. The organs affected by the disease include the gastrointestinal tract, liver, pancreas, and lungs. The cause of the disease is linked to a mutated Cystic fibrosis conductance regulator (CFTR) gene, which is responsible for decoding the CFTR anion channel protein. ‎(2). The CFTR anion channel protein regulates the transport of bicarbonate and chloride across the cell membranes found in the epithelial tissues in various organs within the body. The disease causes the obstruction of the airways as well as inflammation (‎3). The CFTR protein in patients with Cystic fibrosis has some structural and functional abnormalities because of the mutation on the CFTR gene that is located on chromosome 7. Based on the effect of CFTR mutations on transporter structure or function, (Classes I–VI) are recognized as seen in Figure 1.

image22.png
Figure 1 Pathophysiological cascade of respiratory disorder in cystic fibrosis (25,26)

The way the mutation in the CFTR gene affects the biosynthesis of the CFTR protein as well as the effects of the environmental factors, which lead to the phenotypic heterogeneity in Cystic fibrosis varies. The patients suffering from Cystic fibrosis frequently develop inflammation in the airways because of the chronic bacterial colonization. The colonization leads to the production of reactive oxygen species that plays an essential role in the disruption of the metabolism of the volatile organic compounds and energy as well as the degradation of the cell membranes. Some improvement in understanding the pathophysiology of Cystic fibrosis that has led to the improvement in the administration of treatment is also experienced (‎4).

Rosenstein and Zeitlin in 1998 gave comprehensive information about Cystic fibrosis. The paper identifies the disease as an autosomal life-shortening disease that is common in many populations. The paper also gives more information about the diagnosis of the disease. According to the paper, medical professionals need to carry out a timely and accurate diagnosis so that they avoid the incidences of unnecessary testing, provide the specialized and efficient medical services and to provide the correct counseling and subsequent treatment of the condition. The presence of one or more clinical features in addition to 60 mmol/L in the sweat is measured using Pilocarpine Iontophoresis (‎5).

Most diagnoses indicate the presence of the chronic sinopulmonary disease, where almost all the CF diagnosed males have obstructive azoospermia. Majority of the CF patients also have exocrine pancreatic insufficiency as a symptom. According to the paper by Koseintein and Zeitlin in the year 1998, the identification of the clinical variation present in cystic fibrosis has been improved by the ability to measure the bioelectric properties across the nasal epithelium and the ability to measure mutations in the genes. The authors also suggest that Sino pulmonary disease is observed in 2% of the patients diagnosed with cystic fibrosis. Of particular interest to the readers is the obstructive azoospermia, which involves the congenital absence of the vas deferens.

The attention comes from the fact that the individuals suffering from the conditions may not have abnormalities in the pancreatic functions and the respiratory tract. Their sweat chloride may also be healthy, raised, or even intermediate. According to the paper, whether the individuals are supposed to be given a diagnosis or not are still a controversial matter (‎5). The public health officers and their affiliates are conducting the screening of the newborn children, which is possible due to the improvement in technology.

Nazer et al. 1989 documented the occurrence of cystic fibrosis in Saudi Arabia since many people believed in the absence of the disease in Saudi Arabia. The study incorporated a study sample of thirteen children who were diagnosed with cystic fibrosis. The children showed the clinical features in addition to having high levels of chlorides in their sweat (Higher than 60 mmol/L). Before diagnosis, the children had shown symptoms for a period of 1 to 5 years with a mean of 23 months. The clinical manifestations observed included hepatomegaly, abdominal distention, steatorrhoea, recurrent respiratory infections, rectal prolapse, and failure to thrive.

Eight patients who had the symptoms of cystic fibrosis had a family history that suggested that the trait was in the family tree. However, there was no backing of the confirmatory sweat test results. The report, therefore, creates and increases the awareness of the disease in Saudi Arabia. The paper is significant to the current study because it confirms the presence of cystic fibrosis in Saudi Arabia. It is not possible to carry out metabolomics profiling of the cystic fibrosis patients in Saudi Arabia when the disease is not endemic in the area. The confirmation of the incidence of the disease in Saudi Arabia even if it was discovered long ago gives an insight that the disease is present and that the current study can take place efficiently.

Dawson et al. the year 2000 wrote a paper that addressed the historical perspective of Cystic fibrosis in the Middle East. According to the paper, the first case of a reported Arab child was done in the year 1958. The other three cases were reported four years later. The preliminary investigations showed that at least the family of one affected child was from a family that was a pure Arabic Origin for the past four generations. The patterns of recognition of the Cystic fibrosis in the Middle East had a boost after the discovery of more than 800 CFTR gene mutations. The paper by Dawson et al. 2000 is a review of both the indexed medical papers and the unindexed papers to get a glimpse of the history of the disease in Saudi Arabia. The paper is essential to this study because it is necessary to understand the history of a disease even as the researchers and the medical personnel are looking at the available medical interventions to treat the disease.

A wealth of information and reports followed, for 20 years after the initial report of the disease. In the year 1981, Al Uwihare coming from Kuwait said that no one suspected the disease until two patients were dead because no one suspected that the disease could be found in Saudi Arabia (‎7). The discovery made Katznelson give a response that suggested that Cystic fibrosis was not rare among the people in Israel including the Arabs living in Israel. The paper gives more information about all the cases in history. It also gives an overview of the interventions that the physicians used the history and the improvements so far. It is always essential to understand the preexisting methods before trying out the new ones, which are a justification for studying the historical perspective of Cystic fibrosis.

1.1 Techniques of Cystic Fibrosis diagnosis

According to Elborn 2016, the diagnosis of newborn children in the countries that have the high prevalence of the disease reduces the costs of care by reducing the severity of the disease and subsequently reducing the burden of care because the disease is detected in its early stages. Screening of the unborn children also prevents the cases of missed diagnosis as well as delayed diagnosis. Several techniques are used in the diagnosis of the newborn children namely pancreatitis-associated protein (PAP) testing, combining immunoreactive trypsinogen with DNA mutation analysis and Double IRT testing as shown in Figure 2. IRT in dry blood spots (DBS) is the gold-standard method for initial newborn screening for CF.

image23.png
Figure 2 Flowchart of patient recruitment and study participation. PAP, pancreatitis-associated protein; PSP, pancreatic stone protein (‎28)

The method used in the diagnosis is dependent on the economic, ethnic, and geographical issues. The country should therefore, choose a method that best suits its needs as far as the diagnosis of the newborn children is concerned (‎6). Most programs use the IRT then proceed to test for CFTR mutations. The diagnosis also involves taking DBS samples obtained by heel prick in the newborn soon after birth. The sweat test is carried out in the newborn children as the final diagnostic tool, probably for confirmation purposes. The diagnosed children are then referred to Cystic fibrosis care center for support and necessary care. There are cases where the children are asymptomatic at the early ages, but they have a positive IRT test (> 60 ng/mL). Such children are monitored closely since there is a likelihood that the disease might manifest as they grow older Figure 3 (‎6). REF _Ref514069586 \h \* MERGEFORMAT

image1.png
Figure 3 Dried blood spot testing (DBS) (6)

1.2 Techniques of metabolomics

Metabolomics is a discipline that studies the metabolic pathways in the bodies of living organisms as well as measuring the unique biochemical molecules that are formed in the process. The process has been used as a clinical tool in the characterization of the pathological conditions in various cases of diagnosis. Metabolomics involves the assessment of the endogenous metabolites and the quantification of the metabolites in a global and targeted manner from a biological specimen. The molecules include nucleic acids, amino acids, peptides, alkaloids, vitamins, polyphenols, organic acids and inorganic species. The molecules can be used as biomarkers representing the essential phenotype in an organism (‎4). The metabolic fingerprinting technologies that are used in the disease diagnosis gas (GC), and liquid chromatography (LC) coupled to tandem mass spectrometry (MS), Raman and infrared spectroscopy and nuclear magnetic resonance. The metabolomics techniques are in most cases focused on the analytical techniques of NMR and MS that are rich in information. The techniques that have more advantages over the rest are the NMR-based metabolomics. NMR is rapid taking 10-15 minutes, non-invasive and non-destructive meaning that it does not require the maximum treatment of the samples and gives highly replicable results, something that makes the method to be preferred by many professionals. The most significant disadvantage of using the NMR is the fact that it is limited to high concentrated metabolites because of its insensitive mature.

Nonetheless, due to the non-existence of an official CF newborn screening program in Saudi Arabia, the age at diagnosis was found to be at 2.8 ± 3.5 years (‎4). To date, > 1200 mutations and > 280 other variants are known to be associated with CF (‎3). Based on the impact of CFTR mutations on transporter structure and/or function, (Classes I–VI) are recognized. The most common CFTR mutations identified in 75% of CF patients at KFSH&RC were: 1548delG, I1234V, 3120>A, H139L, and “ΔF508” (2, 3). The Saudi Newborn Screening (NBS) program started in November 1989 for congenital hypothyroidism screening on cord blood samples (‎5). More recently, the Saudi NBS program was expanded to include 17 metabolic disorders [6-10]. In 2002, the mandatory “hemoglobinopathies, thalassemias, and G6PD deficiency” pre-marital screening law was enacted. In this study, we aim to identify a more robust biomarker(s) specific for CF based on metabolomics and lipidomics profiling in newborns as well as older individuals with CF with the potential of being integrated into the national newborn screening program in Saudi Arabia. The study will use high resolution untargeted mass spectrometry based metabolomics and lipidomics profiling to identify more reliable, sensitive, and specific biomarker(s) for CF in newborn and adult patients with CF. The identified biomarkers will then be characterized using high-resolution tandem mass spectrometry. A prototype essay will then be developed towards the gold standard LC-MS/MS based routine newborn screening. The biomarker sensitivity, specificity, and predictability will be validated in dried blood spots (DBS). This newly discovered biomarker(s) may serve as an alternative (more reliable) diagnostic biomarker with additional prognostic and therapeutic value in patients with CF.

Chapter 2: Literature Review

Pulmonary disease is the leading cause of morbidity and mortality in patients with cystic fibrosis (CF), where affected individuals develop end-stage lung disease characterized by extensive airway damage (bronchiectasis, cysts, and abscesses) and fibrosis of lung parenchyma. About 15-20% of newborns with CF present with meconium ileus, while the vast majority of patients develop pancreatic insufficiency with malabsorption. Infertility, still a major problem in >95% of males with CF due to azoospermia, is caused by absent or atrophic, fibrotic Wolffian duct structures (‎16). Congenital absence of the vas deferens (CAVD) occurs in men as well without pulmonary or gastrointestinal manifestations of CF. The clinical diagnosis of CF is often delayed because of the non-specificity of the clinically variable symptoms which can mimic several other disorders. CF patients are managed in a multidisciplinary manner that includes, but not limited to treatment of pulmonary complications, anti-inflammatory agents, mucolytic agents, and chest physiotherapy (‎12). Lung or heart transplantation in selected CF patients with severe complications might be performed. Pancreatic insufficient CF patients receive pancreatic enzyme and vitamin replacement therapy depending on the disease severity and the complications spectrum (Genereviews.org).

The diagnosis of CF in non-symptomatic newborns is based on the detection of elevated DBS-IRT level coupled with the detection of at least one CFTR mutant allele followed by demonstrating an abnormal sweat chloride excretion (> 60 mM) and diagnostic bi-allelic CFTR mutations. The diagnostic efficiency is determined by the quality and number of biomarkers available for the particular disease. Typically, for the diagnosis of several inborn errors of metabolism, gas (GC), and liquid chromatography (LC) coupled to mass spectrometry (MS) have been used for many years. Also, population-based screening programs usually rely on high-quality biomarkers. However, there is no biomarker with absolute diagnostic performance and therefore combination of markers (including ratios) is more frequently used in screening and (sometimes) diagnostic strategies. For example, in the newborn screening for PKU, both phenylalanine levels as well as Phe/Tyr ratio are quite important and their combined usage can reduce the false positive/negative detection rates. Screening for medium chain acyl Co-A dehydrogenase (MCAD) deficiency relies on the combined profile of elevated (C6, C8, C10, C10:1) acylcarnitines although they do not carry the same weight (‎4).

The biomarkers in CF are used for diagnosis based on the CFTR protein function and to evaluate aspects of lung disease severity (e.g., inflammation, infection), where the reliable markers efficiently help in monitoring disease and developing a therapy (‎2‎ ). One of the major challenges in biomarker development is the sample collection under various clinical states (such as with/without inflammation or treatment). It is generally believed that inflammation in the airways or lung disease may not be necessarily reflected in metabolic changes found in blood. However, most published studies evaluated inflammatory markers in CF lung disease in adults, and the significance of these markers in pediatric patients is still unknown (‎14). The earliest metabolomics studies on CF were performed (in-vitro) on cells extracted from three cohorts of CF patients, where a reduction in the levels of various purine nucleotides and glucose was the main feature of those profiles. This was thought to explain the regulation of cellular responses via purinergic signaling, and exacerbation of oxidative stress that limit the epithelial cell response to environmental pressure.

In the last decade, metabolomics-based studies have been conducted for several purposes using different types of samples (‎15). A serum metabolic profiling study was able to differentiat between CF and non-CF lung disease using metabolites such as 3-Hydroxybutyrate, Lauryl-carnitine, Acetyl-carnitine, Decanal-carnitine, Octanoyl-carnitine, 2-Hydroxybutyrate, Octanedioate, Azelate, Sebacate, Undecanedioic Bilirubin (Z,Z), Bilirubin (E,E) Biliverdin, Taurolithocholate-3-sulfate, Taurochenodeoxycholate, Glycocholate, Indoleacetate, 4-Hydroxyphenylacetate and Indole-propionate. Low ketone bodies, low-medium chain acyl-carnitines, elevated di-carboxylic acids and decreased 2-hydroxybutyrate from amino acid metabolism are also key metabolic abnormalities found in CF compared to non-CF patients. The role of high-dose vitamin D administration on systemic metabolism in CF adult patients with an acute pulmonary exacerbation was evaluated using metabolomics strategy which showed overall enhancement of catabolism through the differential expression of several amino acids and lipids. Most of the detected metabolites in this study were related to lipid metabolism, oxidants, and markers consistent with abnormalities in bile acids metabolism. Nuclear Magnetic Resonance (NMR)-based analysis of exhaled breath condensate (EBC) samples was used to distinguish between the respiratory phenotypes of CF vs. Primary Ciliary Dyskinesia (PCD). As sweat is an essential biological fluid for the diagnosis of CF based on elevated chloride level, human sweat metabolic profiling studies were recently performed using tandem mass spectrometry (LC-MSMS) for biomarker discovery. Overall, the metabolomics and lipidomics studies for biomarker(s) discovery in patients with CF are categorized (below) into several groups based on the purpose and clinical state, such as inflamation, infection, etc (‎17).

2.1 Biomarkers of inflammation

Patients with CF develop airway inflammation due to chronic bacterial colonization, which causes increased production of reactive oxygen species (ROS), responsible for degradation of cell membranes and disruption of the energy and volatile organic compounds (VOC) metabolism.

Esther et al. (2015) screened 25 cystic fibrotic children with neutrophilic airway inflammation using bronchoalveolar lavage (BAL) fluid samples, where most of CF correlated peaks were generated by metabolites from pathways related to metabolism of purines, polyamines, proteins, and nicotinamide and correlated them with neutrophil counts. Wolak et al. (2009) reported on alanine, valine, taurine, and lactate as discriminants between patients with high versus low airway inflammation using BAL samples (‎25). The pro-oxidative inflammatory CF respiratory tract milieu is known to contain enzymatically and non-enzymatically produced regulatory lipid mediators. The LC-MSMS based profile of polyunsaturated fatty acids was generated using pretreated patient sputum samples. A broad range of both pro- and anti-inflammatory lipid mediators were detected, including PGE2, PGD2, 20-OH-LTB4, 20-COOH-LTB4, TXB2, LTB4, 6-trans-LTB4, 20-HETE, 15-HETE, 11-HETE, 12-HETE, 8-HETE, 9-HETE, 5-HETE, EpETrEs, diols, resolvin E1, 15-deoxy-PGJ2, and LXA4. Most of these oxylipins have not been reported previously in CF secretions. Adenosine has shown a correlation with forced expiratory volume in one second (FEV1) and tracks changes in lung function in EBC samples collected from children with CF. This marker potentially could be used as an airway disease biomarker after correction for urea and phenylalanine. In a stool-based metabolomics study, lipoyl-GMP was identified as a potential novel inflammatory biomarker.

2.2 Biomarkers of infection

As human or microbial origin metabolites have the potential to be valuable biomarkers of the disease state in CF, the breath samples (VOC-based profile) are the ideal sample for the status of lung infection. The volatile organic compounds (VOCs) were profiled in exhaled breath of more than a hundred children with CF using gas chromatography-time of flight mass spectrometry (GC-ToF). This metabolic profile was not only able to distinguish between CF and control individuals, but also between CF patients with and without Pseudomonas colonization (‎16). VOCs were able to distinguish Pseudomonas aeruginosa (PA)-infected from PA-non infected CF patients, where their sensitivity and specificity for PA status were low and high once combined linearly. A proof of concept study based on GCxGC TOF mass spectrometry showed that the presence of a particular pattern of VOCs (rather than a single VOC biomarker) is necessary for bacterial species identification and stage of their growth.

The chronic polymicrobial lung infection is the main health challenge in older children and adult CF patients; where metabolomics was used to study the impact of microbial growth on airway secretion composition to understand the behavior of pathogens and their relationship with other potential colonizing agents. PA is the leading cause of morbidity and mortality in patients with CF, where quorum sensing infochemicals (such as 2-amino acetophenone) were produced during acute and chronic infection in human tissues, including the lungs of CF patients. PA mutants were characterized in sputum samples using NMR, GC x GC & TOF-MS based metabolic profiling. ADDIN EN.CITE The most leading research in CF uses metabolomics techniques for non-invasive metabiome profiling. Recently, a comparative volumetric mapping study of the microbiome and metabolome was performed using human lung affected with CF as a model. ADDIN EN.CITE Metabolites including xenobiotics, PA specific metabolites and host sphingolipids were detected in CF sputa metabolomics microbiomics profiles, whereas the microbiome analysis missed the diagnosis of PA disease.The most differentially expressed molecules in CF vs. healthy control sputum samples were sphingolipids (such as sphingomyelins, ceramides, and lactosylceramide), which led to the recent patenting of ceramide based detection of CF. ADDIN EN.CITE <EndNote><Cite><Author>The</Author><Year>2017</Year><RecNum>63</RecNum><DisplayText>[38]</DisplayText><record><rec-number>63</rec-number><foreign-keys><key app=”EN” db-id=”wr9vvetrzrfe04e0f0npsrx9fda0dz5xwevs”>63</key></foreign-keys><ref-type name=”Generic”>13</ref-type><contributors><authors><author>The, designation of the inventor has not yet been filed</author></authors></contributors><titles><title>Method for the diagnosis of cystic fibrosis</title></titles><dates><year>2017</year></dates><publisher>Google Patents</publisher><urls><related-urls><url>https://www.google.com/patents/EP3236265A1?cl=en</url></related-urls></urls></record></Cite></EndNote>

2.3 Biomarkers for the detection and prediction of exacerbations

Twenty six percent (26%) of CF patients less than 18 years of age, and 44% of those above 18 years old required hospitalization for acute pulmonary exacerbations (APE) (https://www.cff.org/Research/ Researcher-Resources/ Patient-Registry). The frequency of APEs severe enough to require hospitalization adversely impacts the life quality and life expectancy of patients and associated health care costs. However, despite the clinical importance of APEs, there is still a general lack of knowledge regarding their pathophysiology, resulting in non-uniform treatment decisions. A metabolomics-based study on EBC samples to discriminate between stable CF, pre-APE, and APE patients found lactic acid, pyroglutamic acid, and 4-hydroxycyclohexylcarboxylic acid were the main discriminants. The same group of invistigarors used the same approach to study the metabolomics profiles in stable and unstable CF patients. Hypoxanthine, N4-acetylcytidine, N-acetylmethionine, carbohydrate mannose, and cortisol were significantly perturbed in the plasma of CF patients with pulmonary exacerbation vs. clinically well state patients.

2.4 Biomarkers for CF-related diabetes mellitus (CF-RDM)

Diabetes mellitus is also a significant complication of CF as it accelerates the decline in lung function. Studying the metabolic profile of glucose connected pathways is necessary for improving our understanding of the biochemical mechanisms linking diabetes mellitus and CF. NMR and MS based targeted-metabolomic studies of exhaled breath condensate (EBC) from CF patients and healthy individuals were performed. Ion-mobility mass spectrometric analysis of EBC showed glucose level to be higher in individuals with CF-RDM compared to CF non-diabetics. A stool metabolomics study identified an unknown molecule with m/z, 463.247; retention time, 0.570717 min as a major discriminant between healthy controls, pancreas insufficient CF, versus pancreatic sufficient CF patients.

Chapter 3: Methodology

3.1 Study Design and Patient Samples

Patients with a confirmed diagnosis of CF (sweat test and genotyping) who are followed at CF clinic at King Faisal Specialist Hospital and Research Center (KFSHRC) (Riyadh, KSA). Patients were randomly selected from clinic visits to represent a range of mild to moderate CF disease and have different classes of genotypes. The subjects used in this analysis were taken from a larger cohort study of CF. Informed consent for blood collection was obtained from all patients during their clinic visit. The methodology is based on the comparison of two groups of samples: samples of patients suffering from a CF (n=39) and healthy controls (n=30). The Inclusion Criteria in this study are: young and adult patients with the clinical &/or laboratory confirmed diagnosis of Hereditary cystic Fibrosis and able to give informed consent. The exclusion Criteria in this study are: the enrolment in another clinical study in the last 30 days, inability or unwillingness to provide informed consent and being diagnosed with conditions other than those targeted in this study.

Sample Collection: Blood was obtained in standard BD tubes then spun for separation of serum. In addition, EDTA tubes ware collected for DBS cards. Both Serum and DBS ware immediately frozen and stored at -20. Incubation time is very important it may affect metabolmic biomarkers and care was taken to process all samples in a same condition.

3.2 Chemicals and materials:

Enzyme-Linked Immunosorbent Assay (ELISA) kit for cystic fibrosis transmembrane conductance regulator (CFTR) from MyBioSource (San Diego, USA). DELFLA Neonatal IRT Kit (PerkinElmer, Inc.). 99% of Methanol, Chloroform (HPLC) grade, potassium chloride (KCl) 0.2 M and Acetonitrile (ACN)were purchased from Fisher Chemical image2.png( USA). Bistrimethyl-silyl-triflouroacetamide (BSTFA) with 1% trimethylchorosilane (TMCS) was obtained from Thermo Scientific Pte.Ltd.(Waltham,MA).

3.3 Human Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) ELISA Kit

This kit is a ready-to-use microwell, strip plate. Double-antibody Sandwich ELISA (enzyme-linked immunosorbent assay) for analyzing the presence of the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR). The principle of the kit depends on the enzyme-linked immunosorbent assay technology, where 96 wells plate pre-coated with an antibody specific to CFTR. Standards and samples were then added to the appropriate microplate wells with a biotin-conjugated antibody specific to CFTR. Next, Avidin conjugated to Horseradish Peroxidase (HRP) was added to each microplate well and incubated. After TMB substrate solution was added, only those wells that contain CFTR, biotin-conjugated antibody and enzyme-conjugated Avidin exhibited a change in color, See Figure 4. The enzyme-substrate reaction was terminated by the addition of sulphuric acid solution and the color change was measured spectrophotometrically at a wavelength of 450nm ± 10nm. The concentration of CFTR in the samples was then determined by comparing the O.D. of the samples to the standard curve.

All reagents and samples were brought to room temperature. Prepared all reagents, working standards, and samples. We add 100μl of Standard, Blank and Sample per well, cover with a plate sealer and incubate for 1 hour at 37°C. Aspirate the liquid of each well, without washing. Then we add 100μl of Biotinylated Detection Antibody Reagent A (working solution) to each well, cover with a plate sealer and incubate for 1 hour at 37°C. Aspirate the liquid from each well and wash with 350µL of 1× Wash Solution to each well using a squirt bottle, multi-channel pipette 3 times. We allowed each wash to sit for 1-2 minutes before completely aspirating. After the last wash we aspirate to remove any remaining Wash Buffer then invert the plate and tap against clean absorbent paper. Next, we add 100μl of 1x HRP Conjugate working solution to each well, cover with a new plate sealer and incubate for 30 minutes at 37°C.We repeat the aspiration step and wash 5 times. Then we add 90μl of TMB Substrate solution to each well, cover with a new plate sealer and incubate for 10-20 minutes at 37°C. Protect from light. The liquid turn to blue by the addition of Substrate Solution. Finally, we add 50μl of Stop Solution to each well. The blue color will change to yellow immediately. Determine the optical density (OD value) of each well immediately using a microplate reader set to 450 nm.

image3.jpg
Figure 4 Sandwich ELISA Platform Overview

 

3.4 DELFLA Neonatal IRT Kit (Time -resolved fluoroimmunoassay):

This kit is intended for the quantitative determination of human immunoreative trypsinogen(IRT) in blood specimens dried on filter paper as an aid in screening newborns for cystic fibrosis. The principle of the DELFLA Neonatal IRT assay is a solid phase, two site fluoroimmunometric assay based on the direct sandwich technique in which two monoclonal antibodies derived from mice are directed against two separate antigenic determinants on the IRT molecule.Calibrators,controls and samples containing IRT are reacted simultaneously with immobilized monoclonal antibodies directed against a different antigenic site in assay buffer. The assay buffer elutes IRT from the dried blood spots (DBS) on the filter paper disks. The complete assay requires only one incubation step. Enhancement solution dissociation europium ions from the labeled antibody in to solution where they form highly fluorescent chelates with components of the enhancement solution. the fluorescence of each sample is proportional to the concentration of IRT in the sample (Figure 5).

All reagents and samples were brought to room temperature, all reagents were prepared and transferred the required number of microtitration strip to a strip frame. Then we punched out filter paper disks into the wells by using an automatic puncher (diameter disk should be approximately 3.2mm,1/8 inch). Only one disk is added per well. Next, we added 200 μl of diluted Anti-IRT-Eu tracer solution to each well using the recommended Eppendorf Multipette after discarding the first aliquot. Then the frame was fast shake for 10 minutes using DELFIA pasteshake, covered the plate and incubated for 16-24 hours in refrigerator (2-8 °C) without shaking. After the incubation step we removed the solution and filter paper disks from the wells using the DELFIA Washer-Diskremove program 05. Next, we added 200 μl of Enhancement solution directly from the reagent bottle to each well and we shake the frame slowly for 5 minutes. finally, we measured the fluorescence in the frame by using time-resolved fluorometer.

image4.jpg
Figure 5 Neonatal IRT Sandwich type assay principle (©2005-2014, PerkinElmer, Inc.)

3.5 Metabolomics

Samples were prepared and stored as described previously. This included extraction/depletion using an automated MicroLab STAR system (Hamilton Company). Sample preparation was conducted using a proprietary series of organic and aqueous extractions to remove the protein fraction while allowing maximum recovery of small molecules. The biochemical profiles of serum and DBS samples were extracted and analyzed using both gas chromatography (GC) and liquid chromatography (LC) platforms coupled with mass spectrometry (MS). The LC/M S/MS methods were performed in both positive and negative ion modes to maximize the identification of metabolites with diverse chemical properties and gas chromatography/ mass spectrometry (GC/MS) which is more suitable for non-polar metabolites. Quality control was assessed by the use of blanks, derivatization standard and internal standard within each run. Samples will be placed briefly on a TurboVap® (Zymark) to remove the organic solvent. Each sample was then frozen and dried under vacuum. Samples were then prepared for the appropriate instrument, either (LC/MSMS, GC/MS).

The LC/MSMS portion of the platform was based on a Waters ACQUITY UPLC and a Thermo-Finnigan LTQ mass spectrometer, which consisted of an electrospray ionization (ESI) source and linear ion-trap (LIT) mass analyzer. The sample extract was split into two aliquots, dried, then reconstituted in acidic or basic LC-compatible solvents, one aliquot was analyzed using acidic positive ion optimized conditions (gradient eluted using water/methanol containing 0.1% formic acid) and the other using basic negative ion optimized (gradient eluted using water and methanol containing 6.5mM ammonium bicarbonate) conditions.in two independent injections using separate dedicated columns.

Gas Chromatography/Mass Spectrometry (GC/MS) The analysis was performed using a Thermo–Finnigan Trace DSQ fast-scanning single-quadrupole mass spectrometer using electron impact ionization. Analyzed samples underwent vacuum desiccation (minimum 24hrs) followed by derivatization under nitrogen using bistrimethyl-silyl-triflouroacetamide (BSTFA).

Metabolites were identified by automated comparison of the ion features in the experimental samples to a reference library of chemical standard entries that included retention time, molecular weight (m/z), adducts and in-source fragments as well as their associated MS/MS2 spectra were recorded to allow for rapid identification of metabolites in patient samples with high-confidence.

Sample processing for Gas Chromatography/Mass Spectrometry (GC/MS):

Samples extraction done by adding 250 μl methanol and 120 μl chloroform to 100 μl of serum or 5 DBS disks in Eppendorf tubes. Then vortexed for 1 minutes and incubated for 1 hour at room temperature. Next, we added 380 μl chloroform and 90 μl potassium chloride (KCl) 0.2 M and vortexed. Then the extracts were centrifuged at 10000 rpm for 10 minutes at 4°C.The supernatant layer was transferred to centrifuge glass tube and added 20 μl internal standard then vortexed. For evaporation, drying was done under a stream nitrogen. Next, derivatization of the dried extracts by using 100 μl of bistrimethyl-silyl-triflouroacetamide (BSTFA) and 200 μl Acetonitrile (ACN) then vortexed for 1 minutes and incubated for 30 minutes at 90°C.Finally, after the extracts being cooled to room temperature transferred in to a 200 μl conical base inert glass insert placed inside a 2 ml amber autosampler glass vial (Agilent Technologies, Germany) for GC/MS analysis (Figure 6).

image5.png
Figure 6 The flow work of Gas Chromatography/Mass Spectrometry (GC/MS).

The flowchart used for metabolite extraction, data mining, and metabolite identification. This illustrates sample preparation and GC/MS analysis.

 

Sample processing for liquid chromatography /mass spectrometry (MS):

Samples extraction done by adding 1000 ml of extraction solvent (50% ACN /50% MeOH) to 200 μl of serum or adding 1000 ml of extraction solvent (20% ACN /40% MeOH/20% H2O) to 5 DBS disks in Eppendorf tubes. Then vortexed for 1 minutes and shaking at 1000 RPM for 1 hour at 4°C in a Thermo mixer. Next, the extracts were centrifuged at 14000 rpm for 5 minutes at 4°C.Transferred the supernatant layer to centrifuge glass tube. For evaporation, drying was done under a stream nitrogen. Next, reconstitution of the dried extracts by using 100 μl of reconstitute solution then vortexed for 1 minutes and transferred to Eppendorf tubes and centrifuged at 14000 rpm for 10 minutes at 4°C. Finally, transferred the supernatant layer in to a 200 μl conical base inert glass insert placed inside a 2 ml amber autosampler glass vial (Agilent Technologies, Germany) for LC-MS/MS analysis (Figure 7).

image6.png
Figure 7 The flow work of liquid chromatography /mass spectrometry (MS/MS).

The flowchart used for metabolite extraction, data mining, and metabolite identification. This illustrates sample preparation and LC-MS/MS analysis.

3.6 Statistical Analysis

Unpaired Student’s t-test was applied to assess the differences between two groups of rats. All statistical tests carried out using Graph Pad Prism (version 5.0, Graph Pad software, LA Jolle, CA) and the MetaboAnalyst software. Significance levels are considered at P-value 0.05, and values were presented as mean ± standard error of mean (SEM). The significant metabolites were analyzed based on Principal component analysis (PCA) for cluster analysis.Thep-values were presented on the figures as 0.0001 (***),0.001 (**),and 0.05 (*).

Chapter 4: Results

4.1 Clinical Demographics

Clinical samples often have substantial biological variation. To assess biochemical differences with strong statistical significance between CF and non-CF samples , this study included 69 young and adults grouped into two groups: 39 patients of CF and 30 healthy control . Their genotypic characteristics are shown in Table 1. CF patients were classified according to symptoms into: 34 patients with persistent Cough , 3 patients with lung Infections , 6 patients with sinusitis, 2 patients with bowel Irregularities, and 1 patient with positive Sputum culture , their phenotypic profile are shown in Table 2. Also, the percentage of CFTR mutations gene which classified in our study into five classes (class Ⅱ, class Ⅲ, class Ⅳ, class Ⅴ, class Ⅵ) according to the mechanism by which they disrupt the synthesis, traffic and function of CFTR, patients characteristic are shown in Figure 8.

Table 1 Genotype characteristic
Classes No. of patients Age (Avg±SEM) Gender (F%) Class/PT%
Class II 4 21.5 ±1.06 13% 10%
Class III 14 20.4 ±0.944 35% 36%
Class IV 15 22 ± 0.73 30% 38%
Class V 4 20 ± 0.353 13% 10%
Class VI 2 23.5 ± 7.5 9% 5%

CFTR mutations gene classified into five classes: class Ⅱ (10%), class Ⅲ (36%), class Ⅳ (38%), class Ⅴ (10%), class Ⅵ (5%) according to the deactivation in the mechanism of the synthesis, traffic and function. The mean age in the CF group was 21 ± 2 years.

image24.png

 

Figure 8 Patients genotypic classification

Table 2 phenotype profile
Symptoms PT No. CF Total ClassⅡ ClassⅢ Class Ⅳ Class Ⅴ Class Ⅵ
Persistent Cough 34 87% 9% 35% 41% 9% 6%
Lung Infections 3 8% 33% 33% 0% 0% 33%
Sputum culture 1 3% 100% 0% 0% 0% 0%
Sinusitis 6 15% 17% 67% 0% 17% 0%
Bowel Irregularities 1 3% 0% 0% 0% 100% 0%

4.2 Human Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) ELISA:

ELISA Sandwich (enzyme-linked immunosorbent assay) used for analyzing the presence of the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR), where 96 wells plate pre-coated with an antibody specific to CFTR. Serum from 39 patients with CF was tested by ELISA technique against 30 healthy control samples. These results indicate that the absorbance values produced from healthy control are higher than CF patients. In Figure 9 we plot the O.D. value of the standard (Y-axis) against the known concentration of the standard (X-axis).

image7.png

Figure 9 Typical Standard Curve for CFTR, Human ELISA

4.3 Neonatal IRT Screening ELISA:

Among a total of 39 adults CF patients screened within the study period, six patients were found positive IRT test (> 60 ng/mL) shown in Figure 10. Also, we found that the CFTR mutations gene class for these patients were class Ⅳ.

image25.png

Figure 10 IRT mean values according to age of the CF patients. A positive IRT test (> 60 ng/mL) values are high in adults CF Patients.

4.4 GC/MS metabolomics

Targeted Metabolomics study based on gas chromatography/ mass spectrometry (GC/MS) was performed to study the metabolic changes in DBS of CF samples. More than 100 features (metabolites) were investigated in DBS samples to see the difference in their levels between control and CF samples ( Figure 11).

The data of all features in the experiment samples were normalized to the internal standard (an isotopic version of the target analyte), and the response (peak area) is normalized to the sample total sum. Box and Kernal density plots show the overall signals before and after normalization, where most of the features areof the same intensity. The data scale was Log-transformed, and the overall distribution scaled by Patro Scaling (mean-centered and divided by the square root of thestandard deviation of each variable) showed the intensities of most of the features within the normal Gaussian distribution as in Figure 12 . Box plot showed only 50 features due to space limitation while density plot depends on samples.

image8.png
Figure 11 The flow work of human metabolome analysis using DBS sampling protocols.

Overview of metabolomic data generation and data analysis. The flowchart used for metabolite extraction, data mining, and metabolite identification is detailed. This illustrates sample preparation, mass spectrometric analysis, peak extraction/identification and compound quantification, and statistical data analysis for biomarker identification and mapping of biomarkers to metabolic pathways.

image9.png
Figure 12 Box plots and kernel density plots before and after normalization.

The boxplots show at most 50 features due to space limit. The density plots are based on all samples. Selected methods: Row-wise normalization: Normalization to the constant sum; Data transformation: Log Normalization; Data scaling: Pareto Scaling.

After normalizing the signal of metabolites, a student t-test and fold change (FC) evaluation was performed on the dataset, and the results were visualized in volcano plot. Volcano plot is a two-dimension univariate analysis approach used for exploratory data analysis, where it combinedFCandt-test analyses. Figure 13 shows features that are above FC (x-axis) and t-test (y-axis) thresholds1.2and 0.05, respectively. The features changing are significantly upregulated or downregulated by ±2folds difference from control and patients are presented in red at the upper left corner. The important features that were significantly changed (p<0.05) and identified by volcano plot are D-Sorbitol ,3-Indoleacetic acid and phenylalanine.

image10.png
Figure 13 Important features selected by volcano plot. With fold change threshold (x-axis) ± 2 and t-tests threshold (y-axis) 0.05. The red circles represent features above the threshold. Note both fold changes and p-values are log10 and log2 transform.

 

Principal Component Analysis (PCA), unsupervised multivariate statistical method, is the most popular approach to gain information to evaluate the reproducibility of the analytical method and the degree of changes between the study groups.PCAexplains the variance in a data set (X) without referring to class labels (Y), where the data can be summarized in theform of scores which is the average of original variables. Figure 14 shows the PCA score plot representing the distribution among healthy control and CF group in two dimensions. The level of separation suggests that biological changes occur in CF group compared to healthy control.

 

 
 
(B)image11.png image26.png(A)

 

Figure 14 Scores plot between the selected PCs.

(A) Scores plot between the selected PCs in Healthy Control group and CF patients group. (B)Scores plot between the selected PCs in a 3D scatter plot.

Partial Least Squares – Discriminant Analysis (PLS-DA) is the other popular method for classification and identification of metabolites. It is a supervised method has the same principle of PCA used to enhance classification performance that uses multivariate regression techniques to extract via linear combination of original variables (X) and the information that can predict the class membership (Y) shown in Figure 15. Variable Importance in Projection(VIP) is a weighted sum of squares of the PLS loadings taking into account the amount of explained Y-variation in each dimension. VIP score is calculated for each component. The plot in Figure 16. shows the VIP score of each feature, the important features are presented in volcano plot, and their relative concentrations in each group. For instance, D-Sorbitol which has a high VIP score was found to have a high concentration in healthy control group but not in CF patients.

image12.png
Figure 15 Scores plot between the selected components by PLS-DA.

The PLS-DA score plot representing the distribution between Healthy control and CF patients in two dimensions.

image13.png

Figure 16 Score plots of variable importance in projection (VIP).

 

The colored boxes on the right indicate the relative concentrations of the corresponding metabolite in each group (1) and (2), Healthy Control and CF patients, respectively.

Orthogonal partial least squares (O-PLS-DA) is an extension of PLS-DA which seeks to maximize the explained variance between groups in a single dimension or the first latent variable (LV), and separatewithin group variance into orthogonal latent variable. The variable loadings and/or coefficient weights from a validated O-PLS-DA model can be used to rank all variables with respect to their performance for discriminating between groups. This can be used part of a dimensional reduction or feature selection task which seek to identify the top predictors for a given model, See Figure 17.

image14.png
(A)
image15.png
(B)
Figure 17 Scores plot and permutation test between the selected components by O-PLS-DA.

(A) The O-PLS-DA score plot between Healthy control (Green group) and CF patients (Red group) into orthogonal variables. (B) A permutation test performed with 200 random permutations in a PLS-DA model showing R2 (blue color) and Q2 (pink color) values from the permuted analysis (left) significantly lower than the corresponding original values (right).

Receiver Operating Characteristic (ROC) metric other popular method to evaluate classifier output quality and to characterize threshold independent performance of the prediction models. ROC curves typically feature true positive rate on the Y axis, and false positive rate on the X axis. This means that the top left corner of the plot is the “ideal” point – a false positive rate of zero, and a true positive rate of one, see Figure 18 . This is not very realistic, but it does mean that a larger area under the curve (AUC) is usually better. The “steepness” of ROC curves is also important, it is ideal to maximize the true positive rate while minimizing the false positive rate. Figure 19, shows the evaluation measure for multi-class classification (macro-averaging), which gives equal weight to the classification of each label.

image16.png
Figure 18 Receiver Operating Characteristic curve (ROC) plot all the six classifiers.

ROC curve shows that the ten analytes achieved the maximum value for area under Receivers Operating Characteristic curve closely followed by five and twenty analytes. Three analytes had the least ROC.

image27.png image28.png
(A) (B)
image29.png image30.png
(C) (D)
Figure 19 Receiver Operating Characteristic curve (ROC) plot.

The blue area corresponds to the Area Under the curve of the Receiver Operating Characteristic (AUROC). The dashed line in the diagonal we present the ROC curve of a random predictor: (A). (ROC) curves with AUC values for D-Sorbitol which the best biomarker for CF followed by(B) for Phenylalanine, (C) for 3-Indoleacetic acid and(D)for Arachidic acid.

4.5 LCMS/MS metabolomics

Targeted Metabolomics study based on liquid chromatography-tandem mass spectrometry (LC/MSMS) was performed to study the metabolic changes in DBS of CF samples. More than 250 features (metabolites) were investigated in DBS samples to see the difference in their levels between control and CF samples.

The data of all features in the experiment samples were normalized to the internal standard (an isotopic version of the target analyte), and the response (peak area) is normalized to the sample total sum. Box and Kernal density plots show the overall signals before and after normalization, where most of the features are of the same intensity. The data scale was Log-transformed, and the overall distribution scaled by Patro Scaling (mean-centered and divided by the square root of the standard deviation of each variable) showed the intensities of most of the features within the normal Gaussian distribution as in Figure 20. Box plot showed only 50 features due to space limitation while density plot depends on samples.

image17.png
Figure 20 Box plots and kernel density plots before and after normalization.

The boxplots show at most 50 features due to space limit. The density plots are based on all samples. Selected methods: Row-wise normalization: Normalization to the constant sum; Data transformation: Log Normalization; Data scaling: Pareto Scaling.

After normalizing the signal of metabolites, a student t-test and fold change (FC) evaluation was performed on the dataset, and the results were visualized in volcano plot. Volcano plot is a two-dimension univariate analysis approach used for exploratory data analysis, where it combined FC and t-test analyses. Figure 21, shows features that are above FC (x-axis) and t-test (y-axis) thresholds1.2and 0.05, respectively. The features changing are significantly upregulated or downregulated by ±2 folds difference from control and patients are presented in red at the upper left corner. The important features that were significantly changed (p<0.05) and identified by volcano plot are Mannitol, Glutamic acid and Sorbitol.

image18.png
Figure 21 Important features selected by volcano plot.

With fold change threshold (x-axis) ± 2 and t-tests threshold (y-axis) 0.05. The red circles represent features above the threshold. Note both fold changes and p-values are log10 and log2 transformed, respectively. The further its position away from the (0,0), the more significant features.

Principal Component Analysis (PCA), unsupervised multivariate statistical method, is the most popular approach to gain information to evaluate the reproducibility of the analytical method and the degree of changes between the study groups. PCA explains the variance in a data set (X) without referring to class labels (Y), where the data can be summarized in the form of scores which is the average of original variables. Figure 22 (A), shows the PCA score plot representing the distribution among healthy control and CF group in two dimensions. The level of separation suggests that biological changes occur in CF group compared to healthy control. The corresponding loading plot, used to identify biomarkers, is shown in Figure 22 (B), where the features furthest from the origin are changed in their level significantly between the two groups.

image31.png
(A)
image32.png
(B)
Figure 22 Scores plot between the selected PCs.

(A) Scores plot between the selected PCs in Healthy Control group and CF patients group. (B) Loadings plot for the selected PCs.

Partial Least Squares – Discriminant Analysis (PLS-DA) is the other popular method for classification and identification of metabolites. It is a supervised method has the same principle f PCA used to enhance classification performance that uses multivariate regression techniques to extract via linear combination of original variables (X) and the information that can predict the class membership (Y), see Figure 23. Variable Importance in Projection(VIP) is a weighted sum of squares of the PLS loadings taking into account the amount of explained Y-variation in each dimension. VIP score is calculated for each component. The plot in Figure 24. shows the VIP score of each feature, the important features are presented in volcano plot, and their relative concentrations in each group. For instance, Mannitol which has a high VIP score was found to have a high concentration in healthy control group but not in CF patients.

image33.png
(A)
image34.png
(B)
Figure 23 Scores plot between the selected components by PLS-DA.

(A) The PLS-DA score plot representing the distribution between Healthy control and CF patients in two dimensions. (B) Scores plot between the selected components in a 3D scatter plot.

image19.png
Figure 24 Score plots of variable importance in projection (VIP).

The colored boxes on the right indicate the relative concentrations of the corresponding metabolite in each group (1) and (2), Healthy Control and CF patients, respectively.

Orthogonal partial least squares (O-PLS-DA) is an extension of PLS-DA which seeks to maximize the explained variance between groups in a single dimension or the first latent variable (LV), and separate within group variance into orthogonal latent variable. The variable loadings and/or coefficient weights from a validated O-PLS-DA model can be used to rank all variables with respect to their performance for discriminating between groups. This can be used part of a dimensional reduction or feature selection task which seek to identify the top predictors for a given model, See Figure 25.

image35.png
(A)
image36.png
(B)
Figure 25 Scores plot and permutation test between the selected components by O-PLS-DA.

(A) The O-PLS-DA score plot between Healthy control (Green group) and CF patients (Red group) into orthogonal variables. (B) A permutation test performed with 200 random permutations in a PLS-DA model showing R2 (blue color) and Q2 (pink color) values from the permuted analysis (left) significantly lower than the corresponding original values (right).

Receiver Operating Characteristic (ROC) metric other popular method to evaluate classifier output quality and to characterize threshold independent performance of the prediction models. ROC curves typically feature true positive rate on the Y axis, and false positive rate on the X axis. This means that the top left corner of the plot is the “ideal” point – a false positive rate of zero, and a true positive rate of one, see Figure 26 . This is not very realistic, but it does mean that a larger area under the curve (AUC) is usually better. The “steepness” of ROC curves is also important, it is ideal to maximize the true positive rate while minimizing the false positive rate. Figure 27, shows the evaluation measure for multi-class classification (macro-averaging), which gives equal weight to the classification of each label.

image37.png
Figure 26 Receiver Operating Characteristic curve (ROC) plot all the six classifiers.

ROC curve shows that the all analytes achieved the maximum value for area under Receivers Operating Characteristic curve.

image38.png image39.png
(A) (B)
image40.png image41.png
(A) (B)
Figure 27 Receiver Operating Characteristic curve (ROC) plot.

The blue area corresponds to the Area Under the curve of the Receiver Operating Characteristic (AUROC). The dashed line in the diagonal we present the ROC curve of a random predictor: (A). (ROC) curves with AUC values for mannitol which the best biomarker for CF followed by(B) for Glutamic acid, (C) for Sorbitol and(D)for Fructose1,6 Bisphosphate.

A heat map of the most significant features out from t-test is shown in Figure 28. Each feature appears to have different relative concentration in each group.

image20.png
Figure 28 A heat map of significant top features from t-test.

Where the group are; control (Yellow), and CF classes, class Ⅱ (Red), class Ⅲ (Green), class Ⅳ(Blue), class Ⅴ(Light blue) and class Ⅵ(Pink).

 

The Pathway Analysis identifies the most relevant pathways involved under the study condition. Many metabolic pathways were found to involve more than one metabolite from the significant metabolites in this study. The pathway that contains more metabolites (impact of the pathway) from the statistically significant list that has been generated from the t-test, and more statistically significant enriched pathways are considered the most affected pathway under experiment condition. Fructose and mannose metabolism, D-Glutamine and D-glutamate metabolism, pyruvate metabolism, Alanine, aspartate and glutamate metabolism, Nitrogen metabolism, are the most significant enriched pathways P<0.05 with different impacts as shown in Figure 29 and Table 3.

image21.png

 

Figure 29 of Pathway Analysis.

The most relevant pathways involved under experiment conditions.

Table 3: Detailed results from the pathway analysis

 

image42.jpg

Chapter 5: Discussion

Chapter 6: Conclusion and Future Directions

References

1. Banjar HH. Cystic fibrosis: presentation of other diseases, the experience in Saudi Arabia. Journal of Cystic Fibrosis. 2003 Sep 1 ;2(3):155-9.

2. Hudock KM, Clancy JP. An Update on New and Emerging Therapies for Cystic Fibrosis. Expert opinion on emerging drugs. 2017 Oct 2(just-accepted).

3. Bowen SJ, Hull J. The basic science of cystic fibrosis. Paediatrics and Child Health. 2015 Apr 1;25(4):159-64.

4. Nobakht M. Gh BF, Aliannejad R, Rezaei-Tavirani M, Taheri S, Oskouie AA. The metabolomics of airway diseases, including COPD, asthma and cystic fibrosis. Biomarkers. 2015 Jan 2;20(1):5-16.

5. Rosenstein B J, Zeitlin P L. Cystic fibrosis. Lancet 1999.

6. Elborn J S. Cystic fibrosis. Lancet 2016.

7. Dawson KP, Frossard PM. Cystic fibrosis in the Middle East: the historical perspective. Annals of Saudi medicine. 2000 Jan;20(1):20.

8. Ramsey KA, Schultz A, Stick SM. Biomarkers in pediatric cystic fibrosis lung disease. Paediatric respiratory reviews. 2015 Sep 1;16(4):213-8.

9. Gierok P, Harms M, Methling K, Hochgräfe F, Lalk M. Staphylococcus aureus Infection Reduces Nutrition Uptake and Nucleotide Biosynthesis in a Human Airway Epithelial Cell Line. Metabolites. 2016 Nov 9;6(4):41.

10. Quinn RA, Lim YW, Mak TD, Whiteson K, Furlan M, Conrad D, Rohwer F, Dorrestein P. Metabolomics of pulmonary exacerbations reveals the personalized nature of cystic fibrosis disease. PeerJ. 2016 Aug 11;4:e2174.

11. Maiuri L, Raia V, Kroemer G. Strategies for the etiological therapy of cystic fibrosis. Cell death and differentiation. 2017 Nov;24(11):1825.

12. Esther CR, Turkovic L, Rosenow T, Muhlebach MS, Boucher RC, Ranganathan S, Stick SM. Metabolomic biomarkers predictive of early structural lung disease in cystic fibrosis. European Respiratory Journal. 2016 Dec 1;48(6):1612-21.

13. Ferrari E, Monzani R, Villella VR, Esposito S, Saluzzo F, Rossin F, D’eletto M, Tosco A, De Gregorio F, Izzo V, Maiuri MC. Cysteamine re-establishes the clearance of Pseudomonas aeruginosa by macrophages bearing the cystic fibrosis-relevant F508del-CFTR mutation. Cell death & disease. 2017 Jan;8(1):e2544.

14. Wetmore DR, Joseloff E, Pilewski J, Lee DP, Lawton KA, Mitchell MW, Milburn MV, Ryals JA, Guo L. Metabolomic profiling reveals biochemical pathways and biomarkers associated with pathogenesis in cystic fibrosis cells. Journal of Biological Chemistry. 2010 Oct 1;285(40):30516-22.

15. Yang J, Eiserich JP, Cross CE, Morrissey BM, Hammock BD. Metabolomic profiling of regulatory lipid mediators in sputum from adult cystic fibrosis patients. Free Radical Biology and Medicine. 2012 Jul 1;53(1):160-71.

16. Muhlebach MS, Sha W. Lessons learned from metabolomics in cystic fibrosis. Molecular and cellular pediatrics. 2015 Dec 1;2(1):9.

17. Roxo-Rosa M, Davezac N, Bensalem N, Majumder M, Heda GD, Simas A, Penque D, Amaral MD, Lukacs GL, Edelman A. Proteomics techniques for cystic fibrosis research. Journal of Cystic Fibrosis. 2004 Aug 1;3:85-9.

18. Najada AS, Dahabreh MM. Clinical profile of cystic fibrosis. Atypical presentation. Saudi medical journal. 2010;31(2):185-8.

19. Nazer H, Riff E, Sakati N, Mathew R, Majeed-Saidan MA, Harfi H. Cystic fibrosis in Saudi Arabia. European journal of pediatrics. 1989 Jan 1;148(4):330-2.

20. Balch WE. Introduction to section II: comics in the biology of cystic fibrosis. InCystic Fibrosis 2011 (pp. 189-191). Humana Press.

21. Laguna TA, Reilly CS, Williams CB, Welchlin C, Wendt CH. Metabolomics analysis identifies novel plasma biomarkers of cystic fibrosis pulmonary exacerbation. Pediatric Pulmonology. 2015 Sep 1 ;50(9):869-77.

22. Raju SV, Lin VY, Liu L, McNicholas CM, Karki S, Sloane PA, Tang L, Jackson PL, Wang W, Wilson L, Macon KJ. The cystic fibrosis transmembrane conductance regulator potentiator ivacaftor augments mucociliary clearance abrogating cystic fibrosis transmembrane conductance regulator inhibition by cigarette smoke. American journal of respiratory cell and molecular biology. 2017 Jan;56 (1):99-108.

23. Wandro S, Carmody L, Gallagher T, LiPuma JJ, Whiteson K. Making It Last: Storage Time and Temperature Have Differential Impacts on Metabolite Profiles of Airway Samples from Cystic Fibrosis Patients. Systems. 2017 Dec 26;2(6):e00100-17.

24. Joseloff E, Sha W, Bell SC, Wetmore DR, Lawton KA, Milburn MV, Ryals JA, Guo L, Muhlebach MS. Serum metabolomics indicate altered cellular energy metabolism in children with cystic fibrosis. Pediatric Pulmonology. 2014 May 1 ;49 (5):463-72.

25. (25Munck, A., et al., Implementation of the French nationwide cystic fibrosis newborn screening program. J Pediatr, 2008. 153(2): p. 228-33, 233 e1.

26. Banjar, H., Geographic distribution of cystic fibrosis transmembrane regulator gene mutations in Saudi Arabia. East Mediterr Health J, 1999. 5(6): p. 1230-5.

27. Lopes-Pacheco M. CFTR modulators: shedding light on precision medicine for cystic fibrosis. Frontiers in pharmacology. 2016 Sep 5;7:275.

28. Fisher OM, Oberkofler CE, Raptis DA, Soll C, Béchir M, Schiesser M, Graf R. Pancreatic stone protein (PSP) and pancreatitis-associated protein (PAP): a protocol of a cohort study on the diagnostic efficacy and prognostic value of PSP and PAP as postoperative markers of septic complications in patients undergoing abdominal surgery (PSP study). BMJ open. 2014 Mar 1;4(3):e004914.

Get professional assignment help cheaply

Are you busy and do not have time to handle your assignment? Are you scared that your paper will not make the grade? Do you have responsibilities that may hinder you from turning in your assignment on time? Are you tired and can barely handle your assignment? Are your grades inconsistent?

Whichever your reason may is, it is valid! You can get professional academic help from our service at affordable rates. We have a team of professional academic writers who can handle all your assignments.

Our essay writers are graduates with diplomas, bachelor, masters, Ph.D., and doctorate degrees in various subjects. The minimum requirement to be an essay writer with our essay writing service is to have a college diploma. When assigning your order, we match the paper subject with the area of specialization of the writer.

Why choose our academic writing service?

  • Plagiarism free papers
  • Timely delivery
  • Any deadline
  • Skilled, Experienced Native English Writers
  • Subject-relevant academic writer
  • Adherence to paper instructions
  • Ability to tackle bulk assignments
  • Reasonable prices
  • 24/7 Customer Support
  • Get superb grades consistently

 

 

 

 

 

 

 

Place your order
(550 words)

Approximate price: $22

Calculate the price of your order

550 words
We'll send you the first draft for approval by September 11, 2018 at 10:52 AM
Total price:
$26
The price is based on these factors:
Academic level
Number of pages
Urgency
Basic features
  • Free title page and bibliography
  • Unlimited revisions
  • Plagiarism-free guarantee
  • Money-back guarantee
  • 24/7 support
On-demand options
  • Writer’s samples
  • Part-by-part delivery
  • Overnight delivery
  • Copies of used sources
  • Expert Proofreading
Paper format
  • 275 words per page
  • 12 pt Arial/Times New Roman
  • Double line spacing
  • Any citation style (APA, MLA, Chicago/Turabian, Harvard)

Our guarantees

Delivering a high-quality product at a reasonable price is not enough anymore.
That’s why we have developed 5 beneficial guarantees that will make your experience with our service enjoyable, easy, and safe.

Money-back guarantee

You have to be 100% sure of the quality of your product to give a money-back guarantee. This describes us perfectly. Make sure that this guarantee is totally transparent.

Read more

Zero-plagiarism guarantee

Each paper is composed from scratch, according to your instructions. It is then checked by our plagiarism-detection software. There is no gap where plagiarism could squeeze in.

Read more

Free-revision policy

Thanks to our free revisions, there is no way for you to be unsatisfied. We will work on your paper until you are completely happy with the result.

Read more

Privacy policy

Your email is safe, as we store it according to international data protection rules. Your bank details are secure, as we use only reliable payment systems.

Read more

Fair-cooperation guarantee

By sending us your money, you buy the service we provide. Check out our terms and conditions if you prefer business talks to be laid out in official language.

Read more
1
You can contact our live agent via WhatsApp! Via +1 817 953 0426

Feel free to ask questions, clarifications, or discounts available when placing an order.
  +1 (301) 710 0002           + 44 161 818 7126           [email protected]
  + 44 161 818 7126         [email protected]