CONFIDENTIAL
Quantitation of a Therapeutic Insulin Analogue by Immuno-Affinity Extraction LC-MS/MS
(or running a real peptide project with a hybrid assay: questions and some solutions?)
Michael BlackburnCovance Laboratories Ltd
Alnwick, UK
5th EBF Open Symposium14th November 2012
Barcelona, Spain
CONFIDENTIAL
Acknowledgements
• Martin Gerl and his team, Sanofi, Frankfurt : Scientific leadership of this project, supply of IA-gel.
• Dietmar Schmidt, Ronald Schmidt and team, Sanofi, Frankfurt: Dialogue, expertise with insulins and training in Immuno-Affinity extraction.
• Isotope Chemistry Group, Sanofi, Frankfurt: Production of iodated and stable labelled internal standards
• Barbara Bell, Nick Gray and Jennifer Mackie, Alnwick: Method development, sample analysis.
• Stuart McDougall: leadership of the Alnwick team.
CONFIDENTIAL
Insulin Analogues• Glargine, glulisine, lispro, aspart etc, are modified insulins, also known as insulin
analogues (IAs).• IAs have a modified isoelectric point (pI) which reduces solubility at physiological pH.
• Changing pI changes adsorption into body: get fast & slow-acting insulins
Longer Lasting Insulin Development Compound: LoLA
• Molecular Weight 6.2 kDa approximately: modified at A & B chain• Requirement for a reliable LC-MS/MS assay in animal & human plasma • LLOQ requirement of 200 pg/mL in plasma or better
Structure of Human Insulin: 51 amino acids, a large peptide
G I V E Q C C T S I C S L Y Q L E N Y C N
F V N Q H L C G S H L V E A L Y L V C G E R G F F Y T P K T
S S
S
S
S
S
NH2-
NH2-
-COOH
-COOH
CONFIDENTIAL
LoLA Q1 Mass Spectra & Manipulation of Charge StateSAR161271_full_scan #21-33 RT: 0.38-0.61 AV: 13 NL: 2.62E6T: + p ESI Q1MS [ 750.000-1400.000]
800 900 1000 1100 1200 1300 1400m/z
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
Rel
ativ
e A
bund
ance
782.85
1043.56
894.80 1252.16
900.20
905.42883.911002.52871.31 962.29 1202.84 1290.491154.431095.48 1339.36
6+5+7+
8+
20mM ammonium formate/acetonitrile LoLA charge distribution envelope: Methanol/formic acid/ipa 10%
5+
6+
7+
6+
5+
4+•Charge state envelope can be to some extent controlled
using mobile phase composition
•Some difference observed between different source typesWaters Z spray, Thermo, also APi5000 at Frankfurt
Waters Z spray, Quattro Ultima
Thermo Quantum Ultra
CONFIDENTIAL
Direct Assay (physico-chemical assay) Approaches
• A combination of solid phase extraction of plasma (1mL) combined with LC/MS/MS• MS/MS transition chosen depended on cluster chosen for fragmentation• Initially used 1252 > 219• Extraction by Waters Oasis HLB• Limit of quantitation using a direct assay was 500-1000 pg/mL LoLA from 1mL plasma• For all assays used the addition of a surfactant to stop binding of standard solutions
RT: 2.94 - 7.00
3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5Time (min)
0
10
20
30
40
50
60
70
80
90
1000
10
20
30
40
50
60
70
80
90
100
Rel
ativ
e A
bund
ance
500pg/mL Lantus in rat plasma LoLA ISTD
RT: 5.32
6.06
6.305.40 6.346.01 6.383.08
5.56 5.915.103.26 4.493.73 4.434.01 4.823.576.976.61
RT: 4.99
5.224.56 6.375.42 6.276.063.49 4.463.953.30 3.76 5.84 6.976.71
NL: 2.10E2TIC F: + c ESI SRM ms2 [email protected] [174.850-175.150] MS 30March09_013
NL: 2.80E3TIC F: + c ESI SRM ms2 [email protected] [218.950-219.250] MS 30March09_013
Direct physico-chemical assay did not givethe required sensitivity
Instrumentation: TSQ Quantum Ultra
Insulin analogue spiked into rat plasma & extracted
CONFIDENTIAL
Sensitivity Enhancement 1: Use of FAIMS-pSRM
Useful for clusters which fragment poorly
In this example fragmentation is relatively poor; at higher collision energy the molecule ‘unzips’
i.e. produces lots of low intensity fragments
Can use the full signal from the precursor ion for best sensitivity and improve selectivity using an alternativetechnique: Ion mobility spectrometry
Combine a specific, high molecular weight Cluster with ion mobility selection(with some collision energy)
and ‘pseudo’ SRM = high energy Q3 SIM
Or can use a 0 constant neutral loss experiment(0 CNL). Gives an alternative specificity for poorlyfragmenting peptides
CONFIDENTIAL
Enhancement 2: Immuno-affinity Extraction
Cleaner extract than SPE
Columns can be re-used: >20 x
Expensive, a little slower
Throughput = 1 plate/analyst per day
Immunoaffinity image: services.leatherheadfood.com
Procedure adapted to use
multi-channel pipettes to
enhance throughput
CONFIDENTIAL
SPE vs IAC (or small & large)
SPE• Commercially prepared columns• Used once• Can be automated
Procedure• Condition column solvent• Equilibrate column aqueous• Load sample• Apply wash (aqueous solvent)• Elute Compound • Discard column/plate
IAC• Columns prepared in the lab•Columns are re-used•Probably best to run manually
but can use multi-channel pipettesProcedure
•Wash column with elution buffer•Wash column with buffer•Slowly apply sample to column•Wash with buffer•Elute with elution buffer (acid/ethanol)•Wash & prepare for storage
Very Good Tip:
Don’t put IAC gels in the freezerWe used Oasis HLB for initial
method development in a vacuum manifold.
CONFIDENTIAL
Result: Hybrid Immunocapture – LC-MS Approaches
• IAC extracts were cleaner, with absolute recovery approximately 80-90%• Limit of quantitation using immuno-capture assay was 200 pg/mL LoLA from
250uL spiked plasma. Immuno-capture gives a more sensitive assay.• This was true whether using FAIMS-pSRM on a Thermo system or MS/MS
on the API5000 system. • Enzyme digestion (Edman) was investigated as a back-up: did not improve
signal vs ‘top down’ approachRT: 3.03 - 5.99 SM: 5B
3.5 4.0 4.5 5.0 5.5Time (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 A
bund
ance
100pg/mL plasma IAC 60uL Laminar flow
RT: 5.29MA: 2248
5.75
5.66
5.50
3.204.89 5.15
5.073.69 4.71
4.34 4.644.23 4.583.553.40
4.124.06
NL:5.43E2TIC MS 15april09_007
100pg/mL LoLA from spiked plasma
Decision: ‘ Hybrid’ IA- LC-MS approach selected for longer lasting insulin project
CONFIDENTIAL
Internal Standards
• Initial method development used an analogue ISTD (e.g. bovine insulin, insulin glargine)
• For early studies used iodated compoundThis ISTD performed well and we did not see evidence of loss of iodine and conversion to parent compound during the period of use. Recommend regular checking of this.
• GLP Studies used 15N isotopically labelled compound: all nitrogens labelled
• Concentration ISTD 2ng/mL (assay range:0.2 – 5 ng/mL)
CONFIDENTIAL
FAIMS method LLOQ & LC conditions
LoLA
I-LoLA ISTD
LLOQ in animal plasma (0.2 ng/mL)
TSQ Vantage with Turboflow system & FAIMS unit
•Load 100uL of extract onto an Oasis on line extraction column HLB 2.1 x 20mm •Elute onto Hichrom Ace 3 C18, 50 x 2.1 mm, 3 μm•Mob Phase A: 20mM ammonium formate pH 3•Mob Phase B: acetonitrile/0.1% formic acid flow rate: 0.4 mL/min•Sample volume 250 uL; injection volume 100 uL
NOTE: 0.2 ng/mL is still a rather high LLOQ for this work. Some low dose levels were not properly covered.The ideal LLOQ was 100 pg/mL or lower
CONFIDENTIAL
Correlation of MS with LBA Assay• IAC-LC-FAIMS-MS vs ELISA values for 16 samples from discovery
study (discovery LC-MS assay)
R2 = 0.9178
0
1000
2000
3000
4000
5000
6000
0 1000 2000 3000 4000 5000 6000
ELISA
"LC-MS"
Both x & y axes are LoLA concentration pg/mL
CONFIDENTIAL
Acceptance Criteria
•As this was a hybrid assay, method validated using 20/25 ‘biologicals’ criteriaDecision based on assay complexity (immuno-capture extraction)Also we are at the limits of instrument performance due to the size of analyte
•Same criteria used at Frankfurt. These were inspected and accepted by FDA•Analytical Range 0.2 – 5ng/mL
LoLA concentrations rarely higher than 2 ng/mL •As part of method development/validation we checked analytical cross-over between
LoLA and other insulin analogues
IA-LC-FAIMS pSRM method was used for discovery and GLP pre-clinical studies.
Need to move to a clinical assay. Some concerns about matrix variability in human plasmawith pSRM method.
Precision and Accuracy Data for first GLP Study (IAE-LC-FAIMS-pSRM).
Nominal concentration (ng/mL)
Number of samples Mean calculated concentration (ng/mL)
Mean % Difference Total%CV
0.2 18 0.198 -1.06 22.2
0.5 18 0.512 2.30 18.8
2 18 2.08 3.83 9.06
5 18 4.81 -3.79 8.37
CONFIDENTIAL
Product Ion Mass Spectra for long lasting insulin
SAR161271_dau_spec #108-191 RT: 3.42-6.02 AV: 84 NL: 8.87E4T: + p ESI Full ms2 [email protected] [ 100.070-1050.000]
200 300 400 500 600 700 800 900 1000m/z
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
Rel
ativ
e A
bund
ance
219.1
120.1
1043.8
328.9
251.3 639.5
999.1573.7201.1 362.0
555.0293.3 476.3 983.4
379.9 656.2624.3757.2426.2 527.6 901.4862.9681.7 949.3185.3
B (a2)
Non-specific immonium ions should be avoided where possible: eg. 120 PheSpecific ions from B chain a & b series are available for insulins (Thevis & Schanzer, 2007).Lower charged clusters eg 1252 5+ did not fragment so well, 6+ have higher internal energy and fragment more readily. Need to compromise specificity vs ease of fragmentation.
B chain fragmentations Reference:. Thevis and Schanzer Analyst 2007, 132, p287-291
[M+6H+]6+
CONFIDENTIAL
Clinical Studies: LC-MS/MS Method
•Changed mobile phase to boost ratio of 6+ charge state & increase signal
•Used diagnostic 219 product ion(possible because of clean IAC extract)
•15N labelled internal standard available
drug
ISTD
LLOQ in plasma (0.2 ng/mL)
Mob. Phase B:Methanol/Iso-propanol/formic acid 0.5%Mobile Phase A: 0.5% Formic acidColumn Thermo Hypersil Gold 1.9um 20 x 2.1mmAgilent 1200, Stainless steel LC tubingSemi-UPLC method: Pressure approx 5000psi/360barFlow rate 0.3 mL/min500 uL sample volume; 50 uL injection volume
NOTE: 0.2 ng/mL is still a rather high LLOQ for this workSome low dose levels were not properly covered,particularly moving into clinical studies.Ideally LLOQ should be 100 pg/mL or lower
CONFIDENTIAL
Anti-Drug Antibody Effects: Repeat Dose Study
• Immunogenicity - Any biological therapeutic agent may have an unwanted immune response. This can range from having no effect to a severe hypersensitivity reaction. Immunogenicity can also result in neutralisation of the drug ie no clinical benefit, or at least a reduction in response and a change in pharmacokinetics.
• Immune response must be monitored when there is more than 7 days of drug exposure. All biotherapeutics need to be assessed for immunogenicity by measuring ADa – anti drug antibodies.
• Differences in insulins across species may well produce ADaeffects
• LoLA: Repeat dose 28 day preclinical studyDay One Samples: variability of internal standard around 10-30%, recovery highDay 28 samples internal standard recovery was low (< LLOQ of LoLA) and highly variable Evidence of an ADa effect, antibodies were also measured directly
CONFIDENTIAL
Effect of ADa in multi-dose studies
• Antibodies to LoLA formed from 2-3 weeks exposure depending on dose
• These bind to drug & internal standard (iodated LoLA in pre-clinical studies)
• Equilibrium between free (active) and total insulins• Request for a ‘total’ method for bound & unbound insulin• Ab complex is dissociated by acidification with HCl• Leave 1 hour followed by neutralisation and then IAC• Total method has an LLOQ of 1 ng/mL, HLOQ 100 ng/mL• Pre-clinical multi-dose studies used ‘Total’ method
‘free’ LoLA could be assessed by filtration or addition of excess insulin to bind AD-antibodies
CONFIDENTIAL
Metabolite Assays
• Much of insulin activity can be attributed to metabolites of the compound
• Peptides have ‘catabolism’ not metabolismStep-wise loss of amino acids from B chain preserves insulin type activity
example:
• IAC antibody extracts based on insulin epitope on the B chain
will also extract metabolites therefore & we can quantify these and gain extra information regarding compound activity.
M1M2M3
CONFIDENTIAL
Exploratory Metabolite Assay
M1 M2
M3 M4
ISTDLo-La
•These were synthesised in Frankfurt
•Used same LC conditions and extraction procedure as clinical method
•Added MS/MS transitions for major metabolites
•Note for this sample there is no cross-over ofMetabolites > LoLA
•Ran selected samples at higher concentrations to give extra information regarding activity & safety
M1 – M4 5ng/mL, spiked into human plasma, LoLA blank
CONFIDENTIAL
Considerations for LC-MS Assays for Insulins• No waiting required for development of specific antibodies for
immunoassays – get information faster• Greater assay specificity relative to LBA gives confidence and can be used
to generate new informationLevels of different metabolites give insight into action of the analogueNo cross reactivity with other insulin analogues (tested glulisine and 4 other insulin analogues for reactivity) or endogenous insulin (tested at 50 ng/mL)Improved instrumentation (sensitivity, increasing mass range) allows us to reach required LLOQsUsed halogenated & stable labelled internal standard
• IAC gives clean extracts to allow lower LLOQs for insulins and is surprisingly quick and robust
Must consider possible ADa effect multi-dose, esp. across species. So look at ‘total’ or free insulins?
• Acceptance criteria need careful consideration for these assaysSmall molecule limits may not be appropriate or achievable
• Peptide and small protein analysis (<8KDa?) by LC-MS/MS is becoming the standard
CONFIDENTIAL
References• POSTER: TT26B: Experiences on LC-MS Analysis of Large Molecules. D.
Schmidt et al., EBF, Barcelona 2012• POSTER: Quantitation of a Therapeutic Insulin Analogue by Immuno-
Affinity Extraction LC-MS/MS. M Blackburn et al., EBF, Barcelona 2012• Mass spectrometric identification of peptide hormones in doping control
analysis. Thevis and Schanzer., Analyst, 132, 287-291 (2007).• Quantitative Analysis of Biomarkers, drugs and toxins in biological samples
by immunoaffinity chromatography coupled to mass spectrometry or tandem mass spectrometry: A focused review of recent applications. Tsikas, D.I., J. Chromatography B., (2010).
• FAIMS: GLP-Validated quantitation method for a peptide in biological matrix. Klaasen, T., and Romer, A. ASMS Poster Presentation, Denver 2008.
• A Mass Spectrometric Method for quantitation of intact insulin in Blood Samples. Darby, S.M, Miller, M.L., Allen, R.O and LeBeau, M. Journal of Analytical Toxicology, Vol. 25, 2001.
• Clinical significance of insulin antibodies in insulin-treated diabetic patients. Van Haeften, TW, Diabetes Care 12:641-648, 1989.
THANK YOU FOR LISTENING!