AUC vs Peak Value: Impact of Data Analysis Method on the Pharmacological Parameters
of the Aequorin Assay.
Vincent Dupriez, Sandrine Ociepka, Nancy Mac Donald, Laurent Dumortier, Janet Park
M5 Aequorin kinetics - LumiLux
M5 Fluo-4 NW kinetics - CellLux
40000
250000
[Acetylcholine]
[Acetylcholine]
-6
-7
-7.5
-8
-8.5
-9
150000
100000
50000
-6
-6.5
-7
-7.5
-8
-8.5
-9
30000
RLU
200000
20000
10000
0
0
0
5
10
15
20
25
0
30
5
10
15
20
25
pEC50 = 8.81
NECA (Max-Min)
SRIF 28 (Max-Min)
pEC50 = 6.40
800,000
AUC ( RLU)
100,000
600,000
75,000
400,000
50,000
200,000
25,000
0
-8
-7
-6
-5
-4
-3
30,000
150,000
20,000
100,000
10,000
50,000
0
-13 -12 -11 -10
-2
log [NECA], M
0
-9
5
200000
180000
Agonists and antagonists rank order of potency was not
modified by the method used for the analysis of data, as shown
below for the muscarinic M4 receptor agonist experiment (fig6),
and for the Adenosine A2B antagonist experiment. This was also
true for the other receptors analyzed (4 in agonist mode, 3 in
antagonist mode, data not shown).
1,000,000
600,000
400,000
0
-12 -11 -10 -9
10000
-8
-7
-6
-5
-4
-3
Log [ligand], M
AUC
0
20
25
30
0
5
time (s)
10
15
20
25
time (s)
pEC50 Values are Higher When Using
“AUC” rather than “Max-Min”
In order to determine the impact of the data processing method
on the pharmacological parameters, data from agonist and
antagonist aequorin experiments performed on 100 stable
AequoScreen® cell lines co-expressing a GPCR were analyzed
using the “AUC” or the “Max-Min” methods. The agonist assay
affinity values were more affected than the antagonist assay
values. The “AUC” method yielded the highest Z’ values.
pEC50 (mean ± SD)
Z' (mean ± SD)
Window (Highest Agonist Response/Buffer)
Hill Slope
TOP (graphpad fitting) / Digitonin
Highest Agonist Response/Digi
Curves having at least 3 experimental points at
the upper plateau
Antagonist Assays
(7 receptors analyzed)
pIC50 (mean ± SD)
Hill Slope
AUC
AUC
Max-Min
Alloxazine
4.70 ± 0.05
4.64 ± 0.08
MRS 1706
6.61 ± 0.03
6.61 ± 0.03
PSB 1115
5.94 ± 0.08
6.14 ± 0.07
40,000
10,000
0
-12 -11 -10 -9
-6
-5
-4
-3
-2
Log [ligand], M
Max-Min Value
80
Acetylcholine
Oxotremorine
Pilocarpine
Milameline
60
40
20
0
-
-12 -11 -10
-9
-8
-7
-6
-5
-4
-3
-2
25
20
15
10
5
0
-
-12 -11 -10 -9
Log [Ligands] (M)
AUC
-7
Calcium flux assay with the Aequoscreen® method allows
detecting partial agonist activity. We observed that the apparent
efficacy of partial agonists is lower when using the “Max-Min”
method compared to the “AUC” method, as shown below for the
muscarinic M4 receptor. The same observation was made with
the other receptor analyzed (muscarinic M1 receptor, data not
shown).
Max-Min
1.09 ± 0.56 lower value (right-shift) when using
Max-Min compared to AUC
0.59 ± 0.19
0.46 ± 0.25
39
85
1.23
0.73
78%
110%
75%
88%
79%
-8
Partial Agonists’ Apparent Efficacy
Varies with Data Analysis Method
AUC
-8
-7
-6
-5
-4
-3
-2
Log [Ligands] (M)
30%
Max-Min
0.22 ± 0.37 higher value (left-shift) when
using Max-Min compared to AUC
-1.35
-1.24
Acetylcholine
Oxotremorine
Pilocarpine
Milameline
pEC 50
9.07
8.17
5.69
7.00
Efficacy
100%
102%
72%
84%
Hill Slope
1.364
1.069
1.136
1.247
Cells, Cell Culture and AequoScreen® assay:
The 100 stable aequorin cell lines used in this study were from the
PerkinElmer catalog, and were each cultured and used in the aequorin
assay according to their technical datasheet specifications (see
www.PerkinElmer.com). Each receptor was tested in the aequorin
assay in agonist mode with 3 to 8 replicates of each concentration of
the reference agonist. Aequorin assays were performed using
suspension cells, dispensed on the agonists present in the assay
plates, and using either the LumiLux® (PerkinElmer) or the Microlumat
® LB96V (Berthold) readers. Raw kinetics data were exported in excel,
and used to calculate the Area Under the Curve (“AUC” = sum of each
individual timepoint), or to extract the maximal response, and the
minimal response, and calculate their difference (“Max-Min”,
corresponding to the peak height). Fitting of data according to a
sigmoidal dose response model (variable slope) was done using
GraphPad Prism® software version 5.01, and parameters such as
pEC50,pIC50, Hill slope and “TOP” of the curve generated by GraphPad®
software were used to compare the “AUC” and “Max-Min” data analysis
methods. GraphPad® software estimation of the TOP of the curve was
better when using the AUC method, with 98% of curves with a TOP
value being less than 20% bigger than the highest experimental point,
vs 86% of curves for the Max-Min method fitting this criteria.
Fluorescent-dye based adherent assay methodology:
Fluorescent calcium assay was performed with Fluo-4 no wash
(Invitrogen Cat n°F36206) following instructions from the supplier of
the kit. Responses were measured on CellLux® instrument
(PerkinElmer) ( ex. 485nm, em. 535nm).
20,000
Adenosine A2B receptor Aequorin Antagonist Assay, same kinetic data treated
with AUC and Max-Min method
6
Materials and Methods
30,000
30
Kinetics of Fluorescent Calcium and AequoScreen® assays performed on the
same AequoScreen® cell line expressing the muscarinic M5 receptor.
Agonist Assays
(100 receptors analyzed, average values)
pIC50
-2
120000
100000
50,000
Alloxazine
MRS 1706
PSB 1115
800,000
RLU
Max-Min
Max-Min
200,000
20000
140000
4
-4
Rank Order of Potency is not
Changed by the Analysis Method
Max-Min ( RLU)
30000
15
-5
30
Area Under the Curve (AUC)
is used as a measurement of
the Ca2+ signal in the Aequorin assay
220000
10
-6
Impact of analysis method varies according to the receptor
Same data analyzed in AUC or Max-Min mode for 2 receptors
AUC ( RLU)
Max - Min signal
is used as a measurement of
the Ca2+ signal in the Fluo-4 NW assay
5
-7
log [SRIF-28], M
AUC
0
-8
time (s)
time (s)
160000
40,000
200,000
0
-9
pEC50 = 7.15
250,000
AUC ( RLU)
Signal decrease is faster in the aequorin assay compared to the
Fluo-4 assay. Due to slow dissociation of Ca2+ from the
fluorescent dye, kinetics in the fluorescent assay are extended
compared to the real calcium flow. On the contrary, due to
consumption of aequorin substrate during the measurement, the
shape of the signal decrease in the luminescent assay is sharper
than the real calcium decrease. As a result of these features,
the difference between the maximal and minimal signal intensity
is the parameter used in the fluorescent assay, whereas the
integration of the signal (Area under the Curve, AUC) is used in
the luminescent assay. However, past users of the fluo-4 assay
tend to use the Peak Value (“Max-Min”) method in the aequorin
assay as well.
Digi (AUC)
Dig (Max-Min)
SRIF 28 (AUC)
Max-Min (% of Digitonin)
Cells are incubated with coelenterazine,
the co-factor of aequorin. Coelenterazine
enters the cell and combines with
apoaequorin to form active aequorin.
Agonist stimulation of the G-protein
coupled receptor leads to release of Ca2+
from intracellular stores, therefore
increasing cytoplasmic free Ca2+
concentration. This Ca2+ increase leads
to the activation of the catalytic
activity of aequorin, which oxidizes the coelenterazine substrate
to produce CO2 and a flash of light at 466 nm. When working
with GPCRs, aequorin is targeted in mitochondria as it increases
the signal intensity.
pEC50 = 6.55
8
Somatostatin sst5 Aequorin Response
Digitonin (AUC)
Digitonin (Max-Min)
NECA (AUC)
AUC (% of Digitonin)
Principle of
Technology
®
the AequoScreen
Adenosine A2B Aequorin Response
Max-Min ( RLU)
2
Kinetics of the signal in luminescence
and fluorescence assays
Max-Min ( RLU)
Aequorin-based Ca2+ assays represent a new
paradigm in cell-based high throughput screening
(HTS) Ca2+-coupled GPCRs and ion channels. Besides
not being affected by autofluorescent compounds,
the AequoScreen® methodology also presents
advantages in terms of improved quality (Z’, window
and hit confirmation rate), easiness of assay
implementation, and increased throughput.
Although the traditional fluorescent dye-based Ca2+
assays fundamentally measure the same intracellular
event as aequorin, when looking at the kinetics of
the response to agonist stimulation, signal decrease
is faster in the aequorin assay compared to the Fluo4 assay. On one hand, due to slow dissociation of
Ca2+ from the fluorescent dye, kinetics in the
fluorescent assay are extended compared to the real
calcium flow. On the other hand, due to consumption
of aequorin substrate during the measurement, the
shape of the signal decrease in the luminescent
assay is sharper than the true calcium decrease. As a
result of these technical features, the difference
between the maximal and minimal signal intensity
(Max-Min or Peak value) is the preferred parameter
used in the fluorescent assay, whereas the
integration of the signal (Area Under the Curve,
AUC) is more adequate in the luminescent assay.
However, because of the historical spread of the
fluorescent calcium assays, there is a natural
tendency to use the “Max-Min” method to analyze
Aequorin data.
In this work, we present how pharmacological
parameters extracted from aequorin experimental
results, such as EC50, IC50 values and agonist
efficacy, can differ according to the data processing
method, i.e. “Max-Min” method vs “AUC” method.
3
RFU
Introduction
RLU
1
Acetylcholine
Oxotremorine
Pilocarpine
Milameline
pEC50
8.53
7.44
5.16
6.67
Efficacy
100%
108%
52%
78%
Hill Slope
1.351
0.948
1.042
1.619
Muscarinic M4 receptor Aequorin experiment, same kinetic data treated with
AUC and Max-Min method
9
Conclusions
 Calcium flux using AequoScreen® assays is a
convenient and cost-effective method to perform HTS
as well as pharmacological characterization of
compounds.
 The method used to transform the kinetic
luminescence data into sigmoidal dose-response
curves impacts the apparent pharmacological
parameters that can be extracted from these curves,
such as EC50, IC50 and efficacy values. In addition, the
experimental settings (like dispensing speed,
dispensing height, geometry of the reader used,...)
can also modify the kinetics (not shown) and hence
the values of the parameters that can be extracted
from them.
 The method we recommend to analyze aequorin
data is using the Area Under the Curve (AUC) of the
luminescent signal, due to the better sensitivity
(higher pEC50 values), the better assay quality (Z’
value), and the better curve fitting when using this
method.
 As for any assay, the true meaning of parameters
such as EC50, IC50 and efficacy extracted from aequorin
experiments results must be kept in mind when using
them and when comparing them with results from
other assays.
PerkinElmer, Inc., 940 Winter Street, Waltham, MA USA (800) 762-4000 or (+1) 203 925-4602 www.perkinelmer.com