Fundamentals: Bioanalytical LC/MS method validation - fit for purpose Ludmila Alexandrova Staff Scientist, Stanford University Mass Spectrometry (SUMS) [email protected] SUMS SEMINAR SERIES JULY 2, 2020
Fundamentals: Bioanalytical LC/MS method validation - fit for purpose
Ludmila AlexandrovaStaff Scientist, Stanford University Mass Spectrometry (SUMS)
SUMS SEMINAR SERIESJULY 2, 2020
Submit your questions!
Related Seminars in SUMS SeriesTHURSDAY APRIL 2, 2020 - SUMS SEMINAR SERIES
Fundamentals: Measuring concentrations of small molecules using mass spectrometry -theory and practice, Part I
THURSDAY APRIL 23, 2020 - SUMS SEMINAR SERIES
Fundamentals: Measuring concentrations of small molecules using mass spectrometry -theory and practice, Part II
THURSDAY May 21, 2020 - SUMS SEMINAR SERIES
Fundamentals: Applications of LC/MS in small molecule drug discovery
Recording and slides are available at https://mass-spec.stanford.edu/events
3
Presentation OutlineIntroduction
Ø Bioanalysis in Drug Discovery and DevelopmentØ Fit For Purpose Concept in Bioanalysis
Part 1Ø Method Validation: Bioanalytical Assay Performance Attributes and
Requirements
Part 2Ø Case study 1 – Melphalan LC-MS/MS Method Development and
ValidationØ Case study 2 –Determination of Global DNA Methylation Levels in
Clinical Samples by LC-MS/MS
Conclusion
4
5
Introduction
• Toxicology studies
• Clinical studies
• LC/MS bioanalytical methods
LO P1 P2 P3 Market
Non-GLP method GLP/GCP methodParentMetabolite
LI CCS CLS
GLP: Good Laboratory PracticeGCP: Good Clinical Practice
• Pharmacokinetic/ADME studies
• Biomarker research
AbsorptionDistributionMetabolismExcretion
6
It is critical to use well characterized and appropriately validated LC/MS bioanalytical methods to ensure data integrity
7
What Is Bioanalysisq Bioanalytical Method (BA) performs quantitative analysis of drugs and
metabolites in biological matrices (e.g. plasma, blood, serum or urine) in support of pharmacokinetic (PK), toxicokinetic (TK) and clinical studies, pharmacology evaluation, biomarker research, formulation development, etc. Small molecule platform – LC/MS/MS
q Fully validated BA methods support pivotal TK/PK and clinical studies that are used in regulatory submissions
q For non regulatory studies, extent of BA method validation may be defined by the purpose of the study.
8
Why is Bioanalytical Method Validation Important?q High data quality
Ø Reproducible and reliable dataØ Sensitivity establishedØ Stability investigated
q Data used for safety and efficacy assessmentq Data integrity aids to results interpretation q Increasing demand of peer-reviewed scientific journals
Why Not Done Routinely?q Time consuming and labor intensiveq Not always required e.g. screening studies, exploratory studies, endogenous/biomarkers
quantitation
Fit for purpose approach is a useful tool in conserving resources and time to generate reliable data
9
Fit For Purpose (FFP)q “The fit for purpose (FFP) concept states that the level of validation should be
appropriate for the intended purpose of the study”
Ø FFP applies to drugs, their metabolites, endogenous compounds and biomarkers and is used extensively during exploratory stages of drug development
Ø It is important to evaluate study goals and objectives as well as quantification approaches and create validation strategy
Ø Extent of validation and key parameters should be specified and justified in validation plan: e.g. accuracy, precision, stability etc.
Ø Specific validation requirements and acceptance criteria may need to be established for each analyte
Food and Drug administration. Bioanalytical method validation Guidance for industry.https://www.fda.gov/regulatory-information/search-fda-guidance-documents/bioanalytical-method-validation-guidance-industry
10
Method Validation: Bioanalytical Assay Performance Attributes and Requirements
11
BA Method Validation WorkflowoMethod development
• Instrument sensitivity: optimize column chemistry, mobile phase, mode of ionization, SRM, etc.• Sample preparation: PP simple, L-L less matrix impact, SPE more manipulations • Determine curve fitting/weighting; assess accuracy/precision• Key indicators: ion suppression, lot to lot variability
o Validation (guidance driven)• Precision and Accuracy: 3 runs, each a curve + 4 QC levels (n=6) à %CV and % Bias ±15% (or
±20% at LLOQ)• Stability in biological matrix (room temperature, Freeze-Thaw, frozen storage, auto-sampler) and
stock solution (storage)• Specificity (endogenous interference), matrix effect, selectivity, recovery, carryover
o In-study (routine) analysis• Curve, QC (3 levels, n=2), blank, contamination and carryover, IS, peak shape, RT
12
Method Validation: Key Parameterso Small molecule platform – LC-MS/MS in SRM modeo Method Validation plano Bioanalytical assay performance attributes and requirements
o Reference Standards/Internal Standardso Calibration curve, range, regression modelo Sensitivityo QC sampleso Accuracyo Precisiono Dilution effecto Selectivity, Specificityo Recoveryo Stability/Incurred sample reanalysis
o Documentation o Instrumentationo Software
13
Reference Standards/ Internal Standards o Certified reference standards with known identity and purity should be used to
prepare spiking solutions of known concentrations (certificate of analysis, CoA)
o Internal standard does not require CoA if it demonstrates that it does not interfere with quantification of analyte
o Analyte stock and working solution stability should be tested to justify duration and storage conditions
o Isotopic exchange stability of IS (analyte specific)
14
Calibration Curveo Calibration curve is a relationship between the instrument response and
concentration of analyte within selected concentration rangeØ Calibration curve should be prepared in the same matrix as the samples
and extracted side by side with the same procedureØ Choose concentration range of method on the basis of concentration range
expected in particular study: Lower Limit of Quantification (LLOQ) and Upper Limit of Quantification (ULOQ)
Ø LLOQ defines the method sensitivity and is assessed during method validation
Ø Use simplest regression model that describes the concentration – response relationship; appropriate weighting scheme
Ø If more than one analyte, calibration curve should be generated for all analytes
15
Quality Control Sampleso Quality control sample (QC) is a sample spiked with the known amount of
analyte
Ø QCs are prepared in the same matrix as samplesØ QC samples are used to assess precision, accuracy of the assay and
stability of analyte(s)Ø QCs prepared from a separate stock solutionØ It is recommended that blank matrix is screened before QC preparationØ To assess a dilution effect, QCs above Upper Limit of Quantification may
be prepared and analyzed in diluted form and monitored during validation
16
Precision and Accuracyo Precision is a repeatability of the series of measurements which is
expressed as % CVo Accuracy is the ability to measure a true value based on nominal value which is
expressed as % deviation from nominal value (% Bias)
Ø Precision and Accuracy of the method is determined based on QC replicate analysis at the LLOQ, Low, Mid and High QC levels over full calibration range conducted over several days (at least 3 independent runs).
Ø Precision and accuracy is calculated within-run (Intra-assay) and between-run (Inter-assay)
Ø Each run includes a calibration curve and multiple QC concentrations that are analyzed in replicates (≥ 5 replicates per each level).
17
Selectivity, Specificity and Matrix Effecto Selectivity is the ability to differentiate and measure the analyte of interest in the presence of endogenous
compounds Ø Selectivity demonstrated by analyzing blank samples of appropriate matrix from multiple
sources (at least 6) to determine and minimize interferenceØ Selectivity should demonstrate minimal or no response related to interfering component at the RT of
analyte in blank sample (≤ 20% LLOQ; ≤ 5% of the IS response)
o Specificity is the ability to unequivocally differentiate the analyte of interest in the presence of other analytes Ø Potential interferences: compounds structurally similar to analyte, unstable metabolites (e.g.
esters, unstable glucuronides, N-oxides etc.), isomers, impurities, decomposition products, concomitant medications and other xenobiotics
Ø Internal standard evaluated for interference on analyte quantitation
o Matrix effect is a change in the analyte response due to unspecified components in the matrixØ Evaluated in relevant patients or special populations; hemolyzed or lipemic samples or if
stabilizer or enzyme inhibitor is used during sample collectionØ Determine ion matrix suppression or enhancement
18
Recovery, Carryoverq Recovery is the extraction efficiency of analytical process reported as a percentage of the
known amount of analyte carried out through the processing method
o Recovery needs to be determined and optimized to insure it is efficient and reproducible
• It does not need to be 100%, but consistent and reproducible over calibration range (Low, Mid, High)
• Applies to both, analyte and internal standard• Recovery performed by comparing analytical results of extracted samples with
corresponding spiked post-extraction samples, representing 100% recovery
q Carryover is a change in measured concentration due to residual analyte interference from preceding sample o Carryover should be evaluated and eliminatedo If not possible, study samples should not be randomized
19
Stability of Analyte(s) in Biological Matrixq QCs are used for stability assessment
○ Autosampler (re-injection reproducibility)Ø Stability of processed samples at autosampler temperature
○ Bench-top (short-term) matrix stabilityØ Laboratory handling conditions. Kept on the bench top at the same temperature
and for the same duration as study samples ○ Extraction/injection (sample processing stability)
Ø In process sample stability to be used during sample analysis○ Freeze-Thaw matrix stability
Ø Stability assessment after multiple cycles of freezing and thawing according to the same procedure applied to the study samples
○ Long-term matrix stabilityØ Stability of analyte should be demonstrated under the same storage conditions
and duration as study samples
20
Incurred Sample Reanalysis (ISR)q ISR is a repeat analysis of subset of samples which had been analyzed
previously in a separate runq ISR confirms reliability of reported dataq It is used to support precision, accuracy and stability data established
with QCs. Ø E.g. presence of unstable metabolites which could degrade back to
parent compoundq ISR acceptance criteria measured by % difference of the results
between original value (1) and repeat value (2) divided by mean of 1 and 2
q ISR required for regulatory studies
21
Case Study 1
Development and Validation of LC-MS/MS Bioanalytical Method for Determination of Melphalan in Human Plasma
Kim JW, Alexandrova L, DeMarchis E, Lee D, Chien A. (2012) “LC-MS/MS Quantification Of Melphalan Plasma Levels In Children Undergoing Selective Intra-arterial Infusion Of Chemotherapy For Retinoblastoma,” Investigative Ophthalmology & Visual Science. March 2012, 53, 468.
22
Backgroundq Melphalan is an alkylating agent with effective tumoricidal properties
but also severe systemic side effects.
q A minority of subjects developed neutropenia, which suggests that systemic diffusion of the drug does occur.
q Developing a reliable LC-MS/MS assay to determine plasma levels of melphalan following ophthalmic artery infusion is critical in optimizing the benefits of this treatment.
Melphalan LC-MS/MS Method Development
23
SUMS seminar series
Fundamentals: Measuring concentrations of small molecules using mass spectrometry - theory and practice, Part I
Fundamentals: Measuring concentrations of small molecules using mass spectrometry - theory and practice, Part II
Fundamentals: Applications of LC/MS in small molecule drug discovery
Structures
Melphalan Internal Standard: Melphalan-D8
24
LC-MS/MS Method
o TSQ Vantage (Thermo Fisher)o Accela 1250o CTC PAL Autosampler
o ESI+o Column: Polaris A 5 µm C18 50x2.1 mmo Run time 5.6 mino Injection volume – 10 µLo Wash solutions
o Strong – methanolo Weak – 5% methanol/ 95%water
Compound Name
SRM Transition Collision Energy(v)
Melphalan 305→288305→246305→194
122234
Melphalan-D8 313→254313→293
2312
25
Time (min) %A (0.1%FA in water)
%B (0.1%FA in acetonitrile)
Flow rate
(mL/min)
0 80 20 0.25
3 5 95 0.25
3.5 5 95 0.25
3.6 80 20 0.25
5.6 80 20 0.25
Extraction Procedure
26
KimJ_120402_35370_46 - TIC - SM: 5 RT: 0.62 - 4.62 NL: 4.19E7F: + p ESI SRM ms2 305.032 [ 194.054-194.059, 246.042-246.047, 288.060-288.065]
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5Time (min)
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
Rel
ativ
e In
tens
ity
RT: 2.62
1.53 3.03 3.682.191.71 2.44 4.083.473.261.97 4.423.861.261.010.75
KimJ_120402_35370_46 - TIC - SM: 5 RT: 0.60 - 4.60 NL: 1.22E7F: + p ESI SRM ms2 313.145 [ 254.098-254.102, 296.118-296.123]
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5Time (min)
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
Rel
ativ
e In
tens
ity
RT: 2.60
1.54 2.061.84 2.42 3.462.97 3.693.16 4.25 4.523.931.00 1.310.79
o 100 µL plasmao 10 µL melphalan spiking solutionso 10 µL IS (500 ng/mL Melphalan-D8)o 400 µL ice cold methanolo Vortex, centrifugeo Transfer to the injection vialo 10 µL injected
Melphalan
Melphalan-D8
27
Melphalan LC-MS/MS Method Validation
Melphalan LC-MS/MS Method Validation Plan
28
Experiment QC level (ng/mL)LLOQ LQC MQC HQC
2 6 100 3200
Within-run precision and accuracy n=6 n=6 n=6 n=6
Between-run precision and Accuracy (3 runs) n=6 n=6 n=6
Calibration Curve (10 points – 1 per run) Back calculated concentrations
Acceptance criteria ±20% ±15% ±15% ±15%
Selectivity and Specificity 6 different lots (<20% LLOQ)
Recovery 3/3 3/6 3/6Acceptance criteria NA NA NA
Experiment QC level (ng/mL)
LLOQ LQC MQC HQC2 6 100 3200
Short-term stability (bench) n=1 for all calibration points
Acceptance criteria ±20% ±20% ±20% ±20%
Post-preparative Stability (reinjection) n=6 n=6 n=6
Acceptance criteria ±15% ±15% ±15%
Freeze-Thaw Cycles (n=3) n=6 n=6 n=6
Acceptance criteria ±20% ±20% ±20%
Selectivity
6 different lots of blank human plasma were tested
Some endogenous interference at the retention times of the analyte was observed.
Blank plasma from lot#97 was not used for the method validation.
Acceptance criteria: ≤ 20% LLOQ
29
Lot #Measured concentration
(ng/mL) % of LLOQ
95 0.19 9.4
96 0.11 5.6
97 1.39 70.3
1 0 0
2 0 0
3 0 0
KimJ_120329_35370_38 - TIC - SM: 5 RT: 0.57 - 4.57 NL: 1.59E5F: + p ESI SRM ms2 305.032 [ 194.054-194.059, 246.042-246.047, 288.060-288.065]
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5Time (min)
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
Rel
ativ
e In
tens
ity
RT: 2.57
1.54
3.67
3.522.13
1.73
2.932.311.86
4.46
3.09 3.35 4.124.013.20
0.92
KimJ_120329_35370_38 - TIC - SM: 5 RT: 0.58 - 4.58 NL: 1.76E7F: + p ESI SRM ms2 313.145 [ 254.098-254.102, 296.118-296.123]
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5Time (min)
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
Rel
ativ
e In
tens
ity
RT: 2.58
1.59 2.03 2.23 3.181.77 2.97 3.53 4.103.84 4.29 4.480.78 1.13 1.39
It is important to screen blank plasma
endogenous
Representative Calibration Curve Data (3 runs)
30
Nominal Concentrations
(ng/mL) Mean CV (%) Bias (%)
2 2.02 2.39 1.08
4 3.91 2.95 -2.17
10 9.8 2.49 -2.40
20 19.1 1.99 -4.67
50 48.7 2.07 -2.53
100 101 1.99 0.73
200 198 0.291 -0.83
400 420 4.00 4.88
Melphalan
Y = 0.0171088+0.00829092*X R^2 = 0.9991 W: 1/X^2
0 50 100 150 200 250 300 350 400
ng/mL
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
3.6
Area
Rat
ioCalibration curve concentration range: 2 ng/ml – 400 ng/mL Lower Limit of Quantification = 2 ng/mLAcceptance criteria: LLOQ ± 20%; Other levels: ±15%
Calibration curve parameters Mean CV(%)
slope 0.008 2.08
Y-int 0.016 7.13
r2 0.998 0.17
QC Precision and Accuracy Data
Within-run and between-run precision and accuracy were assessed by calculating daily and overall concentration values for six replicates of the LLOQ, Low-, Mid- and High-level validation QC samples extracted and analyzed on the same day and in three separate runs
Acceptance criteria: Ø LLOQ - %CV and % Bias ±20%Ø %CV and % Bias ±15%
31
LLOQ LQC MQC HQC
Conc. (ng/mL) 2 6 100 3200
Aliquot (mL) 0.1 0.1 0.05 0.01
Dilution factor 1 1 2 10
Within-runMean 1.95 6.39 100 2989% CV 10.3 9.6 7.71 8.63
% Bias -2.50 6.53 0.00 -6.59Between-run
Mean 1.98 6.42 104 3289% CV 9.03 5.79 6.41 8.93
% Bias -0.94 6.95 4.04 2.78
Calibration curve range 2 – 400 ng/mLLLOQ = 2 ng/mLMethod ULOQ = 4000 ng/mL
Melphalan Extraction Recovery
Recovery samples o Extracted QC samples (Extracted )o Control samples: Extract blank plasma spiked with IS and analyte added after extraction
(Post-Extracted)Absolute recovery determined based on analyte/IS area ratio
32
QC % Recovery
Low 97.1
Mid 100
High 92.2
Short-Term StabilityBench-top (short-term) matrix stability
Melphalan stored in human plasma on the bench at ambient temperature for ~ 4 hoursSamples at 2ng/mL and 200ng/mL were outside acceptance criteria (>20%)Study plasma samples should be collected and frozen within 4 hours of the sample
collection time point
Autosampler (re-injection reproducibility)Melphalan is stable in human plasma extracts held at 10°C for 24 hoursEither an entire run containing analytical standards or individual samples may be
reinjected as needed during this time period without compromising the accuracy of the determination
Freeze-Thaw CyclesNo significant changes in the measured melphalan concentrations outside the expected
accuracy for the method (% Bias ±20%)Melphalan is stable for at least 3 Freeze-Thaw cycles in human plasma when stored at
-80 ± 10°C.33
Melphalan LC-MS/MS Method Validation Summary
34
Standards QCs (Low, Mid and High) LLOQWithin-run Precision (%CV) - 7.71 – 9.6 10.3
Within-run Accuracy (% Bias) - (-6.59) – 6.53 (-2.50)
Between-run Precision (%CV) 0.29 – 4.0 5.79 – 8.93 9.03
Between-run Accuracy (% Bias) (-4.67) – 2.25 2.78 – 6.95 (-0.94)
Analyte Recovery % 92.2 - 100
Selectivity: Matrix Blanks Low interference (≤20% of LLOQ)
StabilityPost Preparative 24 hours at 10°C in autosamplerShort Term (Bench): 4 hours at room temperature in plasma <4 hours at room temperature in plasma
Freeze-Thaw Cycles 3 cyclesIntermediate Solution Stability (Analyte) < 7 months at -80°C
35
Resultsü Bioanalytical method for determination of melphalan in human plasma
was shown to be accurate, precise and selective over the concentration range from 2 to 4000 ng/mL
ü Melphalan is stable in human plasma under conditions tested
ü This method is suitable for human plasma samples analysis
36
Endogenous/Biomarker Method Validation Recommendations
37
Endogenous/Biomarkers: Method Development and Validation Recommendationso No consensus or guidelines for the assay validation approaches
o Utilize the fit for purpose strategy to identify the most important elements of method development and validation studies
o Same workflow as for drug method validationØ Method development, exploratory or definitive method validation, in-
study sample analysis
o Critical to ensure integrity of data when determining biomarker concentrations
o Need to understand what is being measured, biological relevance, and data limitation in a given assay
38
Endogenous/ Biomarkers: Calibration Curveq Calibration curve selection and preparation are central for the method development,
validation and its applications○ Evaluate suitability of biological matrix
Ø Matrix may contain residual levels of endogenous analytes which will effect accuracy
Ø Levels may vary because of age, gender, disease etc.Ø Screening for biological matrix with low levels of endogenous analyte which will
not interfere with quantification of analyte (≤20% LLOQ)
o Evaluate surrogate matrix (neat, artificial, stripped matrices) Ø Stable isotope internal standards can be useful Ø Investigate if concentration-response relationship of the analyte in the study
sample matrix is similar as that in a substitute matrix
39
Endogenous/Biomarkers: QC Samples/Stabilityo QC Samples
Ø The concentrations of QCs should represent expected study concentrationsØ The endogenous concentration of the analyte in biological matrix evaluated
before QC preparationØ Validation QCs include both: un-spiked and spiked with known amount of
authentic analyteØ Total QC concentrations validated before running samplesØ Alternate QC samples may be used for validation
o Stability (case by case)Ø Short-term and bench-topØ AutosamplerØ Long-term stability
40
Case Study 2
Global DNA Methylation as a Risk Marker for Cancer and Nutritional Health.
Yajing Hu1, Ludmila Alexandrova2, Allis S. Chien2, Megan P. Hitchins1; (1)Departments of Medicine, Stanford Cancer Institute, (2)Vincent Coates Foundation Mass Spectrometry Laboratory, Stanford University, Stanford, CA
41
Background
DNA methylation in mammals occurs at the 5-carbon site of the cytosine pyrimidine ring within CpG dinucleotides
Methylation states can be modified by environmental or life-style factors and nutrients
Epigenetic modifications including DNA methylation are potentially reversible, however, it remains unknown whether dietary interventions resulting in significant weight loss will reverse pathological global methylation states to normal levels
The goal of the study is to develop, optimize and validate liquid chromatography mass spectrometry (LC-MS/MS) method to accurately and reproducibly quantify 5-methyl-2’-deoxycytidine (MdC), 2’-deoxycytidine (dC) and 2’-deoxyguanosine (dG) in hydrolyzed DNA to determine global DNA methylation levels in clinical samples
MdC
Development and Validation of LC-MS/MS Method for Quantification of MdC, dC and dG in Hydrolyzed DNA
42
LC-MS/MS Method Development (Fit For Purpose)
43
SUMS seminar series
Fundamentals: Measuring concentrations of small molecules using mass spectrometry - theory and practice, Part I
Fundamentals: Measuring concentrations of small molecules using mass spectrometry - theory and practice, Part II
Fundamentals: Applications of LC/MS in small molecule drug discovery
44
Nucleoside Structures
MdC (5 Methyl 2’-deoxycytidine)
dC (2’-deoxycytidine)
dG (2’-deoxyguanosine)
15N3-dC
15N5-dG
LC-MS/MS Method
o Shimadzu LC/MS 8030o ESI+o Shimadzu LC-20ADXR Prominenceo Column: Synergy Fusion-RP 80Å, 150x2.0 mm;
4µm particles (Phenomenex)o Column T = 40°Co Flow rate: 0.2 mL/min
• SolventsA: 0.1% formic acid water
• B: 0.1% formic acid acetonitrileo Gradient: 0% to 80% B in 9 minuteso Run time 13 mino Injection volume – 10 µL
CompoundName
SRM Transition
Collision energy (V)
Dwell time (ms)
MdC 241.80>126.10 -12 100
dC 228.00>111.95 -12 50
dG 267.80>151.95 -15 50
15N3-dC 231.00>115.10 -10 50
15N5-dG 275.00>157.10 -10 50
45
DNA Extraction and Hydrolysis Procedure
DNA extraction and HydrolysisØ Buffy coat DNA was extracted using standard
phenol chloroform methodØ 0.5 µg of DNA was then hydrolyzed by a 3-
step enzyme digestion
Calibration curveØ Calibration curve was prepared by spiking
standard solutions into DNA buffer mixture and was treated in the same way as DNA samples including enzymatic hydrolysis.
46
q Internal standards 15N3-dC and 15N5-dG were added to the hydrolysates.
q DNA digests and calibration standards were diluted with water and filtered to remove large proteins.
q Prepared samples were injected into LC-MS
DNA amount and hydrolysis protocol optimized during LC-MS/MS method development
Calibration Curve Ranges and QC Samples
47
Analyte Calibration ranges
MdC 0.746 – 299 ng/mL 3.1 nM - 1.24 µM
dC 18.7 – 1490 ng/mL 82.4 nM – 6.56 µM
dG 18.7 – 1490 ng/mL 70.0 nM – 5.58 µM
QC DNA Methylation Level
QC1-LLOQ Human non methylated DNA
QC2-Mid Pooled human PBL DNA
QC3-High-1 Human fully methylated DNA from HCT 116
QC4-High-2 Human fully methylated DNA from Millipore
PBL: peripheral blood lymphocyte DNA
Non methylated, fully methylated human DNA standards and pooled human PBL DNA ( baseline) were used for the validation
Calculation of Global DNA Methylation Level
• DNA methylation (%) = MdC/(MdC+dC)• DNA methylation (%) = MdC/dG
48
Assay # 1 %MdC=MdC/(MdC+dC) %MdC=MdC/dGQC1
Mean 2.64 2.78CV (%) 2.92 2.55
QC2Mean 4.20 4.43CV (%) 2.74 3.49
QC3Mean 4.99 5.38CV (%) 1.30 1.03
QC4Mean 5.61 6.04CV (%) 4.32 5.68
49
LC-MS/MS Method Validation (Fit For Purpose)
Method Validation Plan
50
Experiment QC for DNA methylation QC1
LLOQQC2Mid
QC3High-1
QC4High-2
Within-run precision n=6 n=6 n=6 n=6
Between-run precision (3 runs) n=4 n=4 n=4 n=4
Calibration Curve (10 points – 1 per run) Back calculated concentrations
Acceptance criteria ±20% ±15% ±15% ±15%
Experiment QC for DNA methylationQC1
LLOQQC2Mid
QC3High-1
QC4High-2
Recovery 3/3 3/3 3/3Acceptance criteria NA NA NA
Selectivity/Specificity/Matrix effectEnsure chromatographic separation for DNA and RNA nucleosides for potential
interference
Post-preparative Stability (reinjection 72hr) n=6
Acceptance criteria ±15%
Freeze-Thaw Cycles (n=3) n=4
Acceptance criteria ±20%
51
Selectivity, Specificity and Matrix effect (Fit For Purpose)
LC-MS nucleoside separationEnsure chromatographic separation for DNA and RNA nucleosides for potential interference
0 50 100 150 200 250Concentration (ng/mL).
0.00
0.25
0.50
0.75
1.00
1.25
Area
Rat
io
Representative MdC Calibration Curve Data: Precision and Accuracy (3 runs)
52
Nominal Concentrations
(nM) Mean CV (%) Bias (%)
3.11 3.35 4.2 7.77
6.22 5.40 6.8 13.0
15.5 16.0 0.7 2.86
31.1 31.7 6.8 1.93
62.2 63.2 10.8 1.77
124 118 7.2 -5.24
311 325 2.8 4.42
498 505 2.0 1.92
622 612 5.5 -1.34
1244 1263 7.3 1.33
MdC
Calibration Curve parameters Mean CV (%)
slope 0.0041 5.53
y-int. 0.0005 7.47
r2 0.9940 0.14
Y = (0.00420620)X + (-7.91350e-005) R^2 = 0.996 W: 1/X^2
Calibration curve concentration range: 3.11 nM – 1.24 µM Lower Limit of Quantification = 3.11 nMAcceptance criteria: LLOQ ± 20%; Other levels: ±15%
QC Precision Data
Within-run precision was assessed by calculating CV values for six replicates of the validation QC samples extracted and analyzed on the same day
Between-run precision was determined by calculating CV values for 4 replicates of validation QC samples extracted and analyzed on 3 different days
Acceptance criteria: Ø LLOQ - %CV ±20%Ø %CV ±15%
53
DNA Methylation %MdC=MdC/(MdC+dC)QC1 QC2 QC3 QC4
non methylated pooled blood
Fully methylated
HCT116
Fully methylated
MPAliquot (ug) 0.5 0.5 0.5 0.5
Dilution factor 1 1 1 1
Within-runMean 2.97 5.00 6.11 6.31% CV 5.1 4.58 2.60 3.45
Between-runMean 2.81 4.48 5.87 5.80% CV 7.2 12.1 13.3 7.7
Extraction Recovery: MdC, dC and dG
Recovery samples o Add analyte and internal standard prior to
hydrolysis and extraction (Extracted )o Control samples: Extract sample spiked with
IS and analyte added after hydrolysis and extraction (Post-Extracted)
Absolute recovery determined based on analyte/IS area ratio
54
nM % Recovery MdC
Low (15.5) 100Mid (125) 103High (622) 110
dCLow (81.9) 100Mid (656) 93.3
High (3290) 107dG
Low (69.7) 103Mid (558) 96.3
High (2790) 107
Freeze-Thaw Stability
DNA methylation values are stable when stored at -80°C for at least 3 freeze/thaw cycles
No significant changes in the measured DNA methylation values outside the expected precision and accuracy of the method were found, indicating that 3 Freeze-Thaw cycles did not affect the sample integrityAcceptance criteria: Ø %CV and % Bias ±20%
55
DNA Methylation %MdC=MdC/(MdC+dC)
QC2pooled blood
Aliquot (µg) 0.5Dilution factor 1
Freeze-Thaw 1Mean 4.20% CV 2.74
Freeze-Thaw 3Mean 4.27% CV 2.96
Global DNA Methylation: Method Validation Summary
56
Standards QCs (Low, Mid and High) LLOQWithin-run Precision (%CV) - 2.6 - 4.58 5.1
Within-run Accuracy (% Bias) - - -
Between-run Precision (%CV) 0.7 – 10.8 7.7 – 13.3 7.2
Between-run Accuracy (% Bias) (-5.24) – 13 - -
Analyte Recovery % 93.3 - 110
Selectivity, Specificity, Matrix effect Low interference
StabilityPost Preparative 72 hours at RT in autosampler
Freeze-Thaw Cycles 3 cyclesIntermediate Solution Stability (Analytes) 4 months at -80°C
57
Pilot Study: Effect of Low Fat vs. Low-Carbohydrate Diet on Global DNA Methylation
Ø DNA methylation levels were determined in 16 paired clinical samples collected during a trial of obese individuals prior to, and following the dietary intervention (Prof. C. Gardner)
Ø Results show trend of reduced methylation level 12 months after intervention (p-value=0.069, Wilcoxon signed ranks test).
DNA methylation (%) MdC/(MdC+dC)
Diet Type Baseline 12 months Difference
Low Carb 4.31 4.05 -0.26
Low Carb 4.18 3.94 -0.24
Low Carb 4.6 3.8 -0.80
Low Carb 4.06 3.92 -0.14
Low fat 4.2 3.9 -0.30
Low fat 3.96 4.07 0.11
Low fat 4.06 4.05 -0.01
Low fat 3.77 3.82 0.05
ResultsLC-MS/MS method to accurately and reproducibly measures global DNA methylation
levels developed and validated.
Fit for purpose approach was used to validate and monitor assay performance.▪ DNA buffer matrix was evaluated and used for calibration curve preparation.▪ Non-methylated and fully methylated human DNA commercial standards were
utilized as quality control samples
LC-MS/MS method was successfully applied to a pilot diet intervention study to investigate global DNA methylation levels as a potential marker to monitor health/disease status
Preliminary results showed trend of reduced methylation levels 12 months after dietary intervention (p-value=0.069, Wilcoxon signed ranks test)
58
ConclusionØ Method validation is an important tool in drug discovery and development
Ø Assay performance parameters by their design may give an idea of method capabilities and limitations that can be experienced during routine sample analysis
Ø Fit for purpose validation is faster and less labor intensive and used extensively in drug discovery and biomarker research
Ø A practical fit for purpose validation delivers quality results to make informed project decisions
59
ReferencesLiterature:
○ Bioanalytical Method Validation; Guidance for Industry -https://www.fda.gov/media/70858/download
○ Handbook of LC-MS Bioanalysis. Best Practices, Experimental Protocols and Regulations by Li W., Zhang J., and Tse F.L.S.
○ Biomarkers in Drug Discovery and Development: A Handbook of Practice, Application, and Strategy, Second Edition by Rahbari R., Van Neiwaal J. and Bleavins M. R.
Standards:○ Cambridge Isotope Labs - www.isotope.com○ Cerriliant - www.cerilliant.com○ Anaspec – www.anaspec.com○ Sigma-Aldrich – www.sigmaaldrich.com
60
Acknowledgements
SUMS staff
Stanford Dean of Research
Vincent and Stella Coates Foundation
61
Proteomics
Ryan Leib, PhDFang Liu, PhD
Rowan Matney, BAKratika Singhal, MS
Protein IDQuantitative – PRM, TMT
Protein interactionsProtein structure
Phosphoproteomics, PTMs
Small Molecule
Ludmila Alexandrova, PhDKarolina Krasinska, MS
Beryl Xia, PhD(Theresa McLaughlin, MS)
Absolute quantitationTargeted metabolomics
PK, Stability studies Metabolite ID Intact protein
Open Access
Theresa McLaughlin, MS(Beryl Xia, PhD)
GC/MSLC/MSHRMS
MALDI-TOF
Director – Allis Chien, Ph.D.
Applications
People
SUMS Group
Accounting Associate – William Lee, BS62
July 9, 2020In-depth shotgun triacylglyceride profilingSpeaker: Matias Cabruja, PhD – Dept. of Genetics, Stanford University School of Medicine
July 16, 2020Improved protein and PTM characterization with a practical electron-based fragmentation on Q-ToFs and ion mobility separationsSpeaker: Prof. Joe Beckman, PhD – Dept. of Biochemistry & Biophysics, Oregon State University
Stay tunedMore webinars on Thursdays, noon-1pm
Details & registration: mass-spec.stanford.edu/events
Requests welcome!
63
http://mass-spec.stanford.edu
64
We are Stanford University Core Lab
iLab help: - SUMS website - William Lee
65