Bioanalysis of Protein Biopharmaceuticals
by Lucia Baljeu-Neuman
MSc Chemistry
Analytical Sciences
Master Thesis
Bioanalysis of Protein Biopharmaceuticals
Based on Signature Peptides
by Liquid Chromatography-Tandem Mass Spectrometry
Somatropin Quantification Using Bovine Fetuin as Internal Standard
by
Lucia Baljeu-Neuman
January 2016
Research performed between 15
thof January and 15
thof August 2015
Daily Supervisor: Examiner:
dr Erik Baltussen dr Wim Th. Kok
1 ACKNOWLEDGEMENTS
I would like to thank Henk Lingeman, PhD, for recommending WIL Research Europe to conduct my research.
My sincere thanks goes to Theo Noij, PhD, who provided me the opportunity to join WIL Research Europe in Den Bosch as intern and to Erik Baltussen, PhD, who kindly guided and advised me during the whole project.
I would also like to thank Daphne de Ruijter and the Bioanalysis team for their precious assistance and constant support in the laboratory. In addition, I would like to thank the personnel of WIL Research Europe for being so collegial and friendly.
My sincere thanks also goes to Chris de Koster, PhD, for guiding my first steps in protein research and to Wim Kok, PhD, for taking over the guidance of my research and helping to finalise this thesis.
2 SUMMARY
Protein biopharmaceuticals are large, heterogeneous molecules produced via recombinant DNA technology. Protein biopharmaceuticals can be divided into therapeutic proteins and monoclonal antibodies. Some of these are used to treat diseases which affect a large number of people like cancer, diabetes and anaemia, while others are used in the treatment of rare diseases like Pompe disease or Hunter’s syndrome.
Nowadays, the number of approved protein biopharmaceuticals is increasing rapidly. Moreover, the patents for some of the protein biopharmaceuticals have expired and others will expire in the next years, therefore a large increase in the number of biosimilars is expected in the coming years.
WIL Research Europe in Den Bosch, The Netherlands, is an independent laboratory which conducts a variety of toxicological, (bio)analytical chemistry, metabolism and pharmaceutical studies. Although protein analysis is not part of the current portfolio, there is considerable interest to implement this technology. Therefore, the aim of this research project was to investigate the possibility of setting up a generic method for protein analysis based on surrogate peptides using Liquid Chromatography coupled to tandem Mass Spectrometry (LC-MS/MS). Surrogate peptides, also called signature peptides, are the result of the enzymatic digestion of targeted proteins. These proteins are first denaturated, reduced and alkylated to enhance the effectiveness of the proteolysis. Subsequently, these peptides are separated using liquid chromatography and identified by tandem mass spectrometry.
In this thesis a sample preparation protocol is presented, without a purification step of the protein or surrogate peptides, that can be used as a generic sample pre-treatment for LC-MS/MS protein biopharmaceuticals analysis based on surrogate peptide(s). This study has focused on the analysis of transferrin, somatropin and bovine fetuin in human plasma to demonstrate the applicability of the generic protocol for a select number of proteins. Furthermore, the generic sample preparation has been optimised and made cost-effective. Finally, the LC-MS/MS method for somatropin quantification using bovine fetuin as internal standard has been validated according to the principles of bioanalytical method validation of the Food and Drug Administration (FDA) and European Medicine Agency (EMA).
3 CONTENTS ACKNOWLEDGEMENTS ... 1 SUMMARY ... 2 CONTENTS ... 3 1. INTRODUCTION ... 4 1.1. Proteins ... 4
1.1.1. Structure and Functions of Proteins ... 4
1.1.2. Post-translational Modifications ... 5
1.1.3. Proteins in Clinical Diagnostics ... 6
1.2. Biopharmaceuticals ... 6
1.2.1. Description of Protein Biopharmaceuticals ... 7
1.2.2. Characteristics of Protein Biopharmaceuticals ... 7
1.2.3. Protein Biopharmaceuticals in Pharmaceutical Industry ... 8
1.3. Protein Analysis ... 9
1.3.1. Intact Protein Analysis ... 9
1.3.2. Peptide Analysis ... 9
1.3.3. Method Validation ... 12
2. EXPERIMENTAL DATA ... 13
2.1. Chemicals and Materials ... 13
2.2. Preparation of the Stock Solutions, Calibration Standards and Quality Controls ... 13
2.3. Equipment and Software ... 14
2.4. Sample preparation ... 14
2.4.1. Protein Denaturation ... 14
2.4.2. Protein Reduction and Alkylation... 15
2.4.3. Protein Proteolysis ... 15
2.5. Surrogate Peptides Identification and Selection ... 15
2.5.1. Surrogate Peptides Identification by Direct MS Infusion ... 15
2.5.2. Surrogate Peptides Identification by Importing MS-settings from Skyline ... 18
2.5.3. Signature Peptide Selection ... 18
2.6. LC-MS/MS Method ... 18
3. RESULTS ... 21
3.1. Protein Denaturation Using Five Different Denaturation Agents ... 21
3.2. Protein Digestion Using Different Qualities and Quantities of Trypsin ... 21
3.3. Protein Digestion at Different Temperature and Incubation Time ... 22
3.4. Surrogate peptides identification and selection ... 24
3.5. Method Validation ... 25
4. DISCUSSION ... 26
5. CONCLUSIONS ... 28
LIST OF ABBREVIATIONS ... 29
APPENDIX 1. HUMAN SERUM TRANSFERRIN - GENERAL FACTS, FASTA SEQUENCE AND STRUCTURE ... 30
APPENDIX 2. (RECOMBINANT) HUMAN GROWTH HORMONE - GENERAL FACTS, FASTA SEQUENCE AND STRUCTURE ... 31
APPENDIX 3. BOVINE FETUIN - GENERAL FACTS, FASTA SEQUENCE AND STRUCTURE ... 32
APPENDIX 4. METHOD VALIDATION: SOMATROPIN QUANTIFICATION USING BOVINE FETUIN AS INTERNAL STANDARD AND LC-MS/MS ANALYSIS ... 33
4 1. INTRODUCTION
1.1. Proteins
All living cells and viruses contain proteins.1 Proteins are macromolecules produced inside the ribosomes of living cells. Proteins are made of amino acids linked through amide bonds as coded in the DNA. There are thousands of different proteins in the human body.2 For example, structural proteins, enzymes, hormones, antibodies, receptors or channels in the cell membrane, transporters of vitamins, metal ions or drugs. Proteins have a vital role in nearly all biological processes.2,3 The structure, functions and post-translational modifications of proteins are described below, followed by a brief presentation of the role of proteins in clinical diagnostics.
1.1.1. Structure and Functions of Proteins
The structural entities of proteins are α-amino acids. These are amino acids which have the amino-group and the carboxyl-group linked at the same carbon atom, known as the α-carbon atom. Amino acids form peptide bonds by covalently linking the α-amino group of one amino acid with the α-carboxyl-group of the next amino acid. The formed chain is known as a peptide. Proteins are made of one or several peptides. The difference between proteins and peptides is that proteins wield certain functions, as a result of their three-dimensional structure.1 Many proteins contain besides amino acids other compounds (prosthetic groups), such as carbohydrates, lipids or metal complexes. These proteins are known as conjugated proteins.1,2
The amino acid sequence is called the primary structure of protein. This arrangement gives the protein its uniqueness. Therefore, even a single amino acid modification can change the characteristics of a protein. For example, by replacing glutamic acid with valine in the hemoglobin molecule, sickle-cell anemia, an inherited disorder of the blood cell occurs.1 Glutamic acid is a polar amino acid while valine is a nonpolar, hydrophobic amino acid. Therefore, hemoglobin’s properties, shape and function are altered. This abnormal hemoglobin, known as hemoglobin S, congregates and forms long, rigid and sickle shaped cells. These can block the capillaries, producing severe pain, physical weakness and damage to the vital organs.1
Once the primary chain is formed, it spontaneously folds and twists due to the hydrogen-bond interactions. This two-dimensional arrangement of the backbone atoms is known as the secondary structure of a protein. The most common secondary structure arrangements are the α-helix and β-pleated sheet. The interactions of the side chains, namely hydrogen-bonds, ionic-bonds, disulphide-bonds and hydrophobic interactions, give a protein its characteristic three-dimensional shape, the tertiary structure. Additionally, proteins may have a quaternary structure. This fourth level of structure involves the association of two or more subunits into a larger assembly. The folded structure, also known as the native state, is commonly the most stable arrangement of proteins.1
Proteins can be classified, based on their shape, as globular protein or fibrous proteins. Globular proteins have a compact spherical shape. The amino acids with nonpolar side groups are situated inside the folded structure, while the amino acids with polar side groups are placed on the outer surface. These proteins, such as, enzymes, hormones or plasma proteins, are water-soluble and usually mobile within the cell. Their function is mainly regulatory. Examples include transport mechanisms, storing oxygen, catalysing biochemical reactions, regulating the metabolism and protecting against infections. Globular proteins are stable over a narrow pH and temperature range. In case of higher temperature or extreme pH, proteins can be denaturated. Protein denaturation can be reversible or irreversible. A denaturated protein loses its three- and two-dimensional structure and thus its function(s).
5
Fibrous proteins, which include collagen, keratin, fibrinogen, troponin and myosin, have a long, fibrillar shape, are water-insoluble and have mostly structural functions. These proteins are more stable when changes in pH or temperature occur. Fibrous proteins provide the structural integrity and strength to cells and tissues.1,2
The specific function of a protein is correlated with its structure. As mentioned above, changes in the primary structure of a protein alter its biological activity. Moreover, modifications in protein structure may also occur during the folding process. It is assumed that the mad cow disease (Bovine Spongiform Encephalopathy) and Creutzfeldt–Jakob disease, the human equivalent, are caused by a protein misfolding. The prion-related protein (PrP), which occurs in the cell membranes of nerve tissues, has a large percentage of α-helix in its structure, while the misfolded form has a large percentage of β-pleated sheets. These β-pleated sheets aggregate forming insoluble clusters (plaques). The plaques are usually formed in brain tissue and have severe consequences, such as nerve degeneration, mental deterioration, dementia, and even death.4
1.1.2. Post-translational Modifications
Post-translational modification (PTM) is a generic name for protein modification that takes place after its translation by ribosomes.5 As mentioned before, in paragraph 1.1.1, proteins are synthesized from amino acids. There are twenty primary amino acids used in protein synthesis. However, due to the PTMs, circa 140 amino acids and amino acids derivatives have been found in the structure of different proteins.5,6 PTMs can occur by peptide bond cleavage or by addition of one or more functional groups. These modifications can take place at the amino terminus, carboxyl terminus or at the amino acid level. Some modifications can be spontaneous (e.g. deamidation, cysteinylation), while other are catalyzed by enzymes (e.g. glycosylation, phosphorylation).6,7 Moreover, PTMs can be reversible or irreversible, rapid (e.g. phosphorylation – dephosphorylation, the on-off switch of many enzymes) or slow (e.g. glycosylation), independent or inter-dependent (e.g. ubiquitylation, protein degradation, follows phosphorylation).5,6,7 Other examples of post-translational modifications are methylation, acetylation, acylation, sumoylation, glycation, and oxidation. It is estimated that more than 300 various PTMs are involved in the cellular activity.7
PTMs regulate protein structure as well as its functions. One protein can undergo various modifications in order to fulfill its functional role.7 These modifications take place at specific sites. For example, phosphorylation, the addition of a phosphate group, takes place on serine, tyrosine or threonine residues while glycosylation, the addition of a glycan moiety, can occur on asparagine (N-glycosylation), serine or threonine (O-glycosylation). Moreover, N-glycosylation requires asparagine (N) residue to be in the vicinity of threonine (T) or serine (S) residue, as in the sequence –N–X–T/S– and O-glycosylation requires the following sequence –R–X–Y–Z–S(Z)–S/T, where R is arginine and X, Y, Z may be any amino acid residue.5 It is estimated that 50-90% of human proteins are modified after translation.6
PTMs regulate protein activity. It can be proven that in case of PTM deviations, these activities are also disturbed. For example, Congenital Disorders of Glycosylation (CDG) are inherited metabolic disorders that are caused by glycosylation defects. By means of underglycosylation or by the presence of abnormal glycans the whole organ system and cellular functions are affected.8
6 1.1.3. Proteins in Clinical Diagnostics
In clinical diagnostics, the study of proteins is used for early detection, diagnostic and therapeutic purposes. By studying the expression of proteins in living cells, reliable data on the health of patients may be revealed. For example, newborn screening allows the detection of phenylketonuria (PKU) disease before its clinical symptoms. PKU, a metabolic disease, is caused by the deficiency of a specific enzyme, namely phenylalanine hydroxylase. This leads to the accumulation of phenylalanine in blood and tissues and untreated leads to mental retardation.9 Other proteins are produced only in case of diseases and are used as biomarkers for diagnosis. Furthermore, there is a variation of protein concentration between health and disease.2
Plasma is one of the most used biological fluids in clinical diagnostics. The concentration of proteins in plasma is high, approximately 75 mg/mL.3 Albumin covers approximately 50% of the plasma proteins, and together with the other abundant proteins, which include immunoglobulins, fibrinogen, transferrin and apolipoproteins, represents circa 95% of the total protein content.3 Consequently, the variation in protein abundance might be up to 10 orders of magnitude.3,10 The most abundant proteins may mask the least abundant proteins making their detection problematic.3 In addition to plasma, proteins of clinical significance are measured in serum, urine, saliva, cerebrospinal fluid, amniotic fluid or feces.2
For clinical diagnostics various analytical methods are used. These can be screening tests, used to identify a disease before its symptoms occur, quantitative tests, used for diagnosis confirmation and for monitoring treatment, or specific tests, used in the diagnosis of specific disorders.2 The utilized techniques can be categorised in immunochemical, electrophoretic, spectrophotometric and mass spectrometric.2 The analytical methods used for routine testing have to be relatively simple and low-cost, fast and robust. More information about protein analysis is given in section 1.3.
1.2. Biopharmaceuticals
Biopharmaceuticals, also known as biologicals or biological medicines, are medicines containing biological material, such as proteins, DNA or RNA.11 Different types of cells or organisms may be used, namely bacteria, yeasts, plants or genetically modified animals.12,13 Their selection is based generally on production costs and on the possibility to induce specific protein modifications. For example, the glycosylation pattern is not the same in different biological systems and bacteria do not produce glycosylation.12
The first biologic approved by US Food and Drug Administration (FDA) was human insulin (Humulin) in 1982. This has been produced in the Escherichia coli bacteria, via recombinant DNA technology.11 Meanwhile, hundreds of different biopharmaceuticals have been authorised for use.14
Some biopharmaceuticals are used to treat high incidence diseases like cancer, diabetes and anaemia. Other biopharmaceuticals, so-called orphan drugs, are used in the treatment of rare diseases which affect only a small number of the population, such as Pompe disease or Hunter’s syndrome.11,14 The biological materials included in biopharmaceuticals
can be proteins, carbohydrates or nucleic acids.13 The research described in this thesis focuses on protein based biopharmaceuticals.
7 1.2.1. Description of Protein Biopharmaceuticals
Protein biopharmaceuticals are not necessarily novel medicines. For example, insulin was extracted from animal sources in the early 1920s and used to treat diabetes.13 Nowadays, the term ‘protein biopharmaceuticals’ refers generally to recombinant proteins. Protein biopharmaceuticals may be classified into therapeutic proteins or monoclonal antibodies (mAbs).13,14 Their activity can be exerted by different mechanisms. For example, by binding non-covalently to a target, like mAbs, by affecting covalent binding, like enzymes, or by non-specific interactions, like serum albumin.15 Furthermore, protein therapeutics can replace deficient or abnormal proteins, enhance an existing process, provide novel function or activity, target a specific molecule or organism or deliver other compounds (e.g. cytotoxic drug, radioactive nuclide) to the target.3,15
During the last few years, the performance of protein biopharmaceuticals has been improved. For example, the 1989 original version of Epoietin (EPO), used in the treatment of anaemia in kidney dialysis patients, has been modified creating the second (2001) and the third generation EPO (2007). The second generation has two extra N-glycosylation sites and the third generation is conjugated to polyethylene glycol (PEG). These modifications have enhanced the half-life and biological activity of EPO.14 Furthermore, to reduce immunogenicity and to increase efficacy of the biopharmaceuticals, chimeric, humanised or fully human gene sequences have replaced the purely murine sequences.3,13,14
1.2.2. Characteristics of Protein Biopharmaceuticals
Protein biopharmaceuticals are complex macromolecules. As a result, their production, purification and long-term storage is challenging. One reason is that by cloning one protein a substantial number of species may be formed. Due to the co- and post-translational modifications which are taking place in the host cell, beside the expected cloned protein, other products are also produced. Therefore, exact copies cannot be created and batch-to-batch differences are generally present.14
Compared with small molecule drugs, biopharmaceutical production is more complex. In the case of small molecules, it is possible to produce exact copies of the original, so-called generic drugs or generics. In the case of biopharmaceuticals, it is not possible to produce identical products, but similar ones, i.e. biosimilars. In order to be accepted as a biosimilar medicine, it needs to have the same safety and efficacy profile as the reference product.11 One example which illustrates how difficult the production of biosimilars is, is presented further on. Genzyme, one of the leading biotech companies, produces acid-α-glucosidase, an enzyme used in the treatment of Pompe disease, and commercializes it under the name Myozyme. To scale-up the production of Myozyme, the capacity of the bioreactor had been increased from 160 L to 2000 L. The glycosylation of the end product has been so different that US FDA did not approve it as being biosimilar. For this product the company had to apply for a new Biologics License Application (BLA). The new biological has been commercialized under the name Lumizyme.14 Another reason which makes working with biopharmaceuticals challenging is that these can undergo undesirable modifications during production, purification and depositing. These modifications can be enzymatic and/or chemical. Regarding product safety and efficacy, the alterations might be critical and non-critical.14
Protein biopharmaceuticals have several advantages over small-molecule drugs. For example, protein biopharmaceuticals act highly specific, interfere less with other biological processes, therefore less side effects are caused. Moreover, therapeutic proteins can provide replacement for absent or dysfunctional proteins and are better tolerated because of the similarity with own proteins.12 In addition, protein biopharmaceuticals have a relatively long serum half-life, therefore they do not have to be administrated daily.13
8
Protein biopharmaceuticals are strictly examined with regard to purity, molecular identity (primary sequence, PTMs, chemical modifications), structure (cysteine bridges, protein folding), quantity, physicochemical properties (thermal stability, solubility, aggregation, degradation), activity (receptor binding, cell-based activity, enzyme activity), pharmacokinetics (ADME, serum half-life concentration) and pharmacodynamics (toxicity, effectiveness). Currently, chromatography and mass spectrometry are the leading methods for protein biopharmaceutical analysis. However, these methods are complemented with other techniques such as capillary electrophoresis, isoelectric focusing, nuclear magnetic resonance, analytical ultracentrifugation, asymmetric flow field fractionation, differential scanning calorimetry, X-ray crystallography, circular dichroism, fluorescence and FTIR spectroscopy. The analysis of biopharmaceuticals can be performed at the protein level (intact protein or large fragments), peptide level, glycan level or at the amino acids level.14
1.2.3. Protein Biopharmaceuticals in Pharmaceutical Industry
The number of approved protein biopharmaceuticals is increasing rapidly. Nowadays, the protein biopharmaceuticals market is evaluated at about 20% of the worldwide pharmaceutical market. Moreover, it is expected that biologicals will cover more than 50% of the new approved drugs in the next ten years.14 Most of the biopharmaceuticals on the market are recombinant proteins. In addition, newly engineered proteins are in development. Among these, multi-specific fusion proteins, proteins with improved pharmacokinetics, brain penetrant antibodies, antibody mixtures and antibody-drug conjugates (ADC).14
In 2013, the three best-selling drugs were adalimumab (Humira), infliximab (Remicade) and rituximab (Rituxan/MabThera), monoclonal antibodies (mAbs) used to treat arthritis.16 Also, from the top 8 best-selling drugs, in the same year, 7 were biopharmaceuticals and 8 of the top 15 were biologicals as well. Among them, the monoclonal antibodies Trastuzumab (Herceptin), for treatment of breast cancer, and bevacizumab (Avastin) for colorectal, lung and kidney cancer.16 The best-selling therapeutic proteins had been insulin glargine (Lantus) for diabetes and pegfilgrastim (Neulasta) for neutropenia (granulocyte disorder).16 The complete list of these best-selling protein biopharmaceuticals is given in Table 1. The patents for some of the biologics have expired and others will expire in a few years, therefore a large increase in the number of biosimilars is expected in the coming years.14
Table 1
Top 8 blockbuster biologicals in 2013 (reproduced from 16)
Brand
name Active ingredient Type Treatment Company
Patent expiry EU/US
Humira adalimumab mAb arthritis Abbott/Eisai Apr 2018/Dec 2016 Remicade infliximab mAb arthritis Merck/Mitsubishi Aug 2014/Sep 2018 Rituxan/
MabThera
rituximab mAb arthritis, non-Hodgkin’s lymphoma
Roche/Biogen-Idec Nov 2013/Dec 2018
Enbrel etanercept mAb arthritis Amgen/Pfizer/Takeda Feb 2015/Nov 2028 Lantus insulin glargine protein diabetes Sanofi 2014/2014
Avastin bevacizumab mAb cancer Roche Jan 2022/Jul 2019 Herceptin trastuzumab mAb breast
cancer
Roche Jul 2014/Jun 2019
9 1.3. Protein Analysis
Protein analysis includes various separation methods, purification techniques and analyses that reveal their identity, characteristics, modifications, purity or quantity. These methods are either gel methods (e.g. SDS-PAGE, capillary gel electrophoresis) or non-gel methods (e.g. off-gel electrophoresis, microarrays, chromatography, mass spectrometry). In turn, these methods may be fluorescent, metabolic, chemical or enzymatic labeled or non-labelled.
Protein properties (physical, chemical, biochemical) play an important role in their analysis. For example, their high molecular weight is used to separate them from smaller molecules by dialysis, ultrafiltration, gel filtration chromatography and density-gradient ultracentrifugation. The ability to bind to specific antibodies, coenzymes, or hormone receptors has been used for immunochemical assays and affinity chromatography. Other important properties used in protein analysis are differential solubility, adsorption to surfaces, electrical charge, enzymatic activity and predisposition to chemical or enzymatic digestion.2
Nowadays, due to the technical developments, liquid chromatography combined with tandem mass spectrometry (LC-MS/MS) is largely used in protein and biopharmaceuticals analysis.3,14 LC-MS/MS allows protein quantitation for which an immunoassay is not available. Moreover, the methods are highly selective, accurate and precise within a widespread dynamic range.3 Mass spectrometry has been used to analyse intact proteins (top-down analysis) or peptides (bottom-up analysis). These two main approaches are presented in the following sections.
1.3.1. Intact Protein Analysis
Intact protein analysis, also known as top-down analysis, refers to examination at the protein level. Top-down MS analysis is used for molecular weight, protein sequence and PTMs (number, position and nature) determination. However, the size of analytes usually causes complications. For example, the fragmentation is less predictable and high mass resolution instruments are required (e.g. FT-ICR, TOF or orbitrap-MS).7,14 Furthermore, the software for data analysis is less advanced as the one for peptide analysis. Despite its limitations, top-down analysis offers important information about the protein’s purity, aggregation, fragmentation, and high order structures.7 Furthermore, tissues can be visualized using mass spectrometry. For example, MALDI-TOF MS was used to identify microbes17 and mass spectrometry imaging (MSI) has been used in clinical research for biomarkers discovery, imaging endogenous metabolites and neurotransmitters, and for molecular detail inside the tumor.10,18
1.3.2. Peptide Analysis
Protein analysis at the peptide level is also known as bottom-up analysis or peptide mapping. In this approach, the proteins are first enzymatic digested. Then the specific peptides, also called signature peptides or surrogate peptides, are quantified. Finally, the concentration is extrapolated to the protein level.3,19 Therefore, the peptide formation is a crucial step for the bottom-up protein analysis. Before LC-MS/MS analysis the surrogate peptides may be purified and enriched to enhance the sensitivity and selectivity of the method.
A variety of pre-treatment methods for the peptide analysis are described in the literature. These consist usually of the following four steps: protein denaturation, reduction, alkylation and enzymatic digestion, as shown in Figure 1 and discussed further on.
10
Protein denaturation can be defined as the unfolding of the protein. As a result, the access of enzymes to the protein cleavage sites is facilitated.3,20,21 As mentioned before, in paragraph 1.1.1, the folded structure of proteins depends on various interactions, therefore protein denaturation can be performed in numerous ways. For example, Proc et al. (2010) used 14 combinations of heat, organic solvents (acetonitrile, methanol, trifluoroethanol), surfactants (SDS, sodium deoxycholate) and chaotropic agents (guanidine, urea) to determine the optimum for 45 plasma proteins digested with trypsin. Nowadays, RapiGest, a surfactant patented and introduced in 2002 by Waters, is one of the widely used denaturation products for in-solution enzymatic protein digestions. This surfactant improves enzymatic digestion and can be easily removed by centrifugation.21 The effect of different denaturation agents on the peptides forming was investigated in this study as well.
Figure 1. General protein sample pre-treatment steps for in-solution trypsin digestion.
Protein reduction facilitates the access of the enzyme to the cleavage sites by breaking the existing disulfide bonds. It is generally accomplished by using an 1,4-dithiothreitol (DTT) solution. Other reduction agents, like β-mercaptoethanol and tris(2-carboxyethyl)phosphine are also mentioned in literature.22 After reduction, the thiol groups (SH) have to be derivatised in order to avoid oxidation, renaturation and formation of new disulfide bonds. The derivatisation is usually done by alkylation with iodoacetic acid or iodoacetamide (IAA).3,20,22 Proteins can be digested by different types of enzymes which cleave at specific amino acids. For example, trypsin generates peptides after cleavage at the carboxyl side of arginine (R) or lysine (Lys, K), if not followed by proline (P), endoproteinase Lys-C cleaves on the C-terminal side of lysine, endoproteinase Glu-C cleaves on C-terminal of glutamic acid (Glu, E) and aspartic acid (D), chymotrypsin cleaves on C-terminal of tyrosine (Y), tryptophan (W) and phenylalanine (F), if not followed by proline, and endoproteinase Asp-N cleaves at the N-terminal of aspartic acid (Asp, D).3,14
DENATURATION (protein unfolding)
heat
organic sovents (ACN, MeOH, CF3CH2OH)
surfactants (RapiGest, SDS, Na deoxycholate) chaotropic agents (urea, guanidine HCl)
REDUCTION (breaking the disulfide bonds) DTT
30-60 min at 50-60⁰C
ALKYLATION(preventing the disulfide bonds (re)formation) IAA
30-120 min at RT, in the dark
TRYPSIN DIGESTION (peptide forming) enzyme:protein ratio 1:20 - 1:100
11
In addition, the quality and quantity of the enzymes determine the efficiency of the protein digestion. Trypsin is commercialised in different formulations, such as proteomics grade and trypsin suitable for cell culture. Proteomics grade trypsin is highly purified and its lysine residues are methylated to prevent autolysis. Moreover the specific activity is about five times higher than the activity of the trypsin suitable for cell culture.23,24 Therefore, proteomics grade trypsin is mostly used for in-solution protein digestion. Further, an enzyme:protein ratio of 1:20 to 1:100 is generally used.3
As mentioned before, protein digestion is a critical step, therefore digestion optimisation is essential for method development. The efficiency of digestion determines the accuracy and precision of the protein quantification. Although the importance of this step is widely acknowledged, there is limited literature on digestion optimisation.3,22 Trypsin digestion is generally done overnight at 37°C, which makes it the most time consuming step. However, this is usually convenient rather than optimal.20 Unconventional methods suggest the use of higher incubation temperature (50-60⁰C), ultrasonic, infrared or microwave energy, alternating electric fields or trypsin immobilised on a solid support to accelerate the protein digestion.22,25 However, monitoring the process can help to determine the optimal incubation time and temperature.
Another critical step is the signature peptide selection. The signature peptide can be predicted using different software or databanks such as Skyline, MASCOT, PeptideAtlas or BLAST.This peptide has to be unique and representative for the target protein, thus free from PTMs and not disposed to modifications. Moreover, the signature peptide has to be stable, has to be rapidly and reproducible formed by enzymatic digestion, and has to ionise and dissociate rapidly by MS/MS analysis. Theoretically, a signature peptide contains 8 to 15 residues in order to restrict the distribution of charge, to achieve sufficient retention by chromatographic separation and to obtain MS/MS fragmentation.However, monitoring a set of signature peptides is preferred for the reliability of the data.3,19,22
An internal standard (IS) is generally used to correct for the variations which may occur during sample preparation, chromatographic separation and detection. For protein quantitation, peptides or intact proteins are used as internal standards. The IS can be stable isotope labelled (SIL) or structural analogue. The SIL-peptides are produced by incorporating an isotope, such as 13C, 15N or 18O, on the selected amino acid residue of the chosen signature peptide. The analogues peptides are created by single conservative amino acid replacements (SCAR), for example adding/subtracting a methylene group in one of the side chains amino acids.3,26 The main drawback of using a peptide as an internal standard is that it does not undergo the digestion process, which is likely one of the steps where the most variability may occur.
Internal standards which follow the complete process are the winged or concatenated peptides, SIL-peptides with extra cleavable groups to the N- and C- terminus, and the intact proteins.3,27,28 Proteins can be SIL-proteins or analogue proteins.3,29 Yang et al. (2007) have developed an assay for the quantification of somatropin and a mAb using bovine fetuin as internal standard.29 In addition, they set analogue IS against stable isotope labelled peptide and concluded that the results were similar. Therefore, bovine fetuin was used in this present research as IS for the quantification of somatropin.
The peptides are normally separated using reverse phase liquid chromatography. The separation is based on the distribution of the analytes between a liquid mobile phase containing water, acetonitrile and 0.1% TFA/FA and a solid stationary phase, often a C18 column and normally run with a gradient program.14 Liquid chromatography is generally used for peptide analysis because the diversity in physicochemical properties of the peptides favors separation.14
12
Further, the peptides are analysed and quantified by MS/MS. Triple quadrupole instruments operating in selected reaction monitoring (SRM) mode are commonly used for peptide quantitation.3,19 The sample is first ionised, usually by electro spray ionisation (ESI), then the first quadrupole (Q1) selects the precursor ion based on m/z ratio. Next, the precursor ion is fragmented in the second quadrupole, the collision induced dissociation (CID), into product ions. Finally, the target product ions are analysed by the third quadrupole (Q3).
1.3.3. Method Validation
Analytical methods have to be validated to ensure robustness and accuracy of the generated data. The guidance for bioanalytical method validation is based on the principles of the FDA30 and EMA31. These guidelines provide general recommendations for bioanalytical method validation. Furthermore, these guidelines define different categories of validation, e.g. full validation, partial validation and cross-validation, and describe when a certain validation type should be performed. The analytical laboratories adapt this principles in their standard operating procedures (SOPs).
Bioanalytical method validation should include fundamental parameters like accuracy, precision, selectivity, sensitivity, reproducibility and stability. Furthermore, depending on the goal and application of the assay other parameters, like matrix effect, carry-over, lower limit of quantification (LLOQ), upper limit of quantification(ULOQ), calibration curve and sample dilution, might be validated as well.30,31
Whether an analytical method is validated or not depends on meeting the validation criteria. The validation criteria of the parameters included in the method validation of this study are presented further on.
A method is considered selective if the peak area of the matrix is ≤ 20% of the peak area of the analyte at the LLOQ level and ≤ 5% of the response of the IS.
Carry-over should be ≤ 20% of the analyte at the LLOQ level and ≤ 5% of the IS response. Calibration curve should be based on minimum six calibration standards. The back calculated concentrations of the calibration standards should be ≤ 20% of the analyte response at the LLOQ level and ≤ 5% of the IS response. At least 75% of the calibration standards must fulfil this criterion. In case replicates are used, the criteria should also be fulfilled for at least 50% of the calibration standards tested per concentration level.
The accuracy and precision should be tested using quality controls (QCs) in minimum 5-fold per concentration level and minimum 4 concentration levels. The concentration levels should be: LLOQ level, 3 times LLOQ level (low QC), about 30-50% of the calibration range level (medium QC) and circa 75% of the highest calibration standard level (high QC). The method is accurate if the mean concentration of the LLQC is within 20% of the nominal values and within 15% for the other QC samples. The accuracy is determined within-run accuracy and between-run.
Precision, as a function of the coefficient of variation (CV), is determined as well within-run and between-run. The CV value should not be higher than 20% for the LLOQ and 15% for the other QC samples.
The influence of the matrix should be investigated by analysing the analyte prepared in minimum six different sources of human plasma and in a solvent without matrix (reference samples). The coefficient of variation of the normalized matrix factor (ratio of the response of the matrix sample to the mean response of the reference samples) should to be ≤ 15% at each concentration level.
The stability of the analyte in matrix should be evaluated using low QC samples and high QC samples. The mean concentration at each level should be within ± 15% of the nominal concentration.
13 2. EXPERIMENTAL DATA
This research was conducted using an LC-MS/MS method. The samples were aqueous protein solutions and spiked plasma; patient samples were not used. The proteins (human serum transferrin, somatropin and bovine fetuin) were diluted, denaturated, reduced, alkylated and enzymatically digested. The signature peptides were chromatographically separated and further fragmented and detected using a triple quadrupole detector. The chemicals, materials, reagents and the used equipment are presented further on, followed by the sample preparation, signature peptide selection and analysis methods.
2.1. Chemicals and Materials
Transferrin human (T3309, purity ≥ 98%), somatropin Chemical Reference Substance (batch 3, 3.86 mg of somatropin monomer per 55 mg vial), fetuin from fetal bovine serum (F2379, suitable for cell culture), trypsin (T6567, proteomics grade and T4799, suitable for cell culture), DL-Dithiothreitol (43819, purity ≥ 99%), iodoacetamide (I1149, purity ≥99%), ammonium bicarbonate (40867, LC-MS grade), sodium dodecyl sulphate (71725, purity ≥ 99%) and sodium deoxycholate (D6750, purity ≥ 97%) were purchased from Sigma-Aldrich (Steinheim, Germany). RapiGest™ SF Surfactant was purchased from Waters (Milford, MA, USA). Formic acid (99%) and acetonitrile (99.97%) were purchased from Biosolve (Valkenswaard, The Netherlands). Water was purified using a Milli-Q Advantage A10 system (Merck Millipore, Darmstadt, Germany).
Human plasma (50/50 mix male/female) and Guinea pig plasma (50/50 mix male/female) were supplied by WIL Research Europe (Den Bosch, The Netherlands), while mouse plasma (50/50 mix male/female) was purchased from Bio Services (Uden, The Netherlands). The human plasmas for validation (3 different batches male and 3 different batches female) were from Sera Lab (UK).
The 1 mL polypropylene 96-(deep)wells plates with V-bottom and silicon microplate covers were from Screening Devices (Amersfoort, The Netherlands) and amber pushcap tubes with thermoplastic elastomer (TPE) caps from Micronic (Lelystad, The Netherlands).
2.2. Preparation of the Stock Solutions, Calibration Standards and Quality Controls
Transferrin stock solutions (2.6 mg/mL) were prepared in brown glass tubes with Milli-Q water and were stored in the refrigerator for maximum 1 week. The concentrations of the human serum transferrin calibration standards were as follows: 2.6, 7.0, 20, 50, 150, 375, 1000 and 2600 µg/mL.
Somatropin stock solutions of 1.0 mg/mL, in 50 mM of NH4HCO3 or in human plasma, were
freshly made as well. The concentrations of somatropin calibration standards were 3.0, 6.0, 15, 30, 75, 150, 375 and 900 µg/mL. Somatropin quality controls were prepared at 3.0, 9.0, 45 and 675 µg/mL and were made in 50 mM of NH4HCO3 and in human plasma.
The bovine fetuin solution (1.0 mg/mL) in Milli-Q water, as well as, the trypsin (1.0 mg/mL), DTT (33mM) and IAA (50mM) solutions in 50 mM NH4HCO3 were freshly prepared.
14 2.3. Equipment and Software
The ACQUITY UPLC was equipped with: sample manager, sampler organiser, column manager, binary solvent manager (all Waters Corporation, Milford, MA, USA). A valco valve (VICI AG International, Switzerland) was used for the connection with the mass spectrometer of an AB Sciex API 4000 (AB Sciex, Ontario, Canada) equipped with Turbo ion spray interface.
The analytical columns, Acquity UPLC BEH C18, 50 2.1 mm ID, 1.7 µm and Acquity UPLC BEH Shield RP18, 50 2.1 mm ID, 1.7 µm, and the inline filters, ASSY frit, 2.1 mm ID, 0.2 µm, were purchased from Waters.
Other laboratory equipment used for the experiments included: stove (WTC Binder, Germany), benchtop centrifuge (Multifuge 3 S-R Heraeus, Thermo Scientific, Germany), Variomag Monoshake (Thermo Electron LED, Germany), infuse pomp (Harvard Apparatus 11, UK), injection syringe (Hamilton 1 mL, Reno, Nevada) and pipettes (Gilson, UK).
The LC-MS/MS system was controlled by Analyst software version 1.6.2 (AB Sciex). Furthermore, Skyline software (free version 64-bit 2.6.0.7176, MacCoss Lab, University of Washington, USA), PeptideAtlas database (developed by Institute of Systems Biology, Seattle, USA) and MASCOT (free version software, Matrix Science website) were used.
2.4. Sample preparation
Two protocols were used for the sample preparation. The first one was based on the Yu and Gilar21 procedure and the second one has been the result of the optimisation process. Both protocols are described below.
Protocol 1
The samples (10 µL) were pipetted into 96-well (1 mL) plate, diluted and denaturated by adding 50 µL of 0.1% RapiGest surfactant in 50 mM ammonium bicarbonate (NH4HCO3).
Then, the plate was placed on a plate shaker and mixed for 1 minute at approximately 300 rpm. Subsequently, 6 µL of 33 mM dithiothreitol (DTT) in 50 mM NH4HCO3 was added,
mixed, as described before, and placed in a stove for 30 minutes at 50°C. Afterwards, the plate was centrifuged 5 minutes at 3761 g, to cool down at room temperature and to collect condensation. 8 µL of 50 mM Iodoacetamide (IAA) in 50 mM NH4HCO3 was added and the
plate was placed in the dark, at room temperature. After 40 minutes, 8 µL of 1 mg/mL Trypsin in 50 mM NH4HCO3 was added, mixed and placed in a stove at 37°C
overnight. After circa 20 hours the enzymatic reaction was stopped using 82 µL 1% FA in 10/90 ACN/Milli-Q water solution (ratio pretreated sample: stop solution was 1:1, v:v).
Protocol 2
The denaturation was done with 50 µL of 10% sodium deoxycholate in 50 mM NH4HCO3 and
the protein incubation with trypsin was 1 hour at 37°C. The other steps were the same as the previous protocol.
2.4.1. Protein Denaturation
Protein denaturation was performed using the following five denaturation agents: heat (95⁰C, 8 minutes)32, RapiGest (0.1%)21, SDS (30%, w/v), sodium deoxycholate
(10%, w/v) and guanidine HCl (6M). The volume of all denaturation agents was 50 µL, except for the samples denaturated by heat. These were first diluted with 50 µL of 50 mM NH4HCO3. Sample treatment was according to protocol 1 (see paragraph 2.4).
15
The samples used for this experiment were: human plasma, 100 µg/mL transferrin solution in human plasma and 50 µg/mL somatropin with IS (10 µL of 0.100 mg/mL bovine fetuin). Human plasma, without spiking, was used to measure transferrin because this is an endogenous protein. The concentration range for a healthy adult is between 2.0 to 3.6 mg/mL.33 As a result the measured peak areas of the spiked and not spiked plasma were similar.
2.4.2. Protein Reduction and Alkylation
Protein reduction was performed be adding 6 µL of 33 mM DTT followed by incubation at 50°C for 30 minutes. Subsequently, the alkylation was performed by adding 8 µL of 50 mM IAA followed by incubation in the dark at room temperature for 40 minutes.
2.4.3. Protein Proteolysis
In this research the proteins were digested by trypsin. The efficiency of different qualities and quantities were tested. Further, the incubation time and temperature were monitored to determine the optimal digestion conditions.
First experiments were conducted with both types of trypsin according to protocol 1 (paragraph 2.4). For these experiments 2.6 mg/mL transferrin aqueous solutions were used. Subsequently, all the other tests were conducted with trypsin suitable for cell culture.
The influence of the trypsin concentration on the digestion efficiency was tested by conducting the following experiment: aqueous transferrin samples (2 mg/mL) were treated with 8 µL of 0.5, 1.0 and 2.0 mg/mL trypsin suitable for cell culture solution.
Further, the influence of different temperatures and different incubation times was tested. One experiment was conducted with transferrin samples (2 mg/mL, aqueous), treated according to the protocol 1, and incubated at 37, 50 and 60°C for 1, 2, 4, 6, 22, 26 and 72 hours. Another experiment was performed with somatropin samples (0.5 mg/mL, 50% human plasma), treated according to the protocol 2 and incubated at 37 and 50°C for 1, 2, 4 and 22 hours.
2.5. Surrogate Peptides Identification and Selection
First, the protein sequences for human serum transferrin, bovine fetuin and, human growth hormone were acquired from the database of Universal Protein Resource (UniProt).34,35 The sequence of somatropin was acquired from the DrugBank database.36 All these sequences are presented in Appendices 1-3. Subsequently, the lists of peptides formed by trypsin digestion were acquired from the Skyline software. Next, the surrogate peptides were identified by direct MS infusion and LC-MS/MS analysis using the MS/MS settings imported from Skyline. Finally, the peptides were checked for their specificity using PeptideAtlas and MASCOT software.
2.5.1. Surrogate Peptides Identification by Direct MS Infusion
Blank and protein digested samples were directly infused, with 1 mL/min flow, in the first quadrupole (Q1) of the mass spectrometer. The following solutions of protein were infused: - transferrin (concentrations of 0.26, 2.6, 26.0, 130 and 2600 µg/mL, in Milli-Q water), - bovine fetuin (1 mg/mL, in Milli-Q water),
- somatropin (1 mg/mL, in 50 mM ammonium bicarbonate).
The spectra of the digested proteins were compared with the blank spectrum and the peaks which were present only in samples were considered as being potential surrogate peptides. Their m/z values were compared with the data from Skyline. If matched, the programmed infusion optimisation process was started.
16
Throughout this automatic procedure, the optimum values for precursor and product ions were determined by increasing and decreasing the declustering potential (DP), collision energy (CE) and collision exit potential (CXP) voltages. At the end of this process a report which includes the optimal voltages was created.
An example of the identification of the somatropin surrogate peptide LEDGSPR by MS infusion is illustrated below. Blank and protein digested samples, both aqueous solutions, were directly infused in the mass spectrometer. Figure 2 shows the full-scan spectrum (Q1 scans) of the (aqueous) blank digested sample, while Figure 3 shows the one of somatropin (1.86 mg/mL, in 50 mM ammonium bicarbonate) digested sample. Comparing the two spectra, it was observed that two peaks, by m/z values 387.4 and 382.3, were present only in the somatropin spectrum. These m/z values correspond to the surrogate peptide LEDGSPR, respectively FETFLR. Therefore, automatic compound optimisation was performed for each peptide, as described in paragraph 2.6.1. In the first case, at the end of the optimisation process, it was established that the highest intensity of the product ion was 531.3. This fragmentation, 387.4++531.3+, corresponds to the LEDGSPRDGSPR fragmentation and confirms the Skyline data, as shown in Figure 4. In the second case, the fragmentations of the parent ion did not match the somatropin surrogate peptide FETFLR.
Figure 2. Mass spectrum of the aqueous blank sample
17 Figure 3. Mass spectrum of the aqueous somatropin sample (full scan in the m/z interval
300-900 Da).The peaks by m/z 387.4 and 382.3 might indicate the presence of the somatropin surrogate peptide LEDGSPR, respectively FETFLR.
Figure 4. Somatropin data by Skyline software.
Tryptic peptides are displayed on the left side. The precursor charge (2+), m/z value (387.4115) and [M+H] value (773.8157) of the peptide LEDGSPR are displayed on the right side, in the small yellow window, together with the product ions. The product ions with the highest intensities, 660.66, 531.54 and 416.45, are given in blue.
18 2.5.2. Surrogate Peptides Identification by Importing MS-settings from Skyline
Protein digests and blank samples were analysed by LC-MS/MS. The MS/MS quantitation methods were built with data imported from Skyline (m/z of the precursor ion and product ion, Q3, DP, CE). The extracted ion chromatograms (XIC) were examined for the presence of signals at the selected m/z values.
2.5.3. Signature Peptide Selection
After identification, the surrogate peptides were verified for specificity using MASCOT and PeptideAtlas software and a set of peptides was chosen for protein identification and quantification by LC-MS/MS analysis.
2.6. LC-MS/MS Method
An LC-MS/MS method was first optimised for the transferrin analysis. The separation of surrogate peptides was done by gradient reversed phase chromatography. During the optimisation process, the optimal gradient, chromatographic column, injection needle, as well as the composition of the mobile phases and the washes for the needle were established. The surrogate peptides detection was done by using electrospray ionisation (ESI) in positive mode and a triple quadrupole detector. Further, an LC-MS/MS method was optimised and validated for the analysis of somatropin in human plasma using bovine fetuin as internal standard.
The carry-over for transferrin was tested by both steel and polyetheretherketone (PEEK) injection needle. The carry-over for the PEEK-needle was lower than 20% of the analyte response, while the one for the steel needle was higher. In order to reduce it even further, higher volumes of strong and weak wash were used, namely 400 µL and 1200 µL. The linear gradient was adjusted in order to obtain the analyte optimal peak shape in minimum time and it is described in Table 2. The column used was Acquity UPLC BEH C18. The mobile phase A was 10/90/0.1 ACN/Milli-Q water/FA and mobile phase B was 90/10/0.1 ACN/Milli-Q water/FA. The entire LC-MS/MS conditions are given in Table 3.
The ESI source of the mass spectrometer was operated in the positive mode. The acquisition parameters for the MS/MS/detection are listed in Table 4.
Table 2
Linear gradient for transferrin and somatropin LC method Time
(min)
Transferrin Somatropin
Mobile phase Mobile phase
A (%) B (%) A (%) B (%) 0.00 85.0 15.0 99.9 0.1 0.30 85.0 15.0 99.9 0.1 1.50 0.1 99.9 50.0 50.0 1.70 0.1 99.9 50.0 50.0 1.80 85.0 15.0 99.9 0.1 2.00 85.0 15.0 99.9 0.1
19 Table 3
LC-MS/MS conditions for transferrin and somatropin analysis
Table 4
MS/MS acquisition parameters
Compound Parent ion (m/z, z=2) Product ion (m/z, z=1) CE (V) DP (V) Dwell time (ms) Transferrin 625.4 776.4 29.0 66.0 100 709.0 1177.5 31.0 126.0 100 Somatropin 338.4 463.5 15.0 55.8 20 387.4 531.5 18.0 59.4 20 IS (bovine fetuin) 408.9 632.6 19.0 61.0 20 738.4 880.0 38.0 85.0 20
The LC method was optimised for the somatropin analysis. The bovine fetuin (IS) peak presented a ‘shoulder’ using the column Acquity UPLC BEH C18, as the first chromatogram of Figure 5 shows. Therefore, an Acquity UPLC BEH Shield RP18 column, which has an embedded carbamate group, was tested. As a result, the two peaks were separated, as the second chromatogram of Figure 5 shows. Therefore, this column was used for the somatropin assay. The linear gradient was adjusted in order to obtain the analyte optimal peak shape in minimum time and it is described in Table 2. The mobile phase A was 100/0.1 Milli-Q water/FA and mobile phase B was 90/10/0.1 ACN/Milli-Q water/FA with a 0.6 mL/min flow. Further, the weak wash was replaced with 100%Milli-Q water. The entire LC-MS/MS conditions are given in Table 3.
The ESI source of the mass spectrometer was operated in the positive mode as well and the acquisition parameters are listed in Table 4.
Column Transferrin: Acquity UPLC BEH C18, 50 2.1 mm ID, 1.7 µm
Somatropin: Acquity UPLC BEH Shield RP18, 50 2.1 mm ID, 1.7 µm Filter inline filter ASSY frit, 2.1 mm id., 0.2 µm
Column temperature 30⁰C Autosampler temperature 4⁰C Injection volume 5 µL Injection loop 10 µL
Needle 30 µL Peek needle
Sample syringe 100 µL
Mobile phase A Transferrin: 10/90/0.1 (v/v/v) ACN/Milli-Q water/FA Somatropin: 100/0.1 (v/v) Milli-Q water/FA
Mobile phase B 90/10/0.1 (v/v/v) ACN /Milli-Q water/FA
Strong needle wash 150/150/150/9 (v/v/v/v) 2-propanol (IPA)/ ACN/MeOH /NH4OH Weak needle wash Transferrin:10 / 90 ACN/Milli-Q water
Somatropin: Milli-Q water
Flow 0.6 mL/min
Ionisation source ESI+ Ion spray voltage 4000 V Source temperature 450⁰C
20 Figure 5. Chromatographic separation of bovine fetuin peptide TPIVGQPSIPGGPVR using an Acquity UPLC BEH C18 column (above) and an Acquity UPLC BEH Shield RP18 column (under).
Additionally, the LC-MS/MS method for the analysis of somatropin in human plasma using bovine fetuin as internal standard was validated. The validation process was based on the principles of FDA and EMA on bioanalytical method validation included in the WIL Research method validation SOP. The following parameters were validated: selectivity, accuracy and precision, calibration line, carry-over, freeze/thaw stability, short term stability at room temperature and the stability of processed samples. Three analytical runs were performed and the results are presented further in paragraph 3.5 and in Appendix 4.
The samples used for validation were: blank plasma with and without IS, 6 different blank human plasma (3 male and 3 female) with and without IS, system suitability test standards (3.0 µg/mL), calibration standards (3.0, 6.0, 15, 30, 75, 150, 375 and 900 µg/mL), quality controls in 5-fold (QC-LLOQ (3 µg/mL), QC-L (9 µg/mL), QC-M (45 µg/mL) and QC-H (675 µg/mL)), quality controls (QC-L and QC-H) in 2-fold prepared with the 6 different human plasma, quality controls (QC-L and QC-H) in 3-fold prepared in 50 mM NH4HCO3.
21 3. RESULTS
In this chapter, the results of the protein denaturation using five different denaturation agents, the digestion optimisation and the identified surrogate peptides of transferrin, somatropin and bovine fetuin are listed. Further, the validation of the LC-MS/MS somatropin quantification method is presented.
3.1. Protein Denaturation Using Five Different Denaturation Agents
In this current study, the proteins were enzymatic digested after being denaturated with RapiGest, SDS, sodium deoxycholate, guanidine HCl and by heat. The samples, in 4-fold, were pipetted in 1.4 mL amber Micronic tubes. Denaturation was achieved by adding 50 µL of one of the reagents mentioned before. The samples which were denaturated by heat were first diluted with 50 µL of 50 mM NH4HCO3. The average peak areas of the surrogate
peptides are presented in Table 5. The most efficient transferrin denaturation was obtained by heating the samples at 95⁰C, closely followed by RapiGest and Na deoxycholate. Guanidine HCl had a low efficiency and no peptide response was found by SDS denaturation. The efficiency of the somatropin and bovine fetuin (IS) denaturation was similar by RapiGest and Na deoxycholate, while all the other denaturation agents were found to be inefficient.
Table 5
Transferrin, somatropin and bovine fetuin peak areas variation with five different denaturation agents Denaturation agent Peak areas (cps) Transferrin peptide SVIPSDGPSVACVK Somatropin peptide LEDGSPR IS peptide ALGGEDVR 95⁰C 2.0*105 no response no response RapiGest 1.6*105 1.8*104 4.4*105
Guanidine HCl 6.0*103 no response no response
SDS no response no response no response
Na deoxycholate 1.3*105 1.7*104 5.0*104
3.2. Protein Digestion Using Different Qualities and Quantities of Trypsin
Transferrin calibration standards (aqueous) were treated with trypsin proteomics grade and with trypsin suitable for cell culture according to protocol 1, as described in paragraph 2.4. The peak areas for the surrogate peptide SASDLTWDNLK, transition 625.3++776.4+, are given in Table 6 and the calibration lines in Figure 6. The peptide formation using trypsin suitable for cell culture was around 80% lower for the lowest and highest calibration standard and around 50% lower for the other calibration standards, except for two outliers (50 and 150 µg/mL).
The influence of different trypsin concentration was tested only with trypsin for cell culture. Transferrin (2 mg/mL, in Milli-Q water) samples were analysed in 3-fold, following the sample preparation from protocol 1. Trypsin concentrations were 0.5, 1.0 and 2.0 mg/mL and the incubation time with trypsin was 26 hours. Two surrogate peptides were monitored, namely SASDLTWDNLK (transition 625.4++776.4+) and SVIPSDGPSVACVK (transition 709.0++1117.5+). Trypsin with the concentration of 1.0 mg/mL was considered the reference. The response of the first peptide was 3% higher when using 0.5 mg/mL trypsin and 7% lower by 2 mg/mL trypsin. The response of the second peptide was 14% higher by using 0.5 mg/mL trypsin and 3% lower by trypsin 2 mg/mL trypsin. In Table 7 are these results listed.
22 Table 6
Transferrin peak areas of the surrogate peptide SASDLTWDNLK (transition 625.3++776.4+) by digestion with trypsin proteomics grade (I) and trypsin suitable for cell culture (II).
Concentration transferrin (µg/mL) Peak area surrogate peptide SASDLTWDNLK (cps)
I II 2.6 1.0*103 8.2*102 7.0 4.9*103 2.8*103 20 2.2*104 1.0*104 50 1.7*104 2.4*104 150 2.5*105 8.6*104 375 7.7*105 3.8*105 1000 1.3*106 8.7*105 2600 2.7*106 2.1*106
Figure 6. Calibration lines for transferrin peptide SASDLTWDNLK by digestion with trypsin proteomics grade and trypsin suitable for cell culture.
Table 7
Transferrin peak areas by digestion with trypsin suitable for cell culture at different concentrations Peak areas transferrin (cps)
peptide SASDLTWDNLK (transition 625.4++776.4+) peptide SVIPSDGPSVACVK (transition 709.0++1117.5+) Conc. trypsin (mg/mL) 0.5 1.0 2.0 0.5 1.0 2.0 Peak area 1 (cps) 327508 304759 295595 513378 445558 443609 Peak area 2 (cps) 325815 339787 293293 524392 474422 438631 Peak area 3 (cps) 326555 309686 301465 523627 449353 440523 Peak area average (cps) 326626 318077 296784 520466 456445 440921 3.3. Protein Digestion at Different Temperature and Incubation Time
The influence of incubation time and temperature variation was tested first on 2 mg/mL transferrin samples. The peak areas of two surrogate peptides, peptide I, SASDLTWDNLK, transition 625.4++776.4+, and peptide II, SVIPSDGPSVACVK, transition 709.0++1117.5+, were monitored at three different temperatures: 37, 50 and 60°C after 1, 2, 4, 6, 22, 26 and 72 hours of incubation time.
0 5 10 15 20 25 30 0 5 10 15 20 25 30 P ea k ar ea ( cps) *105 Concentration (µg/mL) *102
Calibration lines transferrin
trypsin proteomics grade23
The t=0h was actually 5 minutes (0.083 hours) at room temperature. The samples were prepared in 3-fold according to protocol 1 presented in paragraph 2.4. The average values of the peak areas are given in Table 8. Peptide I reached the maximum after an incubation of 22 hours at 37⁰C. At 50⁰C, the highest amount was found after 1 hour and it was about 8% lower than the maximum. At 60⁰C, the highest amount was also after 1 hour, but it was 30% lower than the maximum. Peptide II reached the maximum after 6 hours at 37⁰C. At 50⁰C, the highest amount was found after 2 hours and it was about 10% lower than the maximum. After 1 hour at 60⁰C, the amount was 15% lower than the t=0.083h.
Table 8
Transferrin peak areas variation with time and temperature
Peak areas at 37⁰C (cps) Peak areas at 50⁰C (cps) Peak areas at 60⁰C (cps) Time (h) Peptide I Peptide II Peptide I Peptide II Peptide I Peptide II
0.083*) 210230*) 495364*) 210230*) 495364*) 210230*) 495364*) 1 299044 532113 308017 496098 232481 421627 2 301684 544808 302156 502194 177468 404988 4 316502 540168 304362 474543 200481 356642 6 321150 556097 289553 480272 190359 351921 22 336501 475036 273672 435252 171557 299306 26 318077 456445 263876 383655 197415 239575 *) at room temperature
The samples which were incubated for 72 hours were analysed in a different run, therefore the t=0.083 hours was analysed a second time. The peak areas after 72 hours incubation were lower than the peak areas at t=0.083 hours.
The responses of the transferrin signature peptides were unexpectedly high after 5 minutes at room temperature, therefore another experiment was conducted to verify these results. Human plasma and 2 mg/mL transferrin solution in 50 mM ammonium bicarbonate were prepared in 3-fold, according to protocol 1, kept 5 min at room temperature before the enzymatic reaction was stopped by adding 1% FA in 10/90 ACN/Milli-Q water and analysed by LC-MS/MS. The average responses were: in plasma 102468 counts (peptide I), respectively 442697 counts (peptide II) and in Milli-Q water 1953599 counts (peptide I) and 1948060 counts (peptide II).
Another set of experiments was conducted with 500 µg/mL somatropin, 1 mg/mL bovine fetuin (IS) and somatropin with IS. The samples, in 2-fold, were treated according to protocol 2 (paragraph 2.4) and incubated at two different temperatures, namely 37⁰C and 50⁰C, and monitored after 1, 2, 3, 4 and 22 hours. The average peak areas for somatropin (surrogate peptide LEDGSPR transition: 387.4++531.3+) and bovine fetuin (surrogate peptide ALGGEDVR, transition 408.9++632.6+) are presented in Table 9. Somatropin peptide reached the maximum after 4 hours incubation at 37⁰C and the highest amount at 50⁰C was found after 22 hours (about 4% lower than the maximum). Bovine fetuin peptide reached the maximum after 4 hours incubation at 37⁰C. The highest amount at 50⁰C was found after 1 hour and after 4 hours and it was about 10% lower than the maximum.
24 Table 9
Somatropin and bovine fetuin (IS) peak areas variation with time and temperature Time
(hours)
Peak areas somatropin peptide LEDGSPR
Peak areas bovine fetuin (IS) peptide ALGGEDVR at 37°C at 50°C at 37°C at 50°C 0.083*) 3129*) 3129*) 2975*) 2975*) 1 159295 153369 15796 14686 2 162692 144574 16375 13607 3 160577 142208 15977 13973 4 176302 165279 16680 14686 22 151130 169297 14849 14511 *) at room temperature
3.4. Surrogate peptides identification and selection
The surrogate peptides of transferrin, somatropin and bovine fetuin were identified first by direct MS infusion, then by LC-MS/MS analysis as described in paragraph 2.5.1. and respectively 2.5.2. The list of surrogate peptides identified during this research, the m/z values of the parent ions and the m/z values of the most intensive product ions are given below, in Table 10.
Table 10
Experimentally identified surrogate peptides for transferrin, bovine fetuin and somatropin Protein Peptide sequence
[position in protein sequence]
Parent ion (m/z ,z = 2+) Product ion (m/z, z= 1+) Transferrin HQTVPQNTGGK [553, 563] 584.1 701.1 APNHAVVTR [600, 608] 483.0 682.3 SVIPSDGPSVACVK [46, 59] 709.1 117.5 DSGFQMNQLR [122, 131] 598.6 789.5 FDEFFSEGCAPGSK [494, 507] 789.3 892.4 SASDLTWDNLK [453, 463] 625.3 776.4
Bovine fetuin TPIVGQPSIPGGPVR [333, 347] 738.4 879.5
HTLNQIDSVK [57, 66] 578.1 917.0 EVVDPTK [211, 217] 394.4 460.5 GSVIQK [231, 236] 316.4 388.5 ALGGEDVR [237, 244] 408.9 632.6 QDGQFSVLFTK [120, 130] 635.7 694.8 Somatropin LEDGSPR [127, 133] 387.3 531.3 IVQCR [178, 182] 338.4 463.5 SNLELLR [70, 76] 423.0 530.6 LFDNAMLR [8, 15] 490.4 719.8 DLEEGIQTLMGR [115, 126] 681.8 1005.2 TGQIFK [134, 139] 347.4 407.5
After identification, the peptides were verified for specificity using MASCOT and PeptideAtlas software. MASCOT data base matched the peptides with the protein only if a combination of three or more peptides was introduced in the query. On the other hand, PeptideAtlas software found all individual surrogate peptides of transferrin, five of bovine fetuin (TPIVGQPSIPGGPVR, HTLNQIDSVK, EVVDPTK, ALGGEDVR, QDGQFSVLFTK) and three of somatropin (SNLELLR, LFDNAMLR, DLEEGIQTLMGR) as being specific. However, for the reliability of the data, a set of peptides was chosen for protein identification and quantification.
25 3.5. Method Validation
An LC-MS/MS method was validated for the analysis of somatropin in human plasma using bovine fetuin as internal standard. The samples were treated according to the protocol 2, described in paragraph 2.4. The validation process was based on to the principles of FDA and EMA on bioanalytical method validation included in the WIL Research Europe SOP. The complete validation results are presented in Appendix 4.
Three analytical runs were performed. The following parameters were tested: the influence of the matrix on the analyte, selectivity, carry-over, accuracy and precision, calibration line, freeze/thaw stability, short term stability at room temperature and the stability of processed samples. Two of the most sensitive peptides for each protein were monitored, namely IVQCR and LEDGSPR for somatropin, and ALGGEDVR and TPIVGPSIPGGPVR for bovine fetuin (IS). The results were similar, therefore the results presented further are restricted to the somatropin peptide LEDGSPR (transition LEDGSPRDGSPR, 387.4++531.3+) and IS peptide ALGGEDVR (transition ALGGEDVRGGEDVR, 408.9++632.6+).
The retention time of the peptides was reproducible. Somatropin peptide LEDGSPR had a retention time of 0.90 minutes and the IS peptide ALGGEDVR 0.98 minutes.
The interference of the matrix with the analyte was lower than 20%. The matrix had no interference with the IS, thus the method was found to be selective.
Carry-over was not detected for the internal standard, but was detected for somatropin. The assay was found accurate and precise for the tested levels (3.0, 9.0, 45 and 675 µg/mL). The calibration line was found linear in the range of 3 to 900 µg/mL. Linear regression and an (1/concentration2) weighing factor were used.
The samples were found to be stabile after three freeze/thaw cycles and after being stored for 28 hours at room temperature. Further, the processed samples were found to be stabile after being stored for 47 hours at 4°C.
