British Journal of Pharmacology (2002) 137, 80 – 86
`
`ª 2002 Nature Publishing Group All rights reserved 0007 – 1188/02 $25.00
`www.nature.com/bjp
`
`A comparison of the actions of BIBN4096BS and CGRP8 – 37 on
`CGRP and adrenomedullin receptors expressed on SK-N-MC, L6,
`Col 29 and Rat 2 cells
`
`2Debbie L. Hay, 1Stephen G. Howitt, 1Alex C. Conner, 3Henri Doods, 3Marcus Schindler &
`*,1David R. Poyner
`
`1Pharmaceutical Sciences Research Institute, Aston University, Aston Triangle, Birmingham B4 7ET; 2Department of Metabolic
`Medicine, Imperial College School of Medicine, Hammersmith Hospital, Du Cane Road, London W12 0NN and
`3Cardiovascular Research, Boehringer Ingelheim Pharma KG, 88397, Biberach, Germany
`
`1 The ability of the CGRP antagonist BIBN4096BS to antagonize CGRP and adrenomedullin has
`been investigated on cell lines endogenously expressing receptors of known composition.
`2 On human SK-N-MC cells (expressing human calcitonin receptor-like receptor (CRLR) and
`receptor activity modifying protein 1 (RAMP1)), BIBN4096BS had a pA2 of 9.95 although the slope
`of the Schild plot (1.37+0.16) was significantly greater than 1.
`3 On rat L6 cells (expressing rat CRLR and RAMP1), BIBN4096BS had a pA2 of 9.25 and a
`Schild slope of 0.89+0.05, significantly less than 1.
`4 On human Colony (Col) 29 cells, CGRP8 – 37 had a significantly lower pA2 than on SK-N-MC
`cells (7.34+0.19 (n=7) compared to 8.35+0.18, (n=6)). BIBN4096BS had a pA2 of 9.98 and a
`Schild plot slope of 0.86+0.19 that was not significantly di€erent from 1. At concentrations in
`excess of 3 nM, it was less potent on Col 29 cells than on SK-N-MC cells.
`5 On Rat 2 cells, expressing rat CRLR and RAMP2, BIBN4096BS was unable to antagonize
`concentrations up to 10 mM. CGRP8 – 37 had a pA2 of 6.72 against
`adrenomedullin at
`adrenomedullin.
`6 BIBN4096BS shows selectivity for the human CRLR/RAMP1 combination compared to the rat
`counterpart. It can discriminate between the CRLR/RAMP1 receptor expressed on SK-N-MC cells
`and the CGRP-responsive receptor expressed by the Col 29 cells used in this study. Its slow kinetics
`may explain its apparent ‘non-competive’ behaviour. At concentrations of up to 10 mM, it has no
`antagonist actions at the adrenomedullin, CRLR/RAMP2 receptor, unlike CGRP8 – 37.
`British Journal of Pharmacology (2002) 137, 80 – 86. doi:10.1038/sj.bjp.0704844
`Keywords: CRLR; CGRP; RAMP1; RAMP2; CGRP8 – 37; BIBN4096BS; adrenomedullin; L6; SK-N-MC; Col 29
`Abbreviations: BIBN4096BS, 1-Piperidinecarboxamide, N-[2-[[5amino-1-[[4-(4-pyridinyl)-1-piperazinyl]carbonyl]pentyl]amino]-
`1-[(3,5-dibromo-4-hydroxyphenyl)methyl]-2-oxoethyl]-4-(1,4-dihydro-2-oxo-3(2H)-quinazolinyl); Col 29, Colony
`29; CRLR, Calcitonin receptor-like receptor; RAMP, Receptor activity modifying protein
`
`Introduction
`
`Calcitonin gene related peptide (CGRP) is an abundant, 37
`amino acid neuropeptide (Amara et al., 1982). It is part of a
`peptide family that includes calcitonin, amylin and adreno-
`medullin. CGRP has a complicated pharmacology. The
`peptide fragment CGRP8 – 37 shows a significantly higher
`a(cid:129)nity for CGRP receptors in preparations such as the
`guinea-pig atrium or ileum (pA2 47) compared to tissues
`such as the rat or guinea-pig vas deferens (pA2 *6) (Dennis
`et al., 1990; Quirion et al., 1992; Tomlinson & Poyner, 1996).
`The receptors with a high a(cid:129)nity for CGRP8 – 37 have been
`designated CGRP1 receptors, as opposed to CGRP2 receptors
`that have a lower a(cid:129)nity for this peptide antagonist (Dennis
`et al., 1990; Juaneda et al., 2000). The linear agonists
`[acetamidomethyl-Cys2,7] human aCGRP and [ethylamide-
`Cys2,7] human aCGRP are reported to be CGRP2-selective
`although this is not always observed (Dennis et al., 1989;
`
`*Author for correspondence; E-mail: D.R.Poyner@aston.ac.uk
`
`Dumont et al., 1997; Wisskirchen et al., 1998). The CGRP1
`receptor is a complex formed from a G-protein coupled
`receptor, calcitonin receptor-like receptor (CRLR) and an
`accessory protein,
`receptor activity modifying protein 1
`(RAMP1) (McLatchie et al., 1998). CRLR can associate
`with a second RAMP, RAMP2, to form an adrenomedullin
`receptor.
`Whilst the CGRP1, CGRP2 receptor division has allowed
`the rationalization of much pharmacological data, it may be
`an oversimplification (Poyner & Marshall, 2001). There is no
`molecular correlate for the CGRP2 receptor, although cells
`and tissues expressing CGRP2-receptors also express CRLR
`and RAMP1 (Chakravarty et al., 2000; Rorabaugh et al.,
`2001). There is a 1000 fold spread in the reported a(cid:129)nity
`constants for CGRP8 – 37 (Marshall & Wisskirchen, 2000),
`which is di(cid:129)cult to accomodate within two receptor classes.
`It has never been possible to identify CGRP2 receptors in
`radioligand binding studies (Dennis et al., 1990; Rorabaugh
`et al., 2001).
`
`1
`
`EX2056
`Eli Lilly & Co. v. Teva Pharms. Int'l GMBH
`IPR2018-01426
`
`

`

`D.L. Hay et al
`
`BIBN4096BS action on CGRP and adrenomedullin receptors
`
`81
`
`Many of the problems with classifying CGRP receptors
`are a consequence of having to rely on a single, peptide
`antagonist
`for determining pharmacology. Recently
`a
`number of
`low molecular weight antagonists have been
`described. The best characterized of these is BIBN4096BS
`which arose out of optimization of a dipeptide
`lead
`compound (Doods et al., 2000). This compound had about
`a 200 fold selectivity for primate CGRP receptors (e.g. on
`human neuroblastoma SK-N-MC cells) compared to non-
`primate receptors (e.g. rat spleen)
`(Doods et al., 2000).
`However,
`the pharmacology of
`the rat
`spleen CGRP
`receptor is unclear as CGRP8 – 37 has not been examined
`on this tissue in functional assays. On rat isolated tissues,
`BIBN4096BS showed a 10 fold discrimination between
`CGRP-activated receptors on the rat right atrium and vas
`deferens (Wu et al., 2000). This work also demonstrated that
`BIBN4096BS could act as a potent antagonist against a
`novel receptor on the vas deferens that was activated both
`by [ethylamine-Cys2,7] human aCGRP and adrenomedullin.
`Isolated tissues are likely to contain very complicated
`mixtures of
`receptors;
`for example the guinea-pig vas
`deferens has high a(cid:129)nity binding for CGRP, amylin and
`adrenomedullin (Poyner et al., 1999). Thus it is not clear
`what the molecular composition of the receptors might be in
`the rat spleen, vas deferens and atrium. Accordingly,
`it is
`di(cid:129)cult
`to
`relate
`the
`data
`so
`far
`established
`for
`BIBN4096BS with defined complexes of CRLR and
`RAMPs.
`We have recently investigated the nature of the CGRP
`receptors found in SK-N-MC, L6, Col 29 and Rat 2 cells
`(Choski et al., 2002). This study confirmed that SK-N-MC
`cells expressed CRLR and RAMP1, making this a suitable
`model for a human CGRP1 receptor. Rat L6 cells also
`expressed these components, establishing that
`these are
`suitable models for the rat CGRP1 receptor. Rat 2 cells
`expressed CRLR and RAMP2, making them a model of
`adrenomedullin receptors. Col 29 cells expressed CRLR and
`RAMP1 and had a CGRP1 pharmacology (Choski et al.,
`2002). We and others had previously observed a CGRP2-like
`pharmacology in these cells (Cox & Tough, 1994; Poyner et
`al., 1998).
`In the light of
`this discrepancy,
`it was of
`particular interest to re-examine the nature of the CGRP
`receptor in Col 29 cells using BIBN4096BS to see if it
`discriminates between CGRP-responsive receptors in human
`cells. Accordingly we have examined the behaviour of
`BIBN4096BS on SK-N-MC, L6, Col 29 and Rat 2 cells,
`comparing it with CGRP8 – 37.
`
`Methods
`
`Cell culture
`
`SK-N-MC human neuroblastoma were a gift from Professor
`S. Nahorski, University of Leicester and Col 29 cells were a
`gift from Dr S. Kirkland, Imperial College. L6 and Rat-2
`fibroblasts were purchased from the European Collection of
`Animal Cell Cultures (Porton Down, U.K.). Cells were
`cultured essentially as described previously (Poyner et al.,
`1992; 1998; Coppock et al., 1999). Briefly L6, Rat 2 and Col
`29 cells were grown in Dulbecco’s Modified Eagle Medium
`supplemented with 10% heat inactivated foetal calf serum.
`
`SK-N-MC cells were grown in a 1 : 1 mixture of Dulbecco’s
`Modified Eagle Medium/F12 medium supplemented with
`10% foetal calf serum. Cells were passaged at confluency
`with Trypsin/EDTA (Sigma) and grown for experiments in
`48 well plates at 378C in 5%CO2/air
`in a humidified
`atmosphere.
`
`Measurement of cAMP
`
`An hour prior to experiments, the medium on the cells was
`replaced with Kreb’s solution (for Col 29, SK-N-MC and
`L6 cells) or serum free Dulbecco’s Modified Eagle Medium
`(GIBCO – BRL), both supplemented with 0.1% bovine
`serum albumin and 1 mM isobutyl methyl xanthine. The
`cells were preincubated with BIBN4096BS or CGRP8 – 37 for
`between 30 and 60 min as appropriate before addition of
`increasing concentrations of human (h)aCGRP or
`rat
`adrenomedullin. Incubations were terminated 5 min after
`addition of agonist by aspiration of
`the medium and
`addition of 0.1 ml of absolute ethanol. This was allowed
`to dry down at room temperature and the cAMP was
`extracted by addition of 0.25 ml of assay bu€er containing
`(mM): EDTA, 5, HEPES 20, pH 7.5. The samples were
`agitated for 5 min before 50 ml samples were withdrawn and
`cAMP measured by a radioreceptor assay as described
`previously (Poyner et al., 1992).
`
`Analysis of data
`
`The data from each concentration – response curve was fitted
`to a sigmoidal concentration – response curve to obtain the
`maximum response, Hill Coe(cid:129)cient and EC50 using the
`fitting routine PRISM Graphpad. From the individual
`curves, dose – ratios were calculated and these were used to
`produce the Schild plots shown in Figure 2. The plots were
`fitted by linear regression using PRISM Graphpad. The pA2
`was taken as the x intercept on the Schild plot where the
`slope was unconstrained; the pKb or apparent pKb was taken
`as the intercept where the slope was constrained to 1. For
`CGRP8 – 37, a pA2 was calculated from the dose – ratio
`produced by a single antagonist concentration using the
`formula pA2=log[antagonist]7log (dose ratio71). As com-
`petitive behaviour has previously been demonstrated for
`CGRP8 – 37 on L6, Col 29 and SK-N-MC cells (Poyner et al.,
`1992; 1998), the pA2 can be assumed to be the same as the
`pKb.
`(to
`t-test
`Statistical analysis was either by Student’s
`determine whether the slopes of the unconstrained Schild
`plots were equal to 1) or by one-way ANOVA followed by
`Tukey’s test where several values were being compared.
`Significance was accepted at P50.05; two-tailed tests were
`used throughout. All values are quoted as means+s.e.mean.
`
`Drugs and materials
`
`Ingelheim,
`synthesized at Boehringer
`BIBN4096BS was
`Pharma KG. Human aCGRP was purchased from Neosys-
`tems (Strassbourg, France). Human aCGRP was custom
`synthesised by ASG University (Szedgel, Hungary). Rat
`adrenomedullin was from Bachem (St. Helens, Merseyside,
`U.K.). Human aCGRP8 – 37 and peptidase inhibitors were
`purchased from Calbiochem. Isobutyl methyl xanthine was
`
`British Journal of Pharmacology vol 137 (1)
`
`2
`
`

`

`82
`
`D.L. Hay et al
`
`BIBN4096BS action on CGRP and adrenomedullin receptors
`
`purchased from Sigma (Sigma-Aldrich, Gillingham, Dorset,
`U.K.). Cell culture medium and foetal calf serum were
`purchased from GIBCO – BRL (Life Technologies, Paisley,
`Renfrewshire, U.K.). Other reagents were purchased from
`Sigma or Fisher. BIBN4096BS was dissolved in a small
`volume of 0.1 M hydrochloric acid, the pH adjusted to 7 with
`sodium hydroxide and diluted to give a stock solution of
`100 mM (Wu et al., 2000). Both it and peptides were stored as
`frozen aliquots before use as previously described (Poyner et
`al., 1998).
`
`Results
`
`L6 cells (CRLR+RAMP1, CGRP1 receptor)
`
`A range of BIBN4096BS concentrations from 0.3 nM to 1 mM
`were examined on L6 cells. In the absence of BIBN4096BS,
`haCGRP stimulated cAMP production with a pEC50 of
`8.61+0.03 (n=6). BIBN4096BS caused a rightwards shift in
`the concentration response curve to CGRP, as illustrated in
`Figure 1. At lower concentrations there was no significant
`depression of the maximum response; however at the highest
`concentrations it was not possible to use su(cid:129)cient CGRP to
`examine the maximum response. After calculation of dose –
`ratios, a Schild plot was made of the data (Figure 2). This
`gave a slope of 0.89+0.05, significantly di€erent from 1
`(P50.05). The pA2 was estimated from the x axis intercept as
`9.25. On the same batch of cells, CGRP8 – 37 caused a parallel
`shift
`in the concentration response curve to CGRP;
`the
`dose – ratio was used to calculate an apparent pA2 of 7.81
`(Figure 3), in line with previous estimates for the a(cid:129)nity of
`CGRP8 – 37 on these cells (Poyner et al., 1992).
`
`SK-N-MC cells (CRLR+RAMP1, CGRP1 receptor)
`
`On SK-N-MC cells, haCGRP stimulated cAMP production
`with a pEC50 of 9.46+0.05 (n=23). BIBN4096BS produced a
`rightwards shift in the concentration response curve; as with
`the L6 cells, it was not possible to establish whether CGRP
`could completely overcome
`the
`e€ects of
`the highest
`concentrations of BIBN4096BS (Figure 1). The Schild plot
`gave a slope of 1.37+0.16, significantly greater than 1
`(P50.05). The pA2 estimated from the x intercept was
`estimated as 9.95; however, due to the high slope of the
`Schild plot, this will underestimate the true pKb (Figure 2).
`Increasing the incubation time with BIBN4096BS from
`30 min to 1 h or repeating the incubations in the presence
`of peptidase inhibitors (1 mM AEBSF, 0.8 mM aprotinin,
`50 mM bestatin, 15 mM E-64, 20 mM leupeptin and 10 mM
`pepstatin) had little e€ect on the potency of BIBN4096BS
`(dose – ratio to 0.3 nM BIBN4096BS; control 3.2, 60 min
`preincubation 3.5, with peptidase inhibitors 4.0).
`
`Col 29 cells (CRLR+RAMP1?, CGRP2 receptor?)
`
`On Col 29 cells, haCGRP stimulated cAMP production with
`a pEC50 of 8.38+0.05 (n=6). BIBN4096BS produced a
`rightwards shift in the concentration response curve (Figure
`the Schild plot was 0.86+0.19, not
`1). The slope of
`significantly from 1. The pA2 taken from the x intercept
`was 9.98. As the slope was not significantly di€erent from
`
`British Journal of Pharmacology vol 137 (1)
`
`Figure 1 E€ects of BIBN4096BS on the stimulation of cAMP
`production by human aCGRP in the presence of the indicated
`concentrations of BIBN4096BS in (a) L6 cells, (b) SK-N-MC cells
`and (c) Col 29 cells. Data represent means+s.e.mean of two to five
`experiments. Points were measured in duplicate in each experiment.
`Data are expressed as percentage of maximum cAMP production,
`estimated by fitting each line to a logistic Hill equation as described
`in the Methods. Maximum cAMP values were as follows: L6 cells,
`270+30 pmol per 106 cells; SK-N-MC cells, 240+20 pmol per 106
`cells; Col 29 cells, 170+19 pmol per 106 cells. Basal values were all
`below 10 pmol per 106 cells.
`
`unity, it was possible to constrain this to 1, giving a pKb of
`9.75+0.14. If the slope of the Schild plot for SK-N-MC cells
`was constrained to 1, the resulting intercept of 10.49+0.017
`was significantly greater than the pKb measured on the Col
`29 cells (P50.001).
`We have previously noted considerable variability in the
`a(cid:129)nity of Col 29 cells to CGRP8 – 37 (Poyner et al., 1998;
`Choski et al., 2002). The a(cid:129)nity of CGRP8 – 37 on the Col 29
`cells used in the present study was compared with the SK-N-
`MC cells (Figure 2). Based on the dose – ratios measured in
`these experiments, the apparent pA2 for CGRP8 – 37 on the
`Col 29 cells was 7.34+0.19 (n=7), significantly less (P50.01)
`than that estimated for the SK-N-MC cells (8.35+0.18,
`n=6).
`
`3
`
`

`

`D.L. Hay et al
`
`BIBN4096BS action on CGRP and adrenomedullin receptors
`
`83
`
`Figure 2 Schild transform of concentration – response curves shown
`in Figure 1; (a) fitted to a straight line with unconstrained slope and
`(b) fitted to a straight line with slope constrained to 1.
`
`The nature of the BIBN4096BS-interacting receptor in
`SK-N-MC, Col 29 and L6 cells
`
`The pA2 values for BIBN4096BS as estimated by the x-
`intercept
`from the Schild plots are very similar for all
`three cell
`lines. However, only for the Col 29 cells is the
`Schild slope consistent with strict competitive behaviour;
`the pA2 for the SK-N-MC cells will be an underestimate
`that
`for
`the L6 cells will be an
`the pKb whilst
`of
`overestimate. In Table 1, the data from the Schild plots is
`summarized;
`it can be seen that the dose – ratios for all
`three cell lines are significantly di€erent from each other at
`high concentrations of BIBN4096BS. This argues
`that
`BIBN4096BS is able to distinguish between the receptors
`on each of the three cell lines. Furthermore, BIBN4096BS
`has very slow on and o€ rates;
`this suggests that
`the
`apparent non-competitive behaviour
`is
`simply a kinetic
`artefact (see Discussion). If this is accepted and the slopes
`of the Schild plots for all three cell lines are constrained to
`1, the intercepts are significantly di€erent from each other
`(Figure 2b, Table 1).
`
`Rat 2 cells (CRLR+RAMP2, adrenomedullin receptor)
`
`Human aCGRP was inactive on Rat 2 cells at concentrations
`of up to 1 mM but rat adrenomedullin caused a concentra-
`tion-dependant stimulation of cAMP production with a
`pEC50 of 8.75+0.11 (n=3) (Figure 4). This was weakly
`antagonized by CGRP8 – 37, with a pA2 of 6.72+0.06 (n=3),
`estimated from the shift in the concentration – response curve
`produced by 1 mM of the antagonist. At concentrations of up
`to 10 mM, BIBN4096BS was unable to antagonize adrenome-
`dullin.
`
`Figure 3 E€ects of CGRP8 – 37 on the stimulation of cAMP
`production by human aCGRP in the presence of the indicated
`concentrations of antagonist in (a) L6 cells, (b) SK-N-MC cells and
`(c) Col 29 cells. Data represent means+s.e.mean of four to seven
`experiments. Points were measured in duplicate in each experiment.
`Data are expressed as percentage of maximum cAMP production,
`estimated by fitting each line to a logistic Hill equation as described
`in the Methods. Maximum cAMP values were as follows: L6 cells,
`250+27 pmol per 106 cells; SK-N-MC cells, 230+18 pmol per 106
`cells; Col 29 cells, 180+17 pmol per 106 cells. Basal values were all
`below 10 pmol per 106 cells.
`
`Discussion
`
`This study describes the ability of BIBN4096BS to interact with
`cell lines expressing human and rat CGRP1(CRLR/RAMP1)
`receptors (i.e. SK-N-MC and L6 cells), rat adrenomedullin
`receptors (CRLR/RAMP2, Rat 2 cells) and a potential human
`CGRP2-like receptor (Col 29 cells). The use of cell
`lines
`expressing endogenous receptors has some advantages over
`artificially transfected systems. In particular, it is apparent that
`batches of some cell lines such as HEK293T and Cos 7 cells
`express endogenous CRLR or RAMPs. Some batches of Cos 7
`cells express an endogenous CRLR that accounts for up to 30%
`of the binding of CGRP when they are transfected with human
`RAMP1 (A Conner, unpublished observations). Thus the
`response seen with these cells will reflect a pattern of mixed
`monkey/human CRLR expression.
`
`British Journal of Pharmacology vol 137 (1)
`
`4
`
`

`

`84
`
`D.L. Hay et al
`
`BIBN4096BS action on CGRP and adrenomedullin receptors
`
`Table 1 Analysis of Schild plots for BIBN4096BS on L6, Col 29 and SK-N-MC cells
`
`Cell line
`
`pA2
`
`Schild plot analysis
`y intercept
`Slope
`
`‘Apparent’ pKb
`
`1 nM
`
`Log (dose – ratio), BIBN4096BS
`3 nM
`10 nM
`
`30 nM
`
`9.25
`L6
`SK-N-MC 9.95
`Col 29
`9.98
`
`0.89+0.05
`1.37+0.16
`0.86+0.19
`
`8.24+0.39
`13.60+1.37
`8.59+1.60
`
`9.10+0.05
`10.47+0.11
`9.75+0.10
`
`0.19+0.16 (3)
`1.36+0.21 (5)
`0.82+0.12 (4)
`
`0.79+0.07 (3)
`2.20+0.24 (3)
`1.43+0.31 (2)
`
`1.12+0.20 (3)
`2.57+0.26 (5)
`1.63+0.25 (4)
`
`1.69+0.12 (3)
`3.34+0.03 (2)
`2.19+0.23 (2)
`
`Values represent means+s.e.mean with sample size in brackets. pA2, slopes and y intercepts taken from Schild plots. The y intercepts
`are all significantly di€erent from each other (P50.05). The ‘apparent’ Kb was calculated by constraining the slope of the Schild plot to
`1 and calculating the intercept; as the 95% confidence limits for the L6 and SK-N-MC cells marginally exclude 1 (SK-N-MC, 1.036 to
`1.697; Col 29, 0.786 to 0.995), this will not give the true pKb. The values are all significantly di€erent from each other (P50.05). For
`the dose – ratios, the following are significantly di€erent: 1 nM; SK-N-MC versus L6, (P50.001), Col 29 versus L6 (P50.05), 3 nM; SK-
`N-MC versus Col 29 (P50.05), SK-N-MC versus L6 (P50.001), Col 29 versus L6 (P50.05), 10 nM; SK-N-MC versus L6 (P50.05),
`SK-N-MC versus Col 29 (P50.01), 30 nM; SK-N-MC versus L6 and Col 29 (P50.001), Col 29 versus L6 (P50.05).
`
`to the peptide (Poyner et al., 1992). For both SK-N-MC and
`L6 cells, if the dose – ratios obtained with very low (0.3 nM)
`and high (300 nM, 1 mM) concentrations of BIBN4096BS are
`excluded from the Schild analysis, then the resulting plots
`have slopes that are not significantly di€erent from 1. Thus it
`remains possible that the underlying reason for the apparent
`non-competitive inhibition is related to the experimental
`conditions used in this study.
`L6 cells express adrenomedullin receptors in addition to
`CGRP receptors (Coppock et al., 1996). This may provide an
`additional explanation of
`the apparent non-competitive
`behaviour of BIBN4096BS at
`these cells. Although the
`adrenomedullin receptors are not directly linked to stimula-
`tion of cAMP production, it is possible that at high CGRP
`concentrations, CGRP can activate these receptors and that
`indirectly this leads to some cAMP production that is not
`antagonized by BIBN4096BS.
`For SK-N-MC cells, the pA2 estimated from the x intercept
`with an unconstrained slope (9.95) probably underestimates the
`true a(cid:129)nity. Assuming that BIBN4096BS acts predominantly
`in a competitive fashion at the CRLR/RAMP1 complex, then
`the ‘apparent pKb’
`in Table 1 of 10.57, determined by
`constraining the slope of the Schild plot to 1 is probably a
`better estimate, although even this may be an underestimate.
`BIBN4096BS is a more potent antagonist on L6 cells than
`CGRP8 – 37, in line with what would be predicted from the other
`published studies (Doods et al., 2000; Wu et al., 2000). Given
`the problems with the slope of the Schild plots, it is not possible
`to estimate how much more potent BIBN4096BS is on human
`CRLR/RAMP1 compared to rat CRLR/RAMP1, as the ratio
`depends on the concentration of BIBN4096BS used. Based on
`the data shown in Figure 2a and Table 1, this could vary
`between 5 fold to over 10,000 fold; a comparison of the
`apparent pKb values in Table 1 shows a 24 fold selectivity.
`The results clearly demonstrate that BIBN4096BS shows
`preferential binding to the human CRLR/RAMP1 complex
`in functional assays and that the di€erences in a(cid:129)nity reported
`previously are due to genuine species di€erences within the
`same receptor subtype.
`It has been previously reported that BIBN4096BS has a
`very low a(cid:129)nity at displacing adrenomedullin in radioligand
`binding assays (Doods et al., 2000); however it was a good
`antagonist at an adrenomedullin receptor in the rat vas
`deferens. The inability to antagonize adrenomedullin at the
`CRLR/RAMP2 complex expressed by Rat 2 cells demon-
`strates that it shows at least 1000 fold preference for CRLR/
`RAMP1 in rats. The nature of the adrenomedullin receptor
`
`Figure 4 E€ects of CGRP8 – 37 and BIBN4096BS on the stimulation
`of cAMP production by rat adrenomedullin in Rat 2 cells. Data
`represent means+s.e.mean of three to four experiments. Points were
`measured in triplicate in each experiment. Data are expressed as
`percentage of maximum cAMP production, estimated by fitting each
`line to a logistic Hill equation as described in the Methods.
`Maximum cAMP values were 25+2.7 pmol per 106 cells. Basal
`values were all below 1 pmol per 106 cells.
`
`In confirmation of previous observations (Doods et al.,
`2000; Wu et al., 2000; Edvinsson et al., 2002), BIBN4096BS
`acted as a very potent antagonist at all CGRP-responsive
`receptors tested. The reported pA2 values of 11 and 11.2 from
`the studies of Doods et al. (2000) and Edvinsson et al. (2002)
`are in good agreement with the values from our study at high
`antagonist concentration (41 nM). However, it is apparent
`that in SK-N-MC cells under the experimental conditions
`used in this study, BIBN4096BS does not
`follow strict
`competitive inhibition as the slope of the Schild plot is
`greater than 1. A similar result was found with the L6 cells.
`In this case, the Schild plot slope is less than 1. In neither
`case is the discrepancy large. For SK-N-MC cells, we have
`attempted to investigate whether this may be due to peptidase
`activity (BIBN4096BS contains a peptide bond, albeit highly
`modified) but the cocktail of, peptidase inhibitors used in this
`study made little di€erence. Increasing the incubation time
`also made little di€erence. However, using [3H]-BIBN4096BS,
`slow on/o€-kinetics of
`this
`compound was observed
`compared to that of
`[125I]-CGRP (Schindler & Doods,
`2002). At high or low antagonist concentration, kinetic
`artefacts are more likely to occur. Thus at low concentra-
`tions, even 60 min may not be su(cid:129)cient to allow the drug to
`reach equilibrium. At high concentrations it is possible that
`once the BIBN4096BS is bound, its slow o€-rate means there
`is insu(cid:129)cient time for the CGRP to equilibrate with the
`receptor. It is not practical to increase the incubation time for
`CGRP, as desensitization is observed after 10 min exposure
`
`British Journal of Pharmacology vol 137 (1)
`
`5
`
`

`

`D.L. Hay et al
`
`BIBN4096BS action on CGRP and adrenomedullin receptors
`
`85
`
`in the rat vas deferens antagonized by BIBN4096BS remains
`unknown, but this data suggests that it is most unlikely to be
`a simple CRLR/RAMP2 complex. It is interesting to note
`that against adrenomedullin, CGRP8 – 37 has a pA2 of 6.72;
`greater or equal to that found against CGRP2-like receptors
`(Dennis et al., 1990; Wisskirchen et al., 1998; Marshall &
`Wisskirchen, 2000). Thus it should be used with care,
`particularly to distinguish between adrenomedullin and
`CGRP when the latter acts through the CGRP2 group of
`receptors. It remains a formal possibility that BIBN4096BS
`has a higher a(cid:129)nity against
`the human CRLR/RAMP2
`complex than the rat equivalent examined in this study.
`Sexton et al. (2001) have argued that BIBN4096BS is likely
`to derive its specificity for CGRP over adrenomedullin by
`binding to RAMP1 rather than CRLR. Recently Kane and
`co-workers have produced radioligand binding data to suggest
`that BIBN4096BS interacts with residue 74 of human RAMP1
`(Mallee et al., 2002) and that this is responsible for the human
`versus rat selectivity. The functional data from this study is
`consistent with these findings. It would be unsurprising if the
`binding domains for BIBN4096BS and CGRP do not fully
`overlap, although it is likely that they will share some points
`of contact both on RAMP1 and CRLR. The greater
`selectivity of BIBN4096BS for CRLR/RAMP1 over CRLR/
`RAMP2 compared to CGRP8 – 37 may imply that the latter
`compound undergoes greater interactions with CRLR. As this
`is common to both receptors, its selectivity would be predicted
`to be less than that of BIBN4096BS. It will be of considerable
`interest to compare these compounds on the CRLR/RAMP3
`complex, which functions as a mixed adrenomedullin/CGRP
`receptor (McLatchie et al., 1998; Husmann et al., 2000).
`Col 29 cells are derived from human colonic epithelium
`(Kirkland, 1986). These have been reported to show CGRP2-
`like pharmacology (Cox & Tough, 1994). Recently, we have
`found the cells to express CRLR and RAMP1 and to show a
`pharmacology identical to that of SK-N-MC cells (Choski et
`al., 2002), in contrast to our previous data (Poyner et al., 1998).
`It was not possible to account for the change in behaviour of
`these cells. In the present study a fresh batch of cells was
`examined, although ultimately it was derived from the same
`source (Kirkland, 1985) as those examined previously by Cox
`and Tough and ourselves. CGRP8 – 37 had a pA2 estimated from
`a single antagonist concentration in excess of 7. This would
`normally be considered to be diagnostic of a CGRP1 receptor;
`however, this a(cid:129)nity was significantly less than that measured
`on SK-N-MC cells with the same batch of CGRP8 – 37. To try
`and control the a(cid:129)nity of the Col 29 cells for CGRP8 – 37 a
`variety of manipulations were carried out, including serum
`starving the cells for 24 h before use or exchanging normal and
`heat inactivated foetal calf serum. However, these did not
`
`References
`
`AMARA, S.G., JONES, V., ROSENFELD, M.G., ONG, E.S. & EVANS,
`R.M. (1982). Alternative RNA processing in calcitonin gene
`expression generates mRNA encoding di€erent polypeptide
`products. Nature (Lond)., 298, 240 – 244.
`CHAKRAVARTY, P., SUTHAR, T.P., COPPOCK, H.A., NICHOLL, C.G.,
`BLOOM, S.R., LEGON, S. & SMITH, D.M. (2000). CGRP and
`adrenomedullin binding correlates with transcript levels for
`calcitonin receptor-like receptor (CRLR) and receptor activity
`modifying proteins (RAMPs) in rat tissues. Br. J. Pharmacol.,
`130, 189 – 195.
`
`produce any reproducible e€ect on the behaviour of the Col 29
`cells. It is possible that the a(cid:129)nity for CGRP8 – 37 is determined
`by one or more accessory factors, perhaps acting in concert
`with CRLR and RAMP1, and that irregular expression of these
`factors are the cause of the variable a(cid:129)nity for CGRP8 – 37. It
`should be noted that Cox and Tough reported no e€ects of
`3 mM CGRP8 – 37, whereas in our hands this concentration did
`antagonize CGRP (Poyner et al., 1998). Thus the variation in
`response to CGRP8 – 37 seems to be a long-standing phenom-
`enon.
`On the batch of Col 29 cells used in the current
`experiment, BIBN4096BS proved to be a more potent
`antagonist than CGRP8 – 37. Unlike the SK-N-MC cells, the
`slope of the Schild plot was not significantly di€erent from
`unity. At low concentrations, BIBN4096BS was as potent as
`on the SK-N-MC cells. However at higher concentrations,
`the dose – ratio values and hence the a(cid:129)nity were significantly
`lower on the Col 29 cells compared to the SK-N-MC cells.
`Thus both CGRP8 – 37 and BIBN4096BS can discriminate
`between the CGRP-activated receptors on these cell
`lines.
`Whether it is appropriate to call the receptor on the Col 29
`cells ‘CGRP2’ receptors is less clear, given the apparently
`high a(cid:129)nity for CGRP8 – 37. It is clear that the Col 29 cells
`can change their properties in culture to show a range of
`a(cid:129)nities for CGRP8 – 37 and so they must be used with care.
`The present study does support the concept that in humans
`CGRP can activate a variety of receptors with high a(cid:129)nity.
`However, only the molecular nature of the CGRP1 CRLR/
`RAMP1 receptor is understood. The ‘CGRP2’ receptor may
`be a distinct molecular entity or it may represent activation
`of some other receptor (amylin or adrenomedullin) by CGRP
`(see Poyner et al., 2002 for further discussion of this issue).
`In conclusion,
`this
`study describes
`the behaviour of
`BIBN4096BS on a number of CGRP and adrenomedullin
`receptors of known molecular composition. BIBN4096BS has
`a higher a(cid:129)nity for human CRLR/human RAMP1 compared
`to rat CRLR/rat RAMP1. It has a very low a(cid:129)nity for rat
`CRLR/rat RAMP2. However, under the assay conditions
`tested,
`the
`blockade
`of CGRP-induced
`e€ects
`by
`BIBN4096BS appear, in part, to be insurmountable. It can
`discriminate between the CRLR/RAMP1 complex expressed
`by SK-N-MC cells and the CGRP-activated receptor on Col
`29 cells. This latter receptor may consist of CRLR and
`RAMP1 with an unknown accessory factor. However the
`pharmacology of
`the Col 29 cells appears
`to show
`considerable variation with batch and passage number.
`
`We wish to thank the Wellcome Trust for support of this work.
`D.L. Hay was supported by a MRC studentship.
`
`CHOSKI, T., HAY, D.L., LEGON, L., POYNER, D.R., HAGNER, S.,
`(2002). Comparison of
`the
`BLOOM, S.R. & SMITH, D.M.
`expression of calcitonin receptor-like receptor (CRLR) and
`receptor activity modifying proteins (RAMPs) with CGRP and
`adrenomedullin binding in cell
`lines. Br. J. Pharmacol., 136,
`784 – 792.
`
`British Journal of Pharmacology vol 137 (1)
`
`6
`
`

`

`86
`
`D.L. Hay et al
`
`BIBN4096BS action on CGRP and adrenomedullin receptors
`
`COPPOCK, H.A., OWJI, A.A., AUSTIN, C., UPTON, P.D., JACKSON,
`M.L., GARDINER, J.V., GHATEI, M.A., BLOOM, S.R. & SMITH,
`D.M. (1999). Rat-2 fibroblasts express specific adrenomedullin
`receptors, but not calcitonin-gene-related-peptide receptors,
`which mediate increased intracellular cAMP and inhibit mito-
`gen-activated protein kinase activity. Biochem. J., 338, 15 – 22.
`COPPOCK, H.A., OWJI, A.A., BLOOM, S.R. & SMITH, D.M. (1996). A
`rat skeletal muscle cell line (L6) expresses specific adrenomedullin
`binding sites but activates adenylate cyclase via calcitonin gene-
`related peptide receptors. Biochem. J., 318, 241 – 245.
`COX, H.M. & TOUGH, I.R. (1994). Calcitonin gene-related peptide
`receptors in human gastrointestinal epithelia. Br. J. Pharmacol.,
`113, 1243 – 1248.
`DENNIS, T