Comparison of quantitative assays for a large protein:

ELISA vs. LC-MS/MS

A. Guenzi, G. Fischer, N. Justies, B. Lausecker F. Hoffmann - La Roche Ltd., Basel, Switzerland

ELISA

2

ELISA (proteins and peptides)

• ADVANTAGES:

– Highly sensitive

– Quick and easy to perform

– Inexpensive instrumentation and material

– If anti-idiotypic Abs used, only ‘active’ is detected

• DISADVANTAGES:

– Need for very specific antibodies: animal immunization, selection and purification of

reagents

• For small peptides specific Abs may be difficult to find

– Small differences in analyte sequence may require development of new reagents

– Batch-to-batch variability of reagents

– Each ELISA method needs multiple optimizations

– Each species needs optimization (cyno≠rat≠man≠…)

– Each matrix needs optimization (plasma≠serum≠…)

– Enzyme/substrate reaction is short term so read as soon as possible

3

LC-MS/MS

4

0

10

20

30

40

50

60

70

0 100 200 300 400 500

An

aly

te/I

ST

D a

rea r

ati

o

Nominal concentration (ng/mL)

LC-MS/MS (proteins and peptides)

• ADVANTAGES:

– Highly sensitive

– Quick and easy to perform

– Multi-analyte assays possible

– Species-independent

– Only ‘total’ detectable

• DISADVANTAGES:

– Multiply charged ions may reduce sensitivity

– Need for digestion (proteins) to ‘simplify’ to peptides

– Is a signature peptide formed during digestion?

– Lipophilicity of peptides

– Wide range of concentrations possible (from pg/mL to µg/mL)

– Mass range of MS instruments

5

6

ELISA and LC-MS/MS data complement each other,

they are not in competition

7

CASE STUDY:

Analysis of a protein (300 kDa including lipidation)

by ELISA and LC-MS/MS

The protein

8

A complex, trimeric fusion protein (150 kDa)

Molecular weight: 300 kDa

containing phospholipids

ELISA (PK) assay

9

POD SUBSTRAT

E

Streptavidin-coated 96-well microplate

Biotinylated capture antibody

Digoxigenylated capture antibody

Fab<Digoxigenin>POD

Validation results of the ELISA

cynomolgus plasma

• Minimum required dilution (MRD): 1%

• Validated range (in 100% plasma): 1.02 to 65.0 µg/mL

• Validated range (in assay buffer): 10.2 to 650 ng/mL

• Precision and accuracy (intra-assay): <9.2% and 89.9 – 101.3%

• Precision and accuracy (inter-assay): <4.7% and 99.9 – 111.2%

• Selectivity, specificity: within Guidance criteria

• Dilution linearity and parallelism: no hook effect up to 6.50 mg/mL

• Assay signal stability: readout immediately after addition of

SDS

10

• During application (routine analysis of studies) the

ELISA-measured concentrations did not fit with other

pharmacodynamic observations

• The possibility of exploiting for quantitative purposes

the results obtained by proteomics studies were

considered

11

An orthogonal analytical method (LC-MS/MS)

able to confirm the ELISA data was developed

Starting point for LC-MS assay:

Proteomics qualitative procedure

12

100 µL plasma

+3.9 mL sample buffer Reduction with DTT

Alkylation with

iodoacetamide

Dialyse against

40 mM Tris buffer pH 8.5

3 x 1 L

0.75 mg / mL

Tryptic digestion with 40 µg

Promega trypsin

Dilute sample for LC/MS

analysis on Orbitrap

1 mg / mL

Sample buffer:

8 M urea, 40 mM Tris pH 8.5

2% Chaps

Reduction:

Add 15 mg DTT, vortex

1 h at room temperature

Alkylation:

Add 60 mg iodoacetamide

vortex 1 h at room temperature

(in the dark)

Tryptic digestion:

overnight, room temperature

Starting point for LC-MS:

Identification of signature peptide

• Samples containing proteins (e.g., plasma) are digested

to peptides with trypsin. Peptides are selected as

surrogates of the protein by the following criteria:

The fragmentations (transitions) are unique for one single

protein

The peptide contains no Cys, Met or other commonly

modified residues

The peptides lie in the mass range 800 – 2000 MW

13

MS analysis of protein in plasma (proteomics)

14

RT: 0.00 - 117.99

0 10 20 30 40 50 60 70 80 90 100 110

Time (min)

0

20

40

60

80

100

0

20

40

60

80

100

Re

lative

Ab

un

da

nce

0

20

40

60

80

100

79.25615.4059

79.41615.4064

24.72705.7875

79.12615.4056

31.00479.7685

45.46551.9948 67.15

646.864473.62

559.631924.65

705.7892 79.64615.4043

92.93639.6291

4.16436.1826

114.32445.1201

48.26709.0074

39.54709.0040

49.43709.0073

30.85709.0087

60.55709.0022

74.65709.0103

109.25709.0016

96.65709.0035

1.02709.0119

48.261062.9988

49.051062.999639.75

1063.003530.80

1063.013461.53

1063.002979.49

1063.003811.43

1063.003889.87

1063.0020109.36

1063.0151

NL: 5.72E8

Base Peak MS RABBITPLASMA242_TETRANECTIN_02

NL: 1.36E7

Base Peak m/z= 708.9982-709.0124 F: FTMS + p NSI Full ms [400.00-1650.00] MS RABBITPLASMA242_TETRANECTIN_02

NL: 3.54E6

Base Peak m/z= 1062.9942-1063.0154 F: FTMS + p NSI Full ms [400.00-1650.00] MS RABBITPLASMA242_TETRANECTIN_02

3+

2+

Assay development LC-MS/MS

signature peptide

15

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXKEQQALQTVDEPPQSPWDR

XXXXXX

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

XXXXXX

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

XXXXXX

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

XXXXXX XXXXXXXXXXXXXXXXXXXXXXXXXXXXX

Hinge region

Trypsin cleavage after Arginine (K) or Lysine (R)

Elemental Composition: C91 H139 N26 O33

Parent Charge: 2+

Monoisotopic M/Z: 1062.50326

Parent Charge: 3+

Monoisotopic M/Z: 708.67126

Full ion scan of the signature peptide

16

RABBITPLASMA242_TETRANECTIN_02 #2507-2522 RT: 48.16-48.26 AV: 2 NL: 1.06E7F: FTMS + p NSI Full ms [400.00-1650.00]

400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600

m/z

0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

90

95

100

Rela

tive

Ab

un

da

nce

709.0056z=3

750.8569z=2

642.2901z=2

1063.0047z=2

544.6262z=3

500.9076z=3

799.4124z=2 1129.4793

z=3993.0772

z=?1209.2407

z=?1417.0044

z=31594.7757

z=?

RABBITPLASMA242_TETRANECTIN_02 #2507-2522 RT: 48.16-48.26 AV: 2 NL: 1.06E7F: FTMS + p NSI Full ms [400.00-1650.00]

708.0 708.5 709.0 709.5 710.0 710.5 711.0 711.5

m/z

0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

90

95

100

Rela

tive

Ab

un

da

nce

709.0056z=3

708.6732z=3

709.3383z=3

709.6721z=3

710.0064z=3

708.3408z=3

710.3403z=3

708.0087z=3 710.6711

z=3711.3329

z=3

RABBITPLASMA242_TETRANECTIN_02 #2507-2522 RT: 48.16-48.26 AV: 2 NL: 2.46E6F: FTMS + p NSI Full ms [400.00-1650.00]

1061 1062 1063 1064 1065 1066

m/z

0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

90

95

100

Re

lative

Ab

un

da

nce

1063.0047z=2

1062.5037z=2

1063.5059z=2

1064.0069z=2

1064.5085z=2

1061.8942z=2

1061.2846z=2

1065.0118z=2 1065.9532

z=2

Multidimensional optimization

of the LC-MS/MS assay

Optimization of the following parameters

• Volume of plasma

• Volume of buffer

• Trypsin concentration

• Protease-protein (mainly albumin) ratio

• Precipitating solvent

• Duration of incubation

• Chromatographic conditions

• Addition of internal standard

17

Example: choice of precipitating solvent

in the preliminary step

18

Example: protein precipitation parameters

19

CONCLUSION

Plasma albumin has an effect on

digestion efficiency: it must be

eliminated by precipitation followed

by centrifugation

Internal standard

• Terminal arginine residue labeled with multiple 13C and 15N

• ISTD added before digestion does not compensate for losses during the

initial protein precipitation and for digestion yield

20

C13

C13

C13

C13

C13

N

N15

OH

O

C13

N15

N15

R

Cyno plasma: final sample preparation

21

10 ml Plasma

Protein Precipitation

Add Incubation buffer

& ISTD

Tryptic digestion with 2 mg

Trypsin Protease/protein 1:350

50uL injection with

on-line dilution for LC/MS

analysis on Vantage

0.7 mg Protein

Protein precipitation

Add 25 µL methanol, mix

Incubation buffer:

Add 200 µL 50mM TRIS buffer

pH 8.30, vortex

Add 20 µL of ISTD, vortex

Tryptic digestion:

Add 20 µL of trypsin solution

(100 µg/mL in 50 mM acetic acid)

Vortex

2 hours, +40°C, stop the digestion

via addition of 10 µL of 50% formic

acid and short cooling at -20°C

with organic

Centrifugation & transfer

LC-MS/MS of the signature peptide

• The traditional approach in our group (column-switching with on-

line solid-phase extraction) was followed

• 50 µL injection volume (1.8 µL plasma-equivalents)

• Pump C to dilution pump flow-rate ratio 1:6.6

• Gradient elution with acetonitrile-water-formic acid on the analytical

column (Supelco Ascentis Express C18, 2.0 (i.d.) x 50 mm)

• Detection on a Thermo TSQ Vantage EM triple quad with H-ESI

source 22

Validation results of the LC-MS/MS assay

cynomolgus plasma

• Validated range (plasma): 1.00 to 200 µg/mL

• Precision and accuracy (intra-assay): <7.9% and 100.0 – 106.7%

• Precision and accuracy (inter-assay): <7.9% and 92.5 – 105.6%

• Selectivity, specificity: within Guidance criteria

• Dilution linearity: validated up to 1:100

• Matrix effect: compensated by SLIS, ca. 0.92

• Extraction recovery: between 93.5 and 97.8 (SLIS)

23

It does not cover

digestion!

Chromatograms

24

D:\Xcalibur\...\MS20-20100602_A008 02-Jun-2010 12:29:18 PM C01

RT: 0.00 - 5.00 SM: 7B

0 1 2 3 4 5

Time (min)

0

200

400

600

800

1000

Inte

nsity

0.972.51

1.17 4.013.413.041.67 2.303.67 4.43 4.650.640.33

NL: 1.10E3

TIC F: + c ESI SRM

ms2 708.700

[491.990-492.010]

M S

M S20-

20100602_A002

RT: 0.00 - 5.00 SM: 7B

0 1 2 3 4 5

Time (min)

0

10000

20000

30000

40000

50000

Inte

nsity

0.96 4.022.88 3.752.562.121.28 4.660.06

NL: 5.40E4

TIC F: + c ESI SRM

ms2 711.800

[496.590-496.610]

M S

M S20-

20100602_A002

RT: 0.00 - 5.00 SM: 7B

0 1 2 3 4

Time (min)

0

200

400

600

800

1000

Inte

nsity

RT: 2.93

3.35

0.96

2.491.15 4.01

1.39 2.05 3.67 4.20 4.560.32

NL: 1.06E3

TIC F: + c ESI SRM

ms2 708.700

[491.990-492.010]

M S

M S20-

20100602_A008

RT: 0.00 - 5.00 SM: 7B

0 1 2 3 4

Time (min)

0

10000

20000

30000

40000

50000

Inte

nsity

2.93

0.95 4.023.312.20 2.68 4.341.31 4.771.670.16

NL: 5.42E4

TIC F: + c ESI SRM

ms2 711.800

[496.590-496.610]

M S

M S20-

20100602_A008

RT: 0.00 - 5.00 SM: 7B

0 1 2 3 4 5

Time (min)

0

50000

100000

150000

200000

Inte

nsity

2.92

3.340.97 4.042.471.40 4.302.060.19

NL: 2.20E5

TIC F: + c ESI SRM

ms2 708.700

[491.990-492.010]

M S

M S20-

20100602_A010

RT: 0.00 - 5.00 SM: 7B

0 1 2 3 4 5

Time (min)

0

10000

20000

30000

40000

50000

Inte

nsity

2.92

0.95 4.013.282.502.18 4.271.32 4.880.38

NL: 5.27E4

TIC F: + c ESI SRM

ms2 711.800

[496.590-496.610]

M S

M S20-

20100602_A010

RT: 0.00 - 5.00 SM: 7B

0 1 2 3 4 5

Time (min)

0

200

400

600

800

1000

Inte

nsity

0.963.34

2.484.011.14 3.001.39 2.27 3.64 4.834.540.39

NL: 1.10E3

TIC F: + c ESI SRM

ms2 708.700

[491.990-492.010]

M S

M S20-

20100602_A011

RT: 0.00 - 5.00 SM: 7B

0 1 2 3 4 5

Time (min)

0

10000

20000

30000

40000

50000

Inte

nsity

2.85 4.020.95 3.742.18 4.321.15 4.841.580.53

NL: 5.40E4

TIC F: + c ESI SRM

ms2 711.800

[496.590-496.610]

M S

M S20-

20100602_A011

Blank Cyno plasma

LLOQ= 1 µg/mL

ULOQ = 200 µg/mL

Blank after ULOQ

(carry-over test)

Cross-validation results

ELISA vs. LC-MS/MS

25

0

2000

4000

6000

8000

10000

12000

14000

0 2000 4000 6000 8000 10000 12000 14000

LC-M

S (µ

g/m

L)

ELISA (µg/mL)

+15%

-15%

Identity line

Effect of high dilution

accuracy (both ELISA

and LC-MS/MS)?

Acknowledgements

Non-Clinical Safety, Roche Basel

• Katrin Schliemann

Translational Research Sciences, Roche Basel

• Axel Ducret

Pharma Research, Roche Penzberg

• Julia Heinrich

• Roland Staack

26

We Innovate Healthcare

27