Journal of Pharmacological and Toxicological Methods 71 (2015) 42–45
Contents lists available at ScienceDirect
Journal of Pharmacological and Toxicological Methods journal homepage: www.elsevier.com/locate/jpharmtox
Brief communication
Increasing the flexibility of the LANCE cAMP detection kit Morag Rose Hunter, Michelle Glass ⁎ Department of Pharmacology & Clinical Pharmacology, University of Auckland, Private Bag 92019, Auckland, New Zealand
a r t i c l e
i n f o
Article history: Received 18 August 2014 Accepted 26 October 2014 Available online 3 December 2014 Keywords: cAMP Adenylyl cyclase High throughput screening Cell signalling assays G-protein coupled receptors
a b s t r a c t Introduction: The detection of cAMP signalling is a common endpoint in the study of G-protein coupled receptors. A number of commercially available kits enable easy detection of cAMP. These kits are based on competition for a cAMP binding site on an antibody or cAMP binding protein and as such have a limited dynamic range. Here, we describe the optimisation of the commercially-available LANCE cAMP detection kit (PerkinElmer) to enable detection in cell lysates. This kit has been designed for use with live cells, with detection reagents applied to cells without wash steps. The standard protocol therefore requires that all assay reagents are compatible with the antibody and the final fluorescent detection stage, limiting the range of assay media and test compounds that can be utilised. The entire experiment must be repeated if cAMP levels fall outside the limited dynamic range. Here we describe a modified protocol that enables the assay to be performed on cell lysates, thereby overcoming these limitations. Methods: In this modified protocol, cells are stimulated for a cAMP response in standard media/buffers, washed and then lysed. The cell lysate is then assayed using a modified protocol for the LANCE cAMP detection kit. Samples were tested for stability following a freeze–thaw cycle. Results: The modified LANCE cAMP detection protocol gives a reproducible measurement of cAMP in cell lysate. Lysate samples remain stable when stored at −80 °C. Discussion: Separating the stimulation and detection phases of this cAMP assay allows a vast array of cell stimulation conditions to be tested. The lysate-modified protocol for the LANCE cAMP detection kit therefore increases the flexibility, versatility and convenience of the assay. As samples are insensitive to freeze–thaw, it enables retesting of samples under different dilution conditions to ensure that all samples remain within the dynamic range of the standard curve. © 2014 Elsevier Inc. All rights reserved.
1. Introduction The detection of drug-mediated alterations in cellular cAMP concentrations is a common method of drug discovery. Intracellular cAMP levels are tightly regulated by adenylyl cyclase enzymes that convert adenosine triphosphate to cAMP, as well as by phosphodiesterases that catalyse cAMP degradation. Isoforms of adenylyl cyclase are activated or inhibited via direct interaction with G-protein subunits (stimulated by Gs, inhibited by Gi/o, and therefore under the control of G-protein coupled receptors (GPCRs)), or by changes in intracellular Ca2+ and calmodulin. cAMP in turn regulates a range of cAMP-dependent protein kinases, resulting in the phosphorylation of protein targets (see Hofer, 2012 for a recent review). The last decade has seen a surge in new methodologies for the detection of intracellular cAMP (see Hill, Williams, & May, 2010 for review), with an increasing emphasis on methods adaptable for high throughput screening for drug discovery. Many of these methods are now available
Abbreviations: cAMP, cyclic adenosine monophosphate; GPCR, G-proteincoupledreceptor; TR-FRET, time-resolved fluorescence resonance energy transfer. ⁎ Corresponding author. Tel.:+64 9 9236247. E-mail addresses: [email protected] (M.R. Hunter), [email protected] (M. Glass).
http://dx.doi.org/10.1016/j.vascn.2014.10.008 1056-8719/© 2014 Elsevier Inc. All rights reserved.
in the form of convenient “kits”. One such kit is the LANCE cAMP kit (PerkinElmer). The LANCE cAMP assay is a time-resolved fluorescence resonance energy transfer (TR-FRET) immunoassay, in which biotinylated-cAMP and europium-W8044 chelate-labelled streptavidin compete for cAMP binding sites on an Alexa-Fluor 647-labelled antibody. A light pulse at 340 nm then excites the europium-chelate molecule of the tracer, leading to energy transfer to the Alexa fluorophore, which in turn emits light at 665 nm. Thus, the greater the level of cAMP in the test sample, the higher the level of competition and the lower the signal emitted. Using the manufacturer's protocol for the LANCE cAMP kit, detection reagents are applied to live cells suspended in a minimal assay buffer, and drug stimulation is performed at room temperature for an extended period of time. The fluorescent readout is then converted to a concentration of cAMP per assay point, utilising a cAMP standard curve constructed from pure cAMP and the same detection reagents. As for all such assays, the validity of the data is dependent on all of the samples falling within the linear range of the standard curve. In the case of the LANCE assay, the manufacturers report a high-affinity interaction between cAMP and the antibody, with IC50s in the low nanomolar range (PerkinElmer Life & Analytical Sciences, 2014). The standard curve generates a sigmoidal response curve with a robust span, but with a relatively limited dynamic range. For example, in the
M.R. Hunter, M. Glass / Journal of Pharmacological and Toxicological Methods 71 (2015) 42–45
43
data presented for standard curve measurement on a Victor (Fig. 2, page 17, (PerkinElmer life & Analytical Sciences, 2014)), the assay has a span of approximately 16,000 counts and an IC50 of 2.4 nM, but the linear range of the standard curve (between 10 and 90%) occurs between approximately 0.3 nM and 25 nM. From our own laboratory's experience with this assay, we have found that despite substantial optimisation of conditions including cell number and duration of incubation, data frequently fell on the non-linear portion of the curve, thereby rendering the results meaningless. As the method is an “add and read” approach (which exposes the total cAMP in the well to the antibody) there is no scope within this assay design to reprobe samples which fall outside the linear range. Furthermore, the assay requires cAMP to be generated in an HBSS buffer, on cells in suspension and at room temperature. These are markedly different conditions to that used in most receptor signalling assays, therefore reducing the comparability of results generated across a range of assays. The ability to compare results across multiple assays is an increasingly desirable quality, given the strong interest in functional selectivity between agonist driven pathways for drug discovery (Gesty-Palmer & Luttrell, 2011). We have therefore modified the LANCE protocol to enable it to be utilised on cell lysates. This approach has several advantages. Firstly, as only a small proportion of the lysate generated per well is assayed, samples falling outside of the range of the standard curve can be reassayed at an appropriate dilution to ensure that they fall within the linear range. Furthermore, the assay can be carried out on adherent cells, and is not sensitive to the conditions under which the cells are stimulated, providing the flexibility to match conditions (stimulation media, cell confluency, time and temperature) to other in-house signalling assays. The assay as described also allows cell lysates to be frozen prior to detection, allowing samples from multiple days to be detected simultaneously without loss of cAMP detection.
2013). Cells were grown in standard growth media consisting of DMEM supplemented with 10% FBS, as well as 250 ug/ml zeocin as a selection antibiotic, at 37 °C in 5% CO2. Assay conditions were based on those previously utilised for assaying CB1 receptors by radioimmunoassay (Kearn, Blake-Palmer, Daniel, Mackie, & Glass, 2005). Precise plating and stimulation conditions should be optimised for each cell type. A summary of optimisation considerations is detailed in Table 1. In this case, HEK-hCB1 cells were seeded in poly-L-lysinetreated 96-well plates to achieve 70–80% confluence after 24 h (approximately 60,000 cells/well). Growth media was removed from HEK-hCB1 cells and replaced with DMEM containing 5 mg/ml BSA and 0.5 mM IBMX, and the cells were subsequently incubated at 37 °C in 5% CO2 for 30 min immediately before drug stimulation to allow time for a stable base line to develop in the absence of full serum. BSA was included as a carrier molecule due to the lipophilic nature of cannabinoid ligands, and could be omitted for more hydrophyllic ligands. Drugs were diluted in DMEM containing 5 mg/ml BSA and 0.5 mM IBMX. As CB1 is a Gi-linked receptor, cAMP synthesis was stimulated with forskolin to enable the inhibition of this response to be detected. Drugs were diluted to 2× the desired final concentration and added directly to the cells to minimise cell disruption. Cells were stimulated for 20 min at 37 °C, and plates were placed on ice immediately thereafter to prevent further signalling. The stimulation medium was quickly removed from all wells and the cells were immediately lysed with icecold detection buffer (provided in the kit). This step should be optimised for each cell type/receptor assayed to ensure that the level of cAMP is within the dynamic range of the standard curve, with a suggested starting point of 25–50 μl per well in a 96-well format. Once the detection buffer is applied, the cells should be gently agitated at 4 °C for 15–30 min to ensure full cell lysis.
2. Materials and methods
2.3. Detection of cAMP
2.1. Reagents
The standard curve dilutions were prepared as described in the LANCE protocol. Briefly, at least six standards with final concentrations of between 1 μM and 10 pM of cAMP were prepared in detection buffer from the 50 μM cAMP standard supplied in the LANCE kit. In order to utilise the entire antibody component of the kit using the lysate technique, additional detection buffer is required. This can be made inhouse, as described in the “Reagents” section above. The proportion of sample to antibody was based on the suggested preparation of the standard curve. Thus, 6 μl of lysate sample or cAMP standard was transferred to a half-volume white 96-well plate (Perkin Elmer). Anti-cAMP antibody, diluted 1/100 in detection buffer (as per kit instructions), was then added immediately to each well (6 μl of diluted anti-cAMP antibody/well). The lysate samples and antibody were mixed in the plate and incubated, protected from light, for 30 min prior to the addition of the detection mix (gentle rocking or agitation is optional). The detection mix (europium-streptavidin and biotincAMP) was prepared as per LANCE kit instructions (first, intermediate
Cells were grown and assayed in culture reagents (DMEM, FBS, BSA) supplied by Life Technologies (USA) or Sigma (Australia), on Corning plasticware (Corning, USA). Drugs (isobutylmethylxanthine (IMBX), forskolin, CP55,940) were obtained from Tocris Bioscience (UK) and Sigma Aldrich (Australia). The LANCE cAMP kit and half-area white 96-well plates were purchased from Perkin Elmer (USA). Additional “detection buffer” was made as described in LANCE cAMP kit handbook (50 mM HEPES, 10 mM CaCl2, 0.35% Triton X-100, distilled water and pH adjusted to 7.4). 2.2. Cell stimulation and lysate preparation The method was developed utilising HEK 293 cells stably transfected with the human CB1 receptor, N-terminally tagged with 3hemagglutinin as previously described (HEK-hCB1) (Cawston et al.,
Table 1 Optimisation considerations for assay conditions. Condition
Optimisation considerations
Assay vessel Cell adherence conditions Confluency
Well/plate size, volume of growth media Surface coating of assay vessel, time to adhere Balance between cAMP concentration in lysate, and effect of cell contact inhibition on receptor function Serum starve length, assay buffer/media composition
Media Drug stimulation conditions
Lysis buffer
Presence and concentration of IBMX, presence and concentration of forskolin, length of drug incubation, temperature. Volume of lysis buffer
General notes Dependent on cell line Higher cell density gives higher cAMP concentrations in a fixed volume of detection buffer. Dependent on cell line/receptor requirements. Generally a serum-free period of 30 min to 1 h results in low base line cAMP. These conditions should be determined by the purpose of the experiment. All other detection conditions can be optimised around these factors. Lower volume of lysis buffer gives a higher concentration of cAMP in lysate. Lysis buffer should cover the entire surface of the assay plate to ensure complete lysis.
M.R. Hunter, M. Glass / Journal of Pharmacological and Toxicological Methods 71 (2015) 42–45
A 40000
cAMP standard curve
cell lysate samples
35000 30000 25000 20000
3.1. Optimisation of cell lysis volume HEK-hCB1 cells were incubated in the presence and absence of 1 μM forskolin. Following removal of media, the cells were lysed in either 30 or 50 μl of detection buffer and assayed as described above. As this is a competition assay, increasing cAMP levels result in a decreased TRFRET signal. The cAMP standard curve data was graphed as a sigmoidal curve, using GraphPad Prism 6 software, and unknowns were automatically interpolated from the standard curve. As shown in Fig. 1A, a typical standard curve spanned approximately 17,000 fluorescent units, with an IC50 of 1.2 nM. The linear portion of the curve (between 10 and 90%) therefore ranged from 0.13 nM to 10.8 nM of cAMP per sample. The cell lysate samples were interpolated from the standard curve and adjusted to account for the lysate volume; absolute cAMP per well is shown in Fig. 1B. As can be seen in Fig. 1A, with a lysate volume of 50 μl, the basal concentration of cAMP falls outside of the linear range of the curve; however Prism did return interpolated values for 2 of the 3 replicates (Prism will not interpolate if the value falls above the “top” or below the “bottom” of the curve, although this does not reflect the actual dynamic range of most assays). As can be seen from the table, while the stimulated levels of cAMP are both within the range of the standard curve and comparable when estimated by either method, the basal value for 50 μl lysates is considerably underestimated by this method, demonstrating the importance of ensuring that samples fall within the linear range. As lysis in 30 μl of detection buffer produced levels of cAMP within the linear range of the curve, this was used for further experiments. 3.2. Determination of receptor-mediated cAMP signalling To ensure that this detection method maintained sufficient sensitivity to be useful in typical cAMP signalling assays, a concentration response curve for a cannabinoid receptor agonist was carried out. Cells were treated with 1 μM forskolin and varying concentrations of the CB1 agonist CP55,940, and the cell lysates were then assayed using the lysate-modified LANCE cAMP assay. As demonstrated in Fig. 2, a concentration-dependent inhibition of forskolin-stimulated cAMP accumulation was detected with pEC50 for CP55,940 of 9.48 ± 0.09 (mean ± standard error of the mean, n = 3), similar to the results recently reported for these same cells utilising a cAMP biosensor (pEC50 9.4 ± 0.1, (Cawston et al., 2013)).
30µl lysate 50µl lysate
,5
0u l
0u l al ,5
0u l
Basal (fmol of Forskolin cAMP per well) stimulated (fmol of cAMP per well) 9.453±3.516 230.22±40.98 2.7975±0.7825 209.2±23.155
FS K
B
3. Results
B
Apparent fold stimulation
24.35 74.78
Fig. 1. Representative data showing the effect of altering lysis buffer volume on cAMP levels in cell lysates. Cells were plated at the same confluency, stimulated and then lysed. A) The graph shows the standard curve, as well as raw data counts for cells stimulated with 1 μM forskolin and then lysed in either 30 μl or 50 μl of lysis buffer. B) Comparison of the apparent effect of forskolin, using the cAMP concentrations as interpolated from the standard curve. cAMP values are presented as fmol of cAMP per well, as calculated by adjusting for the lysis volume and dilution in the detection mix. All data is shown as mean ± SEM.
3.3. Determination of lysate stability through freeze–thaw In order to determine the stability of cAMP in the lysis buffer, cells were treated with forskolin (1 μM) as above, and the resulting lysate was probed twice: first immediately after cell lysis and again after 7 days storage at − 80 °C. Paired T-test analysis showed no statistical difference between cAMP concentration measured before or after freezing (p N 0.05, n = 3).
4. Discussion Here we have presented a simple modification of the LANCE cAMP detection kit, which adds increased flexibility and versatility to the kit's standard applications. This method modification process has also 150
fmol cAMP in well
One potential advantage of a lysate-based detection method is that it enables the samples to be reprobed if the initially-detected value falls outside of the standard curve. For this to be practical, lysates need to be stable following freeze–thaw. To test this, cells were stimulated as described above with forskolin, and some lysate immediately assayed for cAMP. The remaining lysate was frozen at − 80 °C for 7 days, warmed to room temperature (with gentle agitation) and re-assayed.
as
-6
,3
-7
B
-8
FS K
-11 -10 -9
log [cAMP], M
al ,3
0
2.4. Stability of lysate during freeze–thaw
0u l
15000
as
dilutions of europium-streptavidin at 1/18, and biotin-cAMP at 1/6; then each intermediate dilution is further diluted 1/125 in detection buffer to give the detection mix), and complexes were allowed to form for at least 15 min at room temperature. When the detection mix had incubated for at least 15 min, and the lysate-antibody samples had been incubated for 30 min, 12 μl of the detection mix was added per well. The plate was then incubated for a further 60 min (or up to overnight) at room temperature (with optional gentle agitation), before detection in a TR-FRET-capable plate reader as per the manufacturer's instructions. In these experiments, we have utilised a Wallac Victor 1420 Multilabel Counter (Perkin Elmer), with excitation at 340 nm and emission at 615 nm and 665 nm.
counts
44
100
50
0 Basal
FSK
-11
-10
-9
-8
-7
-6
log [CP55,940], M Fig. 2. Inhibition of forskolin-stimulated cAMP accumulation in HEK-HA-hCB1 cells, utilising the modified LANCE assay. Representative data from one of three separate experiments, with each drug condition assayed in triplicate and plotted as mean ± SEM.
M.R. Hunter, M. Glass / Journal of Pharmacological and Toxicological Methods 71 (2015) 42–45
been applied in-house to the AlphaScreen cAMP assay kit (Perkin Elmer) (unpublished data), and can presumably be applied to other commercially-available “add and read” detection systems in order to increase reliability and/or cost-effectiveness. All of the available competition assays require conversion to absolute cAMP levels through the use of a standard curve for accurate data analysis (see Hill et al., 2010 for a discussion on the potential for data misinterpretation from direct detection of cAMP without reference to a standard curve). As such, these assays are highly reliant on data falling within the dynamic range of the standard curve. Here, we have shown that a relatively low concentration of forskolin (1 μM) utilises over 50% of the available dynamic range (Fig. 1), and under our standardised assay conditions all forskolin concentrations above 3 μM resulted in raw data which was outside the dynamic range of the assay (data not shown). However, it is common practise to use concentrations of forskolin well above 5 μM for HEK 293 cells, for example (Bitterman, Ramos-Espiritu, Diaz, Levin, & Buck, 2013; Cawston et al., 2013; Wang, Yan, Zheng, He, & Yang, 2014; Wehbi et al., 2013). Therefore, assay design should consider the expected quantity of cAMP generated. Studies activating Gs-linked GPCRs may require the use of different dilution factors for basal and stimulated conditions to gain accurate measures of their actual cAMP levels—something which could not be achieved in the standard protocol of an “add and read”. Our method modification therefore overcomes what is probably the greatest limitation of the LANCE cAMP kit (and other similar kits), in that it allows samples to be diluted and reprobed if they fall outside of the standard curve, without the need to repeat the experiment. Importantly, we have shown here that samples can be stored frozen without loss of detectable signal. This would also enable samples prepared on separate days to be analysed together for convenience and efficiency, a measure which would also conserve reagents because a single standard curve could be utilised for multiple experiments. The ability to reprobe samples also means that this method requires considerably less optimisation than is required for the standard application of this kit, which requires multiple experiments to establish the conditions that produce signals within the dynamic range. Additionally, the small volumes utilised in the detection (24 μl total) are suitable for detection in 384 well plates, allowing for higher throughput or automation. In this example, we have utilised HEK293 cells transfected with the cannabinoid CB1 receptor, but the lysate-modified LANCE cAMP detection method is adaptable for any adherent cell type, with minimal optimisation from existing protocols set up in any laboratory. This technique also permits the use of any cell culture medium during cell stimulation, because media is removed and washed off; therefore the detection method is not affected by media pigments such as phenol red. Indeed, completing
45
cell stimulation before the detection steps allows researchers to test a wide range of experimental paradigms (e.g. salt concentrations) without concern as to whether these will compromise the subsequent detection assay. Increasingly, pharmacological studies have focused on the “functional selectivity” of ligands—this being the relative effect of ligands on different signalling pathways mediated by the same receptor (Kenakin, 2011). Such comparisons are substantially more valid if cell stimulation protocols are kept as similar as possible across different assays, which this adaptation makes feasible, as the choice of stimulation buffer and temperature will not alter the ability to detect signal. In summary, we show here a straightforward adaptation of the manufacturer's recommendation for a commonly used cell signalling assay, which would improve the utility of the cAMP detection kit. Acknowledgements This research was supported by a Royal Society of New Zealand Marsden (#11-UOA-201) Fund Grant to MG. References Bitterman, J. L., Ramos-Espiritu, L., Diaz, A., Levin, L. R., & Buck, J. (2013). Pharmacological distinction between soluble and transmembrane adenylyl cyclases. Journal of Pharmacology and Experimental Therapeutics, 347, 589–598. Cawston, E. E., Redmond, W. J., Breen, C. M., Grimsey, N. L., Connor, M., & Glass, M. (2013). Real-time characterization of cannabinoid receptor 1 (CB1) allosteric modulators reveals novel mechanism of action. British Journal of Pharmacology, 170, 893–907. Gesty-Palmer, D., & Luttrell, L. M. (2011). Refining efficacy: Exploiting functional selectivity for drug discovery. Advances in Pharmacology, 62, 79–107. Hill, S. J., Williams, C., & May, L. T. (2010). Insights into GPCR pharmacology from the measurement of changes in intracellular cyclic AMP; advantages and pitfalls of differing methodologies. British Journal of Pharmacology, 161, 1266–1275. Hofer, A. M. (2012). Interactions between calcium and cAMP signaling. Current Medicinal Chemistry, 19, 5768–5773. Kearn, C. S., Blake-Palmer, K., Daniel, E., Mackie, K., & Glass, M. (2005). Concurrent stimulation of cannabinoid CB1 and dopamine D2 receptors enhances heterodimer formation: A mechanism for receptor cross-talk? Molecular Pharmacology, 67, 1697–1704. Kenakin, T. (2011). Functional selectivity and biased receptor signaling. Journal of Pharmacology and Experimental Therapeutics, 336, 296–302. PerkinElmer Life and Analytical Sciences, I. (2014). LANCE cAMP 384 Kit Manual. In http://www.perkinelmer.com/CMSResources/Images/44-73586MAN_ LANCEcAMP384KitUser.pdf (Ed.), (Vol. 2014). Shelton, CT, USA: PerkinElmer Life and Analytical Sciences, Inc. Wang, Y., Yan, M., Zheng, G. -y, He, L., & Yang, H. (2014). A cell-based, high-throughput homogeneous time-resolved fluorescence assay for the screening of potential [kappa]-opioid receptor agonists. Acta Pharmacologica Sinica, 35, 957–966. Wehbi, V. L., Stevenson, H. P., Feinstein, T. N., Calero, G., Romero, G., & Vilardaga, J. -P. (2013). Noncanonical GPCR signaling arising from a PTH receptor–arrestin–Gβγ complex. Proceedings of the National Academy of Sciences, 110, 1530–1535.
Contents lists available at ScienceDirect
Journal of Pharmacological and Toxicological Methods journal homepage: www.elsevier.com/locate/jpharmtox
Brief communication
Increasing the flexibility of the LANCE cAMP detection kit Morag Rose Hunter, Michelle Glass ⁎ Department of Pharmacology & Clinical Pharmacology, University of Auckland, Private Bag 92019, Auckland, New Zealand
a r t i c l e
i n f o
Article history: Received 18 August 2014 Accepted 26 October 2014 Available online 3 December 2014 Keywords: cAMP Adenylyl cyclase High throughput screening Cell signalling assays G-protein coupled receptors
a b s t r a c t Introduction: The detection of cAMP signalling is a common endpoint in the study of G-protein coupled receptors. A number of commercially available kits enable easy detection of cAMP. These kits are based on competition for a cAMP binding site on an antibody or cAMP binding protein and as such have a limited dynamic range. Here, we describe the optimisation of the commercially-available LANCE cAMP detection kit (PerkinElmer) to enable detection in cell lysates. This kit has been designed for use with live cells, with detection reagents applied to cells without wash steps. The standard protocol therefore requires that all assay reagents are compatible with the antibody and the final fluorescent detection stage, limiting the range of assay media and test compounds that can be utilised. The entire experiment must be repeated if cAMP levels fall outside the limited dynamic range. Here we describe a modified protocol that enables the assay to be performed on cell lysates, thereby overcoming these limitations. Methods: In this modified protocol, cells are stimulated for a cAMP response in standard media/buffers, washed and then lysed. The cell lysate is then assayed using a modified protocol for the LANCE cAMP detection kit. Samples were tested for stability following a freeze–thaw cycle. Results: The modified LANCE cAMP detection protocol gives a reproducible measurement of cAMP in cell lysate. Lysate samples remain stable when stored at −80 °C. Discussion: Separating the stimulation and detection phases of this cAMP assay allows a vast array of cell stimulation conditions to be tested. The lysate-modified protocol for the LANCE cAMP detection kit therefore increases the flexibility, versatility and convenience of the assay. As samples are insensitive to freeze–thaw, it enables retesting of samples under different dilution conditions to ensure that all samples remain within the dynamic range of the standard curve. © 2014 Elsevier Inc. All rights reserved.
1. Introduction The detection of drug-mediated alterations in cellular cAMP concentrations is a common method of drug discovery. Intracellular cAMP levels are tightly regulated by adenylyl cyclase enzymes that convert adenosine triphosphate to cAMP, as well as by phosphodiesterases that catalyse cAMP degradation. Isoforms of adenylyl cyclase are activated or inhibited via direct interaction with G-protein subunits (stimulated by Gs, inhibited by Gi/o, and therefore under the control of G-protein coupled receptors (GPCRs)), or by changes in intracellular Ca2+ and calmodulin. cAMP in turn regulates a range of cAMP-dependent protein kinases, resulting in the phosphorylation of protein targets (see Hofer, 2012 for a recent review). The last decade has seen a surge in new methodologies for the detection of intracellular cAMP (see Hill, Williams, & May, 2010 for review), with an increasing emphasis on methods adaptable for high throughput screening for drug discovery. Many of these methods are now available
Abbreviations: cAMP, cyclic adenosine monophosphate; GPCR, G-proteincoupledreceptor; TR-FRET, time-resolved fluorescence resonance energy transfer. ⁎ Corresponding author. Tel.:+64 9 9236247. E-mail addresses: [email protected] (M.R. Hunter), [email protected] (M. Glass).
http://dx.doi.org/10.1016/j.vascn.2014.10.008 1056-8719/© 2014 Elsevier Inc. All rights reserved.
in the form of convenient “kits”. One such kit is the LANCE cAMP kit (PerkinElmer). The LANCE cAMP assay is a time-resolved fluorescence resonance energy transfer (TR-FRET) immunoassay, in which biotinylated-cAMP and europium-W8044 chelate-labelled streptavidin compete for cAMP binding sites on an Alexa-Fluor 647-labelled antibody. A light pulse at 340 nm then excites the europium-chelate molecule of the tracer, leading to energy transfer to the Alexa fluorophore, which in turn emits light at 665 nm. Thus, the greater the level of cAMP in the test sample, the higher the level of competition and the lower the signal emitted. Using the manufacturer's protocol for the LANCE cAMP kit, detection reagents are applied to live cells suspended in a minimal assay buffer, and drug stimulation is performed at room temperature for an extended period of time. The fluorescent readout is then converted to a concentration of cAMP per assay point, utilising a cAMP standard curve constructed from pure cAMP and the same detection reagents. As for all such assays, the validity of the data is dependent on all of the samples falling within the linear range of the standard curve. In the case of the LANCE assay, the manufacturers report a high-affinity interaction between cAMP and the antibody, with IC50s in the low nanomolar range (PerkinElmer Life & Analytical Sciences, 2014). The standard curve generates a sigmoidal response curve with a robust span, but with a relatively limited dynamic range. For example, in the
M.R. Hunter, M. Glass / Journal of Pharmacological and Toxicological Methods 71 (2015) 42–45
43
data presented for standard curve measurement on a Victor (Fig. 2, page 17, (PerkinElmer life & Analytical Sciences, 2014)), the assay has a span of approximately 16,000 counts and an IC50 of 2.4 nM, but the linear range of the standard curve (between 10 and 90%) occurs between approximately 0.3 nM and 25 nM. From our own laboratory's experience with this assay, we have found that despite substantial optimisation of conditions including cell number and duration of incubation, data frequently fell on the non-linear portion of the curve, thereby rendering the results meaningless. As the method is an “add and read” approach (which exposes the total cAMP in the well to the antibody) there is no scope within this assay design to reprobe samples which fall outside the linear range. Furthermore, the assay requires cAMP to be generated in an HBSS buffer, on cells in suspension and at room temperature. These are markedly different conditions to that used in most receptor signalling assays, therefore reducing the comparability of results generated across a range of assays. The ability to compare results across multiple assays is an increasingly desirable quality, given the strong interest in functional selectivity between agonist driven pathways for drug discovery (Gesty-Palmer & Luttrell, 2011). We have therefore modified the LANCE protocol to enable it to be utilised on cell lysates. This approach has several advantages. Firstly, as only a small proportion of the lysate generated per well is assayed, samples falling outside of the range of the standard curve can be reassayed at an appropriate dilution to ensure that they fall within the linear range. Furthermore, the assay can be carried out on adherent cells, and is not sensitive to the conditions under which the cells are stimulated, providing the flexibility to match conditions (stimulation media, cell confluency, time and temperature) to other in-house signalling assays. The assay as described also allows cell lysates to be frozen prior to detection, allowing samples from multiple days to be detected simultaneously without loss of cAMP detection.
2013). Cells were grown in standard growth media consisting of DMEM supplemented with 10% FBS, as well as 250 ug/ml zeocin as a selection antibiotic, at 37 °C in 5% CO2. Assay conditions were based on those previously utilised for assaying CB1 receptors by radioimmunoassay (Kearn, Blake-Palmer, Daniel, Mackie, & Glass, 2005). Precise plating and stimulation conditions should be optimised for each cell type. A summary of optimisation considerations is detailed in Table 1. In this case, HEK-hCB1 cells were seeded in poly-L-lysinetreated 96-well plates to achieve 70–80% confluence after 24 h (approximately 60,000 cells/well). Growth media was removed from HEK-hCB1 cells and replaced with DMEM containing 5 mg/ml BSA and 0.5 mM IBMX, and the cells were subsequently incubated at 37 °C in 5% CO2 for 30 min immediately before drug stimulation to allow time for a stable base line to develop in the absence of full serum. BSA was included as a carrier molecule due to the lipophilic nature of cannabinoid ligands, and could be omitted for more hydrophyllic ligands. Drugs were diluted in DMEM containing 5 mg/ml BSA and 0.5 mM IBMX. As CB1 is a Gi-linked receptor, cAMP synthesis was stimulated with forskolin to enable the inhibition of this response to be detected. Drugs were diluted to 2× the desired final concentration and added directly to the cells to minimise cell disruption. Cells were stimulated for 20 min at 37 °C, and plates were placed on ice immediately thereafter to prevent further signalling. The stimulation medium was quickly removed from all wells and the cells were immediately lysed with icecold detection buffer (provided in the kit). This step should be optimised for each cell type/receptor assayed to ensure that the level of cAMP is within the dynamic range of the standard curve, with a suggested starting point of 25–50 μl per well in a 96-well format. Once the detection buffer is applied, the cells should be gently agitated at 4 °C for 15–30 min to ensure full cell lysis.
2. Materials and methods
2.3. Detection of cAMP
2.1. Reagents
The standard curve dilutions were prepared as described in the LANCE protocol. Briefly, at least six standards with final concentrations of between 1 μM and 10 pM of cAMP were prepared in detection buffer from the 50 μM cAMP standard supplied in the LANCE kit. In order to utilise the entire antibody component of the kit using the lysate technique, additional detection buffer is required. This can be made inhouse, as described in the “Reagents” section above. The proportion of sample to antibody was based on the suggested preparation of the standard curve. Thus, 6 μl of lysate sample or cAMP standard was transferred to a half-volume white 96-well plate (Perkin Elmer). Anti-cAMP antibody, diluted 1/100 in detection buffer (as per kit instructions), was then added immediately to each well (6 μl of diluted anti-cAMP antibody/well). The lysate samples and antibody were mixed in the plate and incubated, protected from light, for 30 min prior to the addition of the detection mix (gentle rocking or agitation is optional). The detection mix (europium-streptavidin and biotincAMP) was prepared as per LANCE kit instructions (first, intermediate
Cells were grown and assayed in culture reagents (DMEM, FBS, BSA) supplied by Life Technologies (USA) or Sigma (Australia), on Corning plasticware (Corning, USA). Drugs (isobutylmethylxanthine (IMBX), forskolin, CP55,940) were obtained from Tocris Bioscience (UK) and Sigma Aldrich (Australia). The LANCE cAMP kit and half-area white 96-well plates were purchased from Perkin Elmer (USA). Additional “detection buffer” was made as described in LANCE cAMP kit handbook (50 mM HEPES, 10 mM CaCl2, 0.35% Triton X-100, distilled water and pH adjusted to 7.4). 2.2. Cell stimulation and lysate preparation The method was developed utilising HEK 293 cells stably transfected with the human CB1 receptor, N-terminally tagged with 3hemagglutinin as previously described (HEK-hCB1) (Cawston et al.,
Table 1 Optimisation considerations for assay conditions. Condition
Optimisation considerations
Assay vessel Cell adherence conditions Confluency
Well/plate size, volume of growth media Surface coating of assay vessel, time to adhere Balance between cAMP concentration in lysate, and effect of cell contact inhibition on receptor function Serum starve length, assay buffer/media composition
Media Drug stimulation conditions
Lysis buffer
Presence and concentration of IBMX, presence and concentration of forskolin, length of drug incubation, temperature. Volume of lysis buffer
General notes Dependent on cell line Higher cell density gives higher cAMP concentrations in a fixed volume of detection buffer. Dependent on cell line/receptor requirements. Generally a serum-free period of 30 min to 1 h results in low base line cAMP. These conditions should be determined by the purpose of the experiment. All other detection conditions can be optimised around these factors. Lower volume of lysis buffer gives a higher concentration of cAMP in lysate. Lysis buffer should cover the entire surface of the assay plate to ensure complete lysis.
M.R. Hunter, M. Glass / Journal of Pharmacological and Toxicological Methods 71 (2015) 42–45
A 40000
cAMP standard curve
cell lysate samples
35000 30000 25000 20000
3.1. Optimisation of cell lysis volume HEK-hCB1 cells were incubated in the presence and absence of 1 μM forskolin. Following removal of media, the cells were lysed in either 30 or 50 μl of detection buffer and assayed as described above. As this is a competition assay, increasing cAMP levels result in a decreased TRFRET signal. The cAMP standard curve data was graphed as a sigmoidal curve, using GraphPad Prism 6 software, and unknowns were automatically interpolated from the standard curve. As shown in Fig. 1A, a typical standard curve spanned approximately 17,000 fluorescent units, with an IC50 of 1.2 nM. The linear portion of the curve (between 10 and 90%) therefore ranged from 0.13 nM to 10.8 nM of cAMP per sample. The cell lysate samples were interpolated from the standard curve and adjusted to account for the lysate volume; absolute cAMP per well is shown in Fig. 1B. As can be seen in Fig. 1A, with a lysate volume of 50 μl, the basal concentration of cAMP falls outside of the linear range of the curve; however Prism did return interpolated values for 2 of the 3 replicates (Prism will not interpolate if the value falls above the “top” or below the “bottom” of the curve, although this does not reflect the actual dynamic range of most assays). As can be seen from the table, while the stimulated levels of cAMP are both within the range of the standard curve and comparable when estimated by either method, the basal value for 50 μl lysates is considerably underestimated by this method, demonstrating the importance of ensuring that samples fall within the linear range. As lysis in 30 μl of detection buffer produced levels of cAMP within the linear range of the curve, this was used for further experiments. 3.2. Determination of receptor-mediated cAMP signalling To ensure that this detection method maintained sufficient sensitivity to be useful in typical cAMP signalling assays, a concentration response curve for a cannabinoid receptor agonist was carried out. Cells were treated with 1 μM forskolin and varying concentrations of the CB1 agonist CP55,940, and the cell lysates were then assayed using the lysate-modified LANCE cAMP assay. As demonstrated in Fig. 2, a concentration-dependent inhibition of forskolin-stimulated cAMP accumulation was detected with pEC50 for CP55,940 of 9.48 ± 0.09 (mean ± standard error of the mean, n = 3), similar to the results recently reported for these same cells utilising a cAMP biosensor (pEC50 9.4 ± 0.1, (Cawston et al., 2013)).
30µl lysate 50µl lysate
,5
0u l
0u l al ,5
0u l
Basal (fmol of Forskolin cAMP per well) stimulated (fmol of cAMP per well) 9.453±3.516 230.22±40.98 2.7975±0.7825 209.2±23.155
FS K
B
3. Results
B
Apparent fold stimulation
24.35 74.78
Fig. 1. Representative data showing the effect of altering lysis buffer volume on cAMP levels in cell lysates. Cells were plated at the same confluency, stimulated and then lysed. A) The graph shows the standard curve, as well as raw data counts for cells stimulated with 1 μM forskolin and then lysed in either 30 μl or 50 μl of lysis buffer. B) Comparison of the apparent effect of forskolin, using the cAMP concentrations as interpolated from the standard curve. cAMP values are presented as fmol of cAMP per well, as calculated by adjusting for the lysis volume and dilution in the detection mix. All data is shown as mean ± SEM.
3.3. Determination of lysate stability through freeze–thaw In order to determine the stability of cAMP in the lysis buffer, cells were treated with forskolin (1 μM) as above, and the resulting lysate was probed twice: first immediately after cell lysis and again after 7 days storage at − 80 °C. Paired T-test analysis showed no statistical difference between cAMP concentration measured before or after freezing (p N 0.05, n = 3).
4. Discussion Here we have presented a simple modification of the LANCE cAMP detection kit, which adds increased flexibility and versatility to the kit's standard applications. This method modification process has also 150
fmol cAMP in well
One potential advantage of a lysate-based detection method is that it enables the samples to be reprobed if the initially-detected value falls outside of the standard curve. For this to be practical, lysates need to be stable following freeze–thaw. To test this, cells were stimulated as described above with forskolin, and some lysate immediately assayed for cAMP. The remaining lysate was frozen at − 80 °C for 7 days, warmed to room temperature (with gentle agitation) and re-assayed.
as
-6
,3
-7
B
-8
FS K
-11 -10 -9
log [cAMP], M
al ,3
0
2.4. Stability of lysate during freeze–thaw
0u l
15000
as
dilutions of europium-streptavidin at 1/18, and biotin-cAMP at 1/6; then each intermediate dilution is further diluted 1/125 in detection buffer to give the detection mix), and complexes were allowed to form for at least 15 min at room temperature. When the detection mix had incubated for at least 15 min, and the lysate-antibody samples had been incubated for 30 min, 12 μl of the detection mix was added per well. The plate was then incubated for a further 60 min (or up to overnight) at room temperature (with optional gentle agitation), before detection in a TR-FRET-capable plate reader as per the manufacturer's instructions. In these experiments, we have utilised a Wallac Victor 1420 Multilabel Counter (Perkin Elmer), with excitation at 340 nm and emission at 615 nm and 665 nm.
counts
44
100
50
0 Basal
FSK
-11
-10
-9
-8
-7
-6
log [CP55,940], M Fig. 2. Inhibition of forskolin-stimulated cAMP accumulation in HEK-HA-hCB1 cells, utilising the modified LANCE assay. Representative data from one of three separate experiments, with each drug condition assayed in triplicate and plotted as mean ± SEM.
M.R. Hunter, M. Glass / Journal of Pharmacological and Toxicological Methods 71 (2015) 42–45
been applied in-house to the AlphaScreen cAMP assay kit (Perkin Elmer) (unpublished data), and can presumably be applied to other commercially-available “add and read” detection systems in order to increase reliability and/or cost-effectiveness. All of the available competition assays require conversion to absolute cAMP levels through the use of a standard curve for accurate data analysis (see Hill et al., 2010 for a discussion on the potential for data misinterpretation from direct detection of cAMP without reference to a standard curve). As such, these assays are highly reliant on data falling within the dynamic range of the standard curve. Here, we have shown that a relatively low concentration of forskolin (1 μM) utilises over 50% of the available dynamic range (Fig. 1), and under our standardised assay conditions all forskolin concentrations above 3 μM resulted in raw data which was outside the dynamic range of the assay (data not shown). However, it is common practise to use concentrations of forskolin well above 5 μM for HEK 293 cells, for example (Bitterman, Ramos-Espiritu, Diaz, Levin, & Buck, 2013; Cawston et al., 2013; Wang, Yan, Zheng, He, & Yang, 2014; Wehbi et al., 2013). Therefore, assay design should consider the expected quantity of cAMP generated. Studies activating Gs-linked GPCRs may require the use of different dilution factors for basal and stimulated conditions to gain accurate measures of their actual cAMP levels—something which could not be achieved in the standard protocol of an “add and read”. Our method modification therefore overcomes what is probably the greatest limitation of the LANCE cAMP kit (and other similar kits), in that it allows samples to be diluted and reprobed if they fall outside of the standard curve, without the need to repeat the experiment. Importantly, we have shown here that samples can be stored frozen without loss of detectable signal. This would also enable samples prepared on separate days to be analysed together for convenience and efficiency, a measure which would also conserve reagents because a single standard curve could be utilised for multiple experiments. The ability to reprobe samples also means that this method requires considerably less optimisation than is required for the standard application of this kit, which requires multiple experiments to establish the conditions that produce signals within the dynamic range. Additionally, the small volumes utilised in the detection (24 μl total) are suitable for detection in 384 well plates, allowing for higher throughput or automation. In this example, we have utilised HEK293 cells transfected with the cannabinoid CB1 receptor, but the lysate-modified LANCE cAMP detection method is adaptable for any adherent cell type, with minimal optimisation from existing protocols set up in any laboratory. This technique also permits the use of any cell culture medium during cell stimulation, because media is removed and washed off; therefore the detection method is not affected by media pigments such as phenol red. Indeed, completing
45
cell stimulation before the detection steps allows researchers to test a wide range of experimental paradigms (e.g. salt concentrations) without concern as to whether these will compromise the subsequent detection assay. Increasingly, pharmacological studies have focused on the “functional selectivity” of ligands—this being the relative effect of ligands on different signalling pathways mediated by the same receptor (Kenakin, 2011). Such comparisons are substantially more valid if cell stimulation protocols are kept as similar as possible across different assays, which this adaptation makes feasible, as the choice of stimulation buffer and temperature will not alter the ability to detect signal. In summary, we show here a straightforward adaptation of the manufacturer's recommendation for a commonly used cell signalling assay, which would improve the utility of the cAMP detection kit. Acknowledgements This research was supported by a Royal Society of New Zealand Marsden (#11-UOA-201) Fund Grant to MG. References Bitterman, J. L., Ramos-Espiritu, L., Diaz, A., Levin, L. R., & Buck, J. (2013). Pharmacological distinction between soluble and transmembrane adenylyl cyclases. Journal of Pharmacology and Experimental Therapeutics, 347, 589–598. Cawston, E. E., Redmond, W. J., Breen, C. M., Grimsey, N. L., Connor, M., & Glass, M. (2013). Real-time characterization of cannabinoid receptor 1 (CB1) allosteric modulators reveals novel mechanism of action. British Journal of Pharmacology, 170, 893–907. Gesty-Palmer, D., & Luttrell, L. M. (2011). Refining efficacy: Exploiting functional selectivity for drug discovery. Advances in Pharmacology, 62, 79–107. Hill, S. J., Williams, C., & May, L. T. (2010). Insights into GPCR pharmacology from the measurement of changes in intracellular cyclic AMP; advantages and pitfalls of differing methodologies. British Journal of Pharmacology, 161, 1266–1275. Hofer, A. M. (2012). Interactions between calcium and cAMP signaling. Current Medicinal Chemistry, 19, 5768–5773. Kearn, C. S., Blake-Palmer, K., Daniel, E., Mackie, K., & Glass, M. (2005). Concurrent stimulation of cannabinoid CB1 and dopamine D2 receptors enhances heterodimer formation: A mechanism for receptor cross-talk? Molecular Pharmacology, 67, 1697–1704. Kenakin, T. (2011). Functional selectivity and biased receptor signaling. Journal of Pharmacology and Experimental Therapeutics, 336, 296–302. PerkinElmer Life and Analytical Sciences, I. (2014). LANCE cAMP 384 Kit Manual. In http://www.perkinelmer.com/CMSResources/Images/44-73586MAN_ LANCEcAMP384KitUser.pdf (Ed.), (Vol. 2014). Shelton, CT, USA: PerkinElmer Life and Analytical Sciences, Inc. Wang, Y., Yan, M., Zheng, G. -y, He, L., & Yang, H. (2014). A cell-based, high-throughput homogeneous time-resolved fluorescence assay for the screening of potential [kappa]-opioid receptor agonists. Acta Pharmacologica Sinica, 35, 957–966. Wehbi, V. L., Stevenson, H. P., Feinstein, T. N., Calero, G., Romero, G., & Vilardaga, J. -P. (2013). Noncanonical GPCR signaling arising from a PTH receptor–arrestin–Gβγ complex. Proceedings of the National Academy of Sciences, 110, 1530–1535.
