Adrenomedullin and CGRP interact with endogenous ... · Adrenomedullin (AM) and calcitonin gene-related peptide (CGRP) are related peptides with distinct pharmacological proﬁles.

910 Research Article

IntroductionThe calcitonin family of peptide hormones comprises six knownmembers [calcitonin, amylin, two calcitonin-gene-relatedpeptides (CGRP-� and CGRP-�), adrenomedullin (AM) andthe recently discovered intermedin], which are involved incellular proliferation, survival and migration (supplementarymaterial, Fig. S1) (Poyner et al., 2002; Roh et al., 2004). AMand CGRP play significant roles in vascular biology (forreviews, see Hinson et al., 2000; Brain and Grant, 2004).

Circulating AM levels are upregulated in cardiovasculardisease and sepsis (reviewed by Hinson et al., 2000); and local

concentrations are elevated in some tumours (Satoh et al.,1995). Accumulating experimental evidence from animalmodels suggests a significant role for AM in these pathologiessince it protects against hypertension and ischaemia- or injury-induced vascular remodelling, regulates blood vesselpermeability in response to inflammatory mediators, andpromotes tumour angiogenesis (Shindo et al., 2000; Matsui etal., 2004; Kawai et al., 2004; Brell et al., 2005; Oehler et al.,2002; Martinez et al., 2002). Similar to AM, CGRP is asignificant player in maintaining vascular function, in additionto its known role as a neuropeptide, and has been implicated

Adrenomedullin (AM) and calcitonin gene-related peptide(CGRP) are related peptides with distinct pharmacologicalprofiles. Calcitonin-receptor-like receptor (CRLR, nowknown as CL) can function as either an AM receptor or aCGRP receptor, when cotransfected with receptor-activity-modifying proteins (RAMPs) that define ligand-bindingspecificity. The aim of the present study was to determinethe role of endogenously expressed CL (EndoCL) ingenerating endogenous AM and CGRP receptors. Weraised anti-human CL antibody and identifiedmicrovascular endothelial cells (MVECs) as a major CL-expressing cell type in tissues by immunohistochemistry.Cultured MVECs continue to express EndoCL as well asfully active endogenous AM- and CGRP-sensitive receptorsin vitro, as demonstrated by the ability of both peptides toinduce migration and Akt phosphorylation. We thereforetested the hypothesis that endothelial EndoCL can interactwith both AM and CGRP by examining receptorinternalisation and desensitisation (loss of the ability toinduce Akt phosphorylation). We found that agonist-mediated internalisation of EndoCL occurs in response toAM but not CGRP in MVECs. However, AM-inducedEndoCL internalisation was blocked by antagonists of bothAM and CGRP receptors: AM22-52 and CGRP8-37,

respectively. Furthermore, AM-induced EndoCLinternalisation resulted in desensitisation not only of AMbut also of CGRP receptors. Finally, CGRP also induceddesensitisation of both endogenous AM and CGRPreceptors, but did not mediate EndoCL internalisationdespite interaction with this receptor. Thus, EndoCLinteracts with both AM and CGRP, and simultaneouslyacts as a receptor for both peptides (i.e acting as anendogenous AM/CGRP receptor) in endothelial cells.Interaction with either ligand is sufficient to induceEndoCL desensitisation to both AM and CGRP, butdifferential mechanisms are involved since only AMinduces EndoCL internalisation. These novel findingsregarding regulation of EndoCL function in endothelialcells are likely to be of importance in conditions where AMor CGRP levels are elevated, such as cardiovasculardisease, diabetes and inflammation.

Supplementary material available online athttp://jcs.biologists.org/cgi/content/full/119/5/910/DC1

Key words: Adrenomedullin, CGRP, CRLR, Endogenous CL,Endothelial cell

Summary

Adrenomedullin and CGRP interact with endogenouscalcitonin-receptor-like receptor in endothelial cellsand induce its desensitisation by differentmechanismsLeonid L. Nikitenko1,2,3,*, Nicola Blucher1, Stephen B. Fox4, Roy Bicknell2, David M. Smith5 andMargaret C. P. Rees1

1Nuffield Department of Obstetrics and Gynaecology, The University of Oxford, John Radcliffe Hospital, Oxford, OX3 9DU, UK2Molecular Angiogenesis Laboratory, Cancer Research UK, Weatherall Institute of Molecular Medicine, The University of Oxford, John RadcliffeHospital, Oxford, OX3 9DU, UK3Cancer Research UK Viral Oncology Group, Wolfson Institute for Biomedical Research, University College London, London, WC1E 6BT, UK4Nuffield Department of Clinical Laboratory Sciences, The University of Oxford, John Radcliffe Hospital, Oxford, OX3 9DU, UK5AstraZeneca, CVGI, Alderley Park, Macclesfield, Cheshire, SK10 4TG, UK*Author for correspondence (e-mail: [email protected])

Accepted 7 November 2005Journal of Cell Science 119, 910-922 Published by The Company of Biologists 2006doi:10.1242/jcs.02783

Jour

nal o

f Cel

l Sci

ence

911Desensitisation of endogenous endothelial CL

in pathophysiology of subarachnoid haemorrhage, migrane andcoronary artery disease (reviewed by Brain and Cambridge,1996). However, despite the proposed role for AM and CGRPin various pathologies, the therapeutic potential of regulatingtheir action in humans remains unclear. This is a result of thelimited information about the distribution of endogenous AMand CGRP receptors, as well as mechanisms regulating theirfunction (Hay et al., 2004; Kuwasako et al., 2004).

Effects of AM and CGRP are mediated throughheterodimeric receptors composed of calcitonin-receptor-likereceptor (CRLR, now known as CL) and one of the threereceptor-activity-modifying proteins (RAMPs), asdemonstrated in receptor system reconstitution models(McLatchie et al., 1998; Poyner et al., 2002). CL belongs tothe family B of seven-transmembrane (7TM) G-protein-coupled receptors (GPCRs). The RAMP family comprisesthree members (RAMP1, RAMP2 and RAMP3) that share lessthan 30% sequence identity but a common topologicalorganisation (reviewed by Sexton et al., 2001). RAMPs areessential for terminal glycosylation, cell-surface targeting andligand-binding selectivity of CL (McLatchie et al., 1998). Theglycosylation state of the CL receptor is crucial to itsproperties. Only mature, fully glycosylated CL species areexpressed at the cell surface and are selectively recognised byboth AM and CGRP (Hilairet et al., 2001a). CL-RAMPheterodimer formation at the cell surface defines the ligand-binding selectivity of the CL glycoprotein (McLatchie et al.,1998). RAMP1 promotes the expression of CGRP receptor,whereas co-expression of RAMP2 or RAMP3 with CL leadsto the formation of AM receptors, termed AM1 and AM2,respectively (McLatchie et al., 1998; Poyner et al., 2002).

On the basis of CL-RAMP cotransfectant models as reportedabove, it has been proposed that the known effects of AM andCGRP on vascular cells are mediated throughpharmacologically distinct receptors that also mediate theactions of the endogenous CL (EndoCL) receptor in a similarmanner (Kamitani et al., 1999; Kim et al., 2003). However, nointeraction of either peptide with EndoCL in vascular cells hasyet been demonstrated. Furthermore, whereas pharmacologicalstudies suggest the heterogeneity among endogenous AM andCGRP receptors, their nature, and a role for CL in particular,remains unclear (reviewed by Sexton et al., 2001) (Poyner etal., 2002; Kuwasako et al., 2004; Hay et al., 2004). Forexample, the CL-RAMP1 heterodimer matches thepharmacology of the CGRP1 receptor in cell lines and tissues,and is mainly characterised by high affinity for CGRPantagonists such as CGRP8-37 and BIBN4096BS (Poyner et al.,2002). However, in some tissues, another CGRP-responsivereceptor is present, the CGRP2 receptor, which has a lowaffinity for CGRP antagonists. There is no simple molecularcorrelate for this receptor, but it has been suggested that thephenotype might be a result of expression of complexes suchas CL-RAMP3 (AM2 receptor), since it has appreciableaffinity to CGRP8-37 (Hay et al., 2004).

The lack of consistent data using ideal reconstitutionsystems [i.e. equivalent species of RAMPs and CL proteins orunmodified (untagged) cDNA constructs] makes it difficult tojudge the significance of pharmacological studies ontransfected receptor heterodimers in relation to the endogenousreceptor system (Sexton et al., 2001; Hay et al., 2004; Hay etal., 2005). Cellular components other than just CL and RAMPs

could also be important for the expressed phenotype and theproperties of CL-associated receptor subtypes, such as themagnitude of agonist-mediated responses and mechanisms ofinternalisation and desensitisation (for reviews, see Sexton etal., 2001; Hay et al., 2003) (Evans et al., 2000; Bomberger etal., 2005). Furthermore, little is known about the distributionof CL and RAMPs in tissues. In view of the above, and sincethe role for CL-RAMP heterodimers in generating functionalendogenous AM and CGRP receptor subtypes has yet to bedemonstrated, we have searched for a model to study EndoCL.

We initially investigated the distribution of CL in humantissues, and then analysed its properties (glycosylation state,subcellular localisation and function both in normal conditionsand when agonist levels are elevated) specifically inmicrovascular endothelium, as this is a major cell typeexhibiting endogenous expression of the receptor in vivo. Here,we demonstrate that the endothelial EndoCL can interact witheither AM or CGRP. We show that these interactions result indesensitisation of both endogenous AM and CGRP receptors,irrespective of which agonist was used for initial stimulation,and that this occurs through different mechanisms, sinceEndoCL internalisation was induced by AM but not by CGRP.We conclude that AM- or CGRP-induced desensitisation ofEndoCL, as a key endothelial AM/CGRP receptor, might haveadverse effects on endothelial cell biology in diseasesassociated with upregulated levels of either of these two ligands,such as cardiovascular disease, neoplasia and inflammation.

ResultsGPCR CL is predominantly expressed by microvascularendothelium in vivo and in vitroTo examine CL distribution in human tissues, we have raisedan anti-human CL antibody and characterised its specificityusing transient transfection of human embryonic kidney(HEK293T) cells expressing functional human (h)CL-RAMP1and hCL-RAMP2 receptor heterodimers and various tissues(Fig. 1). Polyclonal anti-hCL antibody LN-1436 specificallyrecognised both transiently expressed and endogenouslyexpressed hCL receptor (Fig. 1). HEK293T cells transientlyexpressing hCL produced a distinct band of ~45 kDa (Fig. 1A).Signal of similar size was observed in the majority of tissuelysates (Fig. 1B). Co-expression of hCL with RAMP1 orRAMP2 in HEK293T caused the ~45 kDa band to disappearand a diffuse band of about 50-55 kDa to appear (Fig. 1A).This suggests that antibody LN-1436 can also recognise theterminally glycosylated form of the CL receptor (Hilairet et al.,2001a). This band was also present in tissues lysates (Fig. 1B).

We then used endoglycosidases F and H to differentiatebetween N-linked glycoproteins to confirm the nature of CLspecies that were detected by the antibody LN-1436 (Fig. 1).When the 45 kDa band (representing hCL expressed in theabsence of RAMPs; supplementary material, Fig. S2, lane 1)was treated with endoglycosidase F, a single immunoreactivespecies of 37 kDa was seen after SDS-PAGE (supplementarymaterial, Fig. S2, lane 2). Thus, the ~45 kDa form is aglycoprotein (supplementary material, Fig. S2, lane 1). The~55 kDa form (derived from co-expression of hCL withRAMPs; supplementary material, Fig. S2, lanes 4,7) was alsoreduced to a 37 kDa form by treatment with endoglycosidaseF, demonstrating that the additional mass units representcarbohydrate residues (supplementary material, Fig. S2, lanes

Jour

nal o

f Cel

l Sci

ence

912

5,8). The ~55 kDa form is resistant to endoglycosidase H,indicating that hCL has been terminally glycosylated, anevent normally associated with transit through the Golgicomplex and the production of mature glycoproteins(supplementary material, Fig. S2, lanes 6,9). The ~45 kDaform is sensitive to endoglycosidase H, indicating that it hasnot been terminally glycosylated. Thus, the antibody LN-1436 specifically recognises unglycosylated, core-glycosylated and mature transiently expressed andendogenously produced hCL.

We then used tissue microarray (TMA) technology(Kononen et al., 1998; Fox et al., 2004) to study hCLexpression and localisation in tissues. The receptor is widelyexpressed in normal human tissues, including endometrium(Fig. 2A), myometrium (Fig. 2B), endometrialadenocarcinoma (Fig. 2C), corpus luteum (Fig. 2D), cervix(Fig. 2E,F), kidney, lung and others (supplementary material,Fig. S3A-D). In the majority, it is localised in microvascularendothelium (Fig. 2; supplementary material, Fig. S3), asfurther demonstrated by double immunofluorescence on frozensections (supplementary material, Fig. S4). Furthermore, insome tissues, expression of hCL is not exclusive toendothelium in humans. Thus, hCL immunoreactivity wasfound in epithelium in cervix (Fig. 2F), kidney, adrenal glandand in some other organs (supplementary material, Fig. S3A-D, and summarised in supplementary material, Table S1). Ourdata support findings by others (Hagner et al., 2002),expanding the range of human organs investigated. Our dataalso demonstrate the heterogeneity of CL expression invascular beds in various organs, since standardised analysisusing our anti-hCL antibody and TMA was performed(summarised in supplementary material, Table S1).

We then asked whether microvascular endothelial cells(MVECs) isolated from various human organs (skin, lung,endometrium and myometrium) maintain expression ofglycosylated EndoCL in vitro (i.e. reflecting RAMP co-

Journal of Cell Science 119 (5)

expression and the functional state of the receptor). We alsoanalysed its expression in other cell types from the non-endothelial cell lineage (epithelial, and vascular and non-vascular smooth muscle cells) in vitro. Highly endothelial-specific expression of hCL was observed in vitro(supplementary material, Fig. S5A). All MVECs studiedexpressed core-glycosylated and mature EndoCL, although theamount varied (supplementary material, Fig. S5B).

Correlations with the presence of the terminallyglycosylated form of EndoCL

Correlation with RAMP mRNA expressionThe presence of the terminally glycosylated form of EndoCLin MVECs directly suggests the presence of translatedendogenous RAMP proteins, since CL could only beglycosylated and transported to the cell surface in the presenceof these accessory proteins (McLatchie et al., 1998). Weinvestigated which endogenous RAMPs are co-expressed withEndoCL in microvascular endothelium. Unfortunately, noanti-human RAMP antibodies that reliably distinguish RAMPisoforms (1, 2 and 3) or demonstrate expression ofendogenous monomers of the expected molecular weight(which should be observed in the case of a heterodimeric CL-RAMP receptor) are available. Therefore, we studied themRNA expression of RAMPs. We demonstrated that thepresence of the terminally glycosylated form of EndoCL in allMVECs studied correlates with the expression of RAMP2 andRAMP3 mRNAs only (supplementary material, Fig. S6).None of the MVECs studied expressed RAMP1 mRNA(supplementary material, Fig. S6). Since all studied MVECsdemonstrated expression of mature glycosylated EndoCL(supplementary material, Fig. S5) and a similar pattern ofRAMP expression (supplementary material, Fig. S6), wefocused our studies of the subcellular localisation andtrafficking of EndoCL on human dermal MVECs (hDMVECs)(Fig. 3).

Fig. 1. Transiently expressed andendogenous hCL species. The humanendothelial CL cDNA was cloned intothe pcDNA 3.1 expression vector andtermed hCLpcDNA. To examine thefunctionality, the resulting vector wastransiently cotransfected with RAMPsinto HEK293T cells and theintracellular cAMP concentrations weremeasured. The intact HEK293T cellslacked functional AM and CGRPreceptors, because they showed littlecAMP response to agonist stimulation.By contrast, in HEK293T cellscotransfected with hCL-RAMP, thereceptors were fully functional in intracellular cAMP production depending on the ligand specificity of the RAMPs (data not shown). (A) Totalcell or (B) tissue lysates were analysed by SDS-PAGE under reducing conditions, and immunoblots were probed using polyclonal anti-hCLantibody LN-1436. (A) The antibody specifically recognises hCL in HEK293T cells transfected with hCLpcDNA alone or together withRAMP1 or RAMP2 (CL+RAMP1 or CL+RAMP2; functional CGRP and AM receptors, respectively). HEK293T cells non-transfected(MOCK) or transfected with empty pcDNA vector (pcDNA) or RAMP1 or RAMP2 alone do not express hCL. The ~40-45 kDa hCL species(open diamonds; core-glycosylated receptor) are present in HEK293T cells transfected with hCLpcDNA only. The ~55 kDa hCL species (blackdiamonds; mature fully glycosylated receptor) is only produced when the receptor is co-expressed with RAMPs. (For details of deglycosylationexperiments confirming the origin of hCL species, see supplementary material, Fig. S2.) (B) The antibody also recognises endogenous hCLspecies expressed in tissues. Arrowheads, deglycosylated (~37 kDa) form of the receptor. The immunoblots are representative of twoindependent experiments. For loading controls, the membrane was re-probed with an antibody against �-actin.

Jour

nal o

f Cel

l Sci

ence


Correlation with cell-surfaceexpression

The presence of the terminally glycosylatedform of EndoCL suggests cell-surfaceexpression of this receptor in hDMVECs.We analysed subcellular localisationof EndoCL in hDMVECs byimmunofluorescence. Predominant surfacestaining was demonstrated by co-expression with the plasma membranemarker CD31 (Fig. 4A). IndividualhDMVECs also displayed a punctuatedstaining pattern widely distributed in thecytoplasm as well as concentrated in aperinuclear region (Fig. 4B-D). To identifythe intracellular compartments ofaccumulation of the endogenous receptor,immunofluorescence for markers ofcellular structures and different organellesin the endocytotic pathway was performed.Endogenous hCL protein did not showsignificant overlap with the Golgi complexmarker GM130 (Fig. 4B); but colocalisedwith calnexin, the endoplasmic reticulummarker (Fig. 4C). Intracellular EndoCL andmannose-6-phosphate receptor (M6PR),the late endosome marker, showed verysimilar distribution patterns with significantoverlap of both diffuse cytosolic andperinuclear vesicles (supplementarymaterial, Fig. S7E), although a few hCL-containing vesicles that lacked M6PR werealso observed. Similarly, endogenous hCLwas colocalised with lysosomes (Fig. 4D).

Fig. 2. Localisation of EndoCL in human tissues. Localisation of hCL was assessed by (A-F) immunohistochemistry on paraffin sections fromTMAs using primary LN-1436 and secondary alkaline phosphatase-conjugated antibodies, and detected with Vector Red (red colour). Cellnuclei were counterstained with haematoxylin (blue colour). Note the predominant CL expression in microvascular endothelium in (A)endometrium, (B) myometrium, (C) adenocarcinoma, (D) corpus luteum, and (E and F) cervical stroma (arrows); and in (F) cervical epithelium(arrowheads). (For hCL expression in other tissues, see supplementary material, Figs S3 and S4.)

Fig. 3. Expression of endogenous hCL species and RAMP mRNA in endothelial cells.(A) Expression of endogenous hCL species was analysed by immunoblotting. hDMVEClysates were treated with endoglycosidase F (F, lane 2), endoglycosidase H (H, lane 3) orvehicle (–, lane 1) before SDS-PAGE under reducing conditions and immunoblotting withpolyclonal anti-hCL antibody LN-1436. Arrowheads, deglycosylated (~37 kDa); opendiamonds, core-glycosylated (~45 kDa); black diamonds, mature fully glycosylated(~55 kDa) forms of the receptor. The ~55 kDa hCL species are reduced to a ~37 kDa hCLband after endoglycosidase F treatment, but are resistant to endoglycosidase H. Forloading controls, the membrane was reprobed with an antibody against �-actin. Theimmunoblot is representative of two independent experiments. (B) Expression of CL andRAMP mRNAs in primary hDMVECs (lane 1) was analysed by RT-PCR. The set ofprimers for detection of �-actin was used as a loading control. RNA sample from kidney(lane 2) served as a positive control. Numbers to the right indicate PCR fragment size.(For details of a full RT-PCR screen of all endothelial cell lines used in the present study,see supplementary material, Fig. S5.)

Jour

nal o

f Cel

l Sci

ence

914

By contrast, the receptor did not show marked overlapwith early sorting endosomes, indicated by EEA1(supplementary material, Fig. S7D).

Thus, although EndoCL was predominantlylocalised at the cell surface, a significant portion accumulatedin recycling endosomes, lysosomes and endoplasmic reticulum(Fig. 4; supplementary material, Fig. S7). This suggests thatendogenous endothelial-cell-expressed hCL might cycleconstitutively between the cell surface and recyclingendosomes. EndoCL expression was found at apical, basal andlateral cell surfaces in hDMVECs by confocal microscopy(Fig. 4E), demonstrating a lack of preferential localisation atthe cell surface.

Correlation with the presence of both functionalendogenous AM and CGRP receptors

We then investigated whether the expression of matureglycosylated EndoCL correlates with the presence offunctional endogenous AM or CGRP receptors inmicrovascular endothelium. The functional state of bothreceptor subtypes was investigated using cAMP, Aktphosphorylation and migration assays. We found that bothendogenous AM and CGRP receptors are fully active inhDMVECs in vitro, as demonstrated by the comparablemagnitude of increase in cAMP production (supplementarymaterial, Fig. S8) and migration rates (Fig. 5; supplementarymaterial, Fig. S9) in response to both ligands, and the inductionof Akt phosphorylation as described in more detail below.cAMP accumulation and migration were stimulated by bothpeptides (10 nM to 10 �M) (Fig. 5; supplementary material,Figs S8, S9). Akt phosphorylation was also induced by bothAM and CGRP as described in more detail below. Similar tohDMVECs, fully active endogenous AM and CGRP receptorswere found in human myometrial MVECs (hMMVECs) invitro, as demonstrated by the comparable magnitude ofagonist-mediated migration (data not shown).


AM and CGRP interact with endothelial EndoCL andinduce its desensitisation by different mechanismsThe presence of fully active receptors for both AM and CGRPsuggest that endogenous hDMVEC-expressed CL mightinteract with both peptides. We tested this hypothesis byinvestigating the dynamics of EndoCL internalisation inhDMVECs in response to AM or CGRP (10 nM to 1 �M).

Internalisation of EndoCLAgonist-mediated internalisation of endothelial cell-surface-expressed endogenous receptor was observed in response to AMbut not CGRP (Fig. 6). Time-course studies demonstrated theexistence of an early (within 5-15 minutes) and ligand-specificregulation of EndoCL internalisation by AM. A significant

Fig. 4. Subcellular localisation of EndoCL receptor inendothelial cells. (A-D) Intracellular distribution of EndoCLin hDMVECs was assessed by immunofluorescence usingLN-1436 antibody and markers of individual cellularstructures and organelles: (A) plasma membrane, (B) Golgi,(C) endoplasmic reticulum and (D) lysosomes (see Materialsand Methods). The appropriate FITC- (for the detection ofhCL; left, first image) or Texas Red- (for the detection ofsubcellular structures and organelles; centre, second image)conjugated secondary antibodies were used. DAPI was usedto counterstain cell nuclei. Colocalized structures (antigens;right, third image) appear in yellow (as indicated by yellowarrows) as determined by overlay of images. Non-colocalisedstructures appear in green (green arrows) and red (redarrows). Figures are representative of three independentexperiments. (For details of a full immunofluorescence screenof intracellular distribution of EndoCL, see supplementarymaterial, Fig. S7.) (E) Localisation of EndoCL at the cellsurface in hDMVECs. Reconstructed immunofluorescentconfocal pictures of EndoCL in hDMVECs. The en facepicture is the collapsed serial of the XY plane confocal imagealong the Z direction. At the top is the XZ cross-sectionpicture with the apical side facing up and the basal side facingdown. To the right of the en face picture is the YZ cross-section picture with apical side on the right and basal side onthe left.

Jour

nal o

f Cel

l Sci

ence


proportion of EndoCL was initially (15-30 minutes) targeted toearly sorting endosomes and then to lysosomes upon exposureto the 10 nM to 1 �M ligand (Fig. 6). Internalisation of thereceptor occurred in all hCL-expressing cells (supplementarymaterial, Fig. S10). Recycling of the internalised EndoCL inMVECs was not efficient, with a significant proportion ofendocytosed hCL being targeted to the degradative pathwaythrough lysosomes. The receptor reappeared on the cell surfaceonly two hours after the removal of agonist (Fig. 6). By contrast,

CGRP (10 nM to 1 �M) failed to demonstrate any changes insubcellular localisation of endogenous receptor (Fig. 6). Thesedata suggest that EndoCL interacts with AM and generatesendogenous AM receptor only. A similar specificity of EndoCLinternalisation in response to AM but not CGRP was observedin hMMVECs (data not shown).

Kuwasako et al. reported ligand-specific regulation of greenfluorescent protein (GFP)-tagged CL (CL-GFP) internalisationby RAMPs (Kuwasako et al., 2000). Thus, agonist-mediated

Fig. 5. Functional endogenous AM and CGRP receptors inendothelial cells. Human DMVECs were seeded into the upperchamber of Transwell inserts. Test peptide [human AM, CGRP andamylin (AMY)] was added to the lower chamber. After 24 hoursincubation, the cells were labelled with a fluorescent dye(calcein–AM). Only those labelled cells that have migrated throughthe pores of the FluoroBlok membrane can be detected. The numberof cells migrated to the lower surface of the insert was determined bythe measurement of fluorescence of invaded cells in a fluorescenceplate reader with bottom reading capabilities. Full microvascularendothelial growth medium (Full Medium) and VEGF were used aspositive controls. Each point represents the mean ± s.e.m. of fiveseparate experiments (*P<0.05; **P<0.01; ***P<0.001). (Forimages, see supplementary material, Fig. S9.)

Fig. 6. Dynamics of internalisation and recycling of EndoCL in response to AM and CGRP. Subcellular distribution of EndoCL in hDMVECsbefore and after exposure to ligand (100 nM; AM or CGRP) for 30 minute. Cells were processed for immunocytochemistry immediately afterexposure, or 2 hours after the complete removal of the agonist from the culture medium (see images labelled 120 mins). Colocalisation ofEndoCL with the endothelial cell surface (CD31; top row), early sorting endosome (EEA1; middle row) or lysosome (LAMP1; bottom row)markers (see Materials and Methods). Only merged images are presented. Colocalised structures (antigens; yellow arrows) appear in yellow asdetermined by overlay of the images. Non-colocalised structures appear in green (green arrows) and red (red arrows). The figures arerepresentative of three independent experiments.

Jour

nal o

f Cel

l Sci

ence

916

internalisation of CL-GFP was observed in RAMP1cotransfectants (by CGRP or AM), RAMP2 contransfectants(AM only) and RAMP3-contransfectants (AM only), whichoccurred with similar kinetics. These findings suggest that, inour study, endothelial EndoCL exists as a heterodimer withRAMP2 (AM1 receptor) or RAMP3 (AM2 receptor) (Fig. 6).This is supported by our findings on predominant expressionof RAMP2 and RAMP3 but not RAMP1 mRNA in these cells(Fig. 3). However, in the same study, Kuwasako et al. reportedthat only the CL-RAMP1 complex shows similar functionalresponses to stimulation by both AM and CGRP (Kuwasako etal., 2000). Furthermore, other groups reported that transfectedCL-RAMP3 complex (AM2 receptor) has appreciable affinityto CGRP (Poyner et al., 2002). Thus, it could not be excludedthat EndoCL also generates an endogenous CGRP receptor(either CL-RAMP1 or CL-RAMP3 complex) in endothelialcells. In this case, the contribution of endogenous RAMP1 tothe expression of EndoCL-associated CGRP receptor alsocould not be excluded because mRNA levels do not alwaysreflect the protein levels and because the presence of theefficient RAMP1 translational machinery in endothelial cellshas been previously demonstrated (Hay et al., 2003; Muff etal., 1998).

To test the hypothesis that EndoCL also generates anendogenous CGRP receptor, we investigated the dynamics ofendogenous AM and CGRP receptor desensitisation, and theeffect of AM and CGRP receptor antagonists on EndoCLinternalisation. Agonist-promoted internalisation of GPCRsis often related to the loss of receptor activity (Koenig et al.,1997; Krupnick and Benovic, 1998). Since EndoCL isinternalised after exposure to AM, we tested whether onlyendogenous AM or both endogenous AM and CGRPreceptors lose their activity (desensitise) under theseconditions. Similarly, we tested the activity of both receptorsubtypes after stimulation with CGRP (when EndoCL is notinternalised).


Desensitisation of endogenous AM and CGRPreceptors

Activation of Akt in endothelial cells plays an important rolein AM- and vascular endothelial growth factor (VEGF)-induced angiogenesis (Ilan et al., 1998; Kim et al., 2003). Weused an Akt phosphorylation assay to determine Akt activityand to investigate the dynamics of desensitisation ofendogenous AM and CGRP receptors.

Both AM (Fig. 7, lanes 2-4) and CGRP (Fig. 7, lanes 11-13) (100 nM) increased Akt phosphorylation at Ser473 asearly as 10 and 15 minutes, respectively, compared with thecontrol (Fig. 7, lane 1), and to a similar degree to VEGF (Fig.7, lane 10). Desensitisation assays revealed rapid (within 15minutes) and stable desensitisation of endogenous endothelialAM and CGRP receptors. Re-challenge with either agonist(AM, Fig. 7, lanes 5-8; or CGRP, Fig. 7, lanes 14-17) after30 minutes (with ligand depleted from the medium)demonstrated a lack of repeated Akt phosphorylation, andtherefore suggests a sustained receptor desensitisation.Furthermore, reciprocal re-challenge of receptors (AMreceptor with CGRP, Fig. 7, lane 9; and CGRP receptor withAM, Fig. 7, lane 18) revealed the loss of endogenous receptoractivity.

Both AM and CGRP interact with endothelial EndoCLSince both endogenous AM and CGRP receptors desensitisein response to either agonist, this suggests that both AM andCGRP interact with the same receptor in endothelial cells. Thatthis receptor is EndoCL was further supported by the use ofAM22-52 and CGRP8-37, which are putative antagonists of AMand CGRP receptors, respectively (Poyner et al., 2002). Bothtruncated peptides blocked AM-induced internalisation ofEndoCL in hDMVECs (Fig. 8). AM22-52 completely abolished,whereas CGRP8-37 attenuated, internalisation of the receptor,supporting the view that both AM and CGRP interact withendothelial EndoCL.

Fig. 7. Desensitisation of endogenousAM and CGRP receptors inendothelial cells. The dynamics ofdesensitisation of endogenous AMand CGRP receptors in hDMVECswas investigated using an Aktphosphorylation (Akt-P) assay todetermine Akt activity. In controlgroups, cells were exposed to 100 nMAM (AM, lanes 2-4) or CGRP(CGRP, lanes 11-13) for indicatedtimes. In pretreatment groups(Pretreatment with AM orPretreatment with CGRP; –/+indicate if pretreatment wasperformed), cells were exposed to100 nM AM (lanes 5-9) or CGRP(lanes 14-18) for 15 minutes, andthen to a ligand-depleted medium fora further 30 minutes. In pretreatmentgroups, cells were then re-challengedwith either the same agonist (100 nM) (lanes 5-8 for AM, and lanes 14-17 for, respectively) or reciprocal agonists (100 nM) (CGRP for AM-pretreated cells, lane 9; and AM for CGRP-pretreated cells, lane 18) to examine receptor desensitisation. VEGF-treated cells (10 ng/ml for 10minutes; lane 10) were used as a positive control.

Jour

nal o

f Cel

l Sci

ence


DiscussionIn the present study, we have described specific properties ofthe endogenous GPCR CL (EndoCL) expressed in humanendothelium. We have shown its ability to interact with bothAM and CGRP, and we have shown that interaction with eitherligand is sufficient to induce receptor desensitisation to bothAM and CGRP. Moreover, two different mechanisms areinvolved, and they seem to be equally important in regulatingthe function of endothelial EndoCL.

Endothelial cells present a unique model for studyingEndoCL without the need to create ideal reconstitutionsystems, since these cells ubiquitously express matureglycoprotein (RAMP-associated, functional receptor) in vivoand in vitro, as well as functional endogenous AM and CGRPreceptors. It has been hypothesised that the regulation of acellular CGRP to AM responsiveness might be mediated bychanges in RAMP expression, which would effectivelydesensitise a cell to CGRP and sensitise it to AM or vice versa(reviewed by Sexton, 2001). Recent studies have demonstratedthat RAMP mRNA expression is dynamically regulated inseveral models of pathophysiological challenge and/or drugtreatment (reviewed by Sexton, 2001); however, whether thisleads to a switch between receptor subtypes was notinvestigated. Our study suggests an alternative mechanism ofEndoCL regulation.

Our findings suggest that the mechanisms regulatingendothelial EndoCL function are based primarily on its abilityto interact with both agonists (AM and CGRP), as well asdifferential consequences of such interactions, and not on aswitch between AM and CGRP receptor subtypes (Fig. 9). Ourfindings that CL can generate AM or CGRP receptors parallelresults from other studies (McLatchie et al., 1998), but thefinding that endothelial EndoCL simultaneously acts as areceptor for both peptides (i.e. plays a role as an endogenous

AM/CGRP receptor) is new (Fig. 9). Furthermore, our studynot only supports previous reports suggesting that both AMand CGRP interact with the same receptor in endothelial cells(Shichiri et al., 1999), but also demonstrates for the first timethat this receptor is EndoCL.

It remains to be investigated which RAMP forms aheterodimer with endothelial EndoCL, because its propertiesdo not entirely match those observed for CL-RAMPcotransfectants (McLatchie et al., 1998; Kuwasako et al.,2000). For example, RAMP mRNA expression profile (lack ofRAMP1 mRNA) and internalisation studies suggest thatendothelial EndoCL should be exhibiting AM1 and/or AM2receptor phenotype. But because EndoCL also interacts withCGRP, CL-RAMP3 is likely to be a predominant receptorcomplex in endothelial cells, since appreciable affinity toCGRP has been demonstrated for this heterodimer intransfected cell models (Hay et al., 2005). Given the datasuggesting a potential role for G-protein coupling in theexpression of RAMP-induced phenotype (reviewed by Sextonet al., 2001), it is possible that other components of the cellularbackground could also be important in defining the observedphenotype and properties of endothelial EndoCL. This isbecause the CL-RAMP3 complex, in contrast to CL-RAMP1and to endothelial EndoCL, does not exhibit similar functionalresponses to AM and CGRP peptides (Kuwasako et al., 2000).

Nevertheless, irrespective of the RAMP isoform involved,fully glycosylated EndoCL glycoprotein (functional receptor)is produced in endothelial cells, as demonstrated using anti-hCL antibody (summarised in Fig. 9). Interaction with eitherligand is sufficient to induce its desensitisation to both AM andCGRP, but different mechanisms are involved. Thus, AM-induced complete desensitisation of endogenous AM/CGRPreceptors coincides with EndoCL internalisation and itstargeting to the degradation pathway with subsequent

Fig. 8. Interaction of AM and CGRP with endothelial EndoCL. Subcellular distribution of EndoCL in hDMVECs before treatment (Control)after exposure to AM (AM, 100 nM for 15 minutes) with or without antagonists (AM22-52 and CGRP8-37, 1 �M) was assessed byimmunofluorescence. Cells were processed for immunofluorescence immediately after exposure. Immunofluorescence was performed usinganti-hCL antibody LN-1436 and FITC-conjugated secondary antibody. DAPI was used to counterstain cell nuclei. Figures are representativemerged images of two independent experiments. Cell surface (green arrows) and internalised (green arrowheads) receptor portions areindicated.

Jour

nal o

f Cel

l Sci

ence

918

reappearance on the cell surface only 2 hoursafter an initial challenge. By contrast, CGRP-induced desensitisation of endogenousAM/CGRP receptors is associated withretention of EndoCL at the cell surface.

To our knowledge, there are no data onendogenous AM or CGRP receptordesensitisation in endothelial cells orunderlying mechanisms. Agonist-promotedinternalisation is common to a large numberof GPCRs and is often related to the loss ofreceptor activity (Koenig et al., 1997;Krupnick and Benovic, 1998). Thus, ourfindings that AM-induced EndoCLinternalisation results in desensitisation ofendogenous AM receptors parallel resultsfrom other studies, including those using CL-RAMP receptor system reconstitution models(Kuwasako et al., 2000; Hilairet et al., 2001b),but the finding that it also results in loss of cellsensitivity to CGRP is new. It is notable thatCGRP is also able to interact with endothelialEndoCL and to induce desensitisation of bothendogenous CGRP and AM receptors, thussupporting the view that both peptides actthrough the same receptor. However, themechanism of CGRP-induced desensitisationof EndoCL does not involve receptorinternalisation and therefore requires furtherinvestigation. For example, it might involveGPCR kinases (GRK). Seven GRKs havebeen identified to date, and they play a pivotalrole in desensitisation of GPCRs without lossof their cell-surface expression (Penn et al.,2000). Thus, a role for GRK6 in the regulationof CGRP-mediated desensitisation of porcineCL transiently expressed in HEK293T cellshas been suggested (Aiyar et al., 2000). Inendothelial cells, desensitisation of GPCRsmight involve GRK2 (Liu et al., 2005). Thus,the activated endothelin receptor (ETB), acanonical GPCR, recruits GRK2 that bindsdirectly to Akt and inhibits its phosphorylation in endothelialcells (Liu et al., 2005). We speculate that a similar mechanism


could also be involved in CGRP-induced desensitisation (lossof ability to induce Akt phosphorylation) of endothelial

Fig. 9. Properties of endothelial EndoCL and theassociated AM/CGRP receptor system.(I) Endothelial EndoCL is expressed on the cellsurface as a mature glycoprotein (N-linkedcarbohydrate residues are schematicallyrepresented as grey boxes). Both AM and CGRPinduce Akt phosphorylation (Akt-P).(IIA) Receptor is internalised and (III) targeted fordegradation upon activation by AM, but (IIB)remains at the cell surface after interaction withCGRP. Interaction with either AM (I-IIA-III) orCGRP (I-IIB) results in (IV) desensitisation of thereceptor to both peptides, irrespective of whichagonist was used for initial stimulation, and a lossof activity through distinct alternativemechanisms.

Jour

nal o

f Cel

l Sci

ence


EndoCL. However, irrespective of the mechanism involved inEndoCL desensitisation, our findings suggest that endogenousendothelial AM and CGRP receptors might lose their activitythrough sustained desensitisation after interaction with eitherligand, rather than switching between receptor subtypes/phenotypes, since EndoCL simultaneously acts as a receptorfor both peptides.

Desensitisation of EndoCL as a key endothelial AM/CGRPreceptor might have adverse effects on endothelial cell biologyin pathologies where levels of either peptide are elevated, suchas cardiovascular disease, neoplasia and inflammation (whereAM levels are elevated); and subarachnoid haemorrhage andmigrane (where CGRP levels are upregulated). This is becauseAM signalling is essential in endothelial cell biology, asdemonstrated in AM-‘depleted’ or -‘deficient’ endothelium invivo and in vitro (Shindo et al., 2001; Hippenstiel et al., 2002).Thus, lack of basal-cell-surface AM signalling throughEndoCL as a key endothelial AM/CGRP receptor couldcontribute to changes in endothelial cell morphology and theirdetachment from the basal membrane in vitelline vessels foundin AM-homozygote knockout mice (Shindo et al., 2001).Blockade or insufficient AM cross-talk between endothelialcells through juxtacrine loops mediated by laterally expressedEndoCL might contribute to hyperpermeability of theendothelial monolayer in response to stress stimuli, such aspolymorphonuclear-leukocyte-derived oxygen metabolites(Hippenstiel et al., 2002). Apically expressed EndoCL mightmediate endothelium-dependent AM-induced vasodilatation(Hinson et al., 2000; Zhang and Hintze, 2001). Similarly,EndoCL could be important in mediating CGRP-mediatedproliferation (Haegerstrand et al., 1990) and migration (presentstudy) of endothelial cells. However, further studies arerequired to dissect these mechanisms in order to understand therole of GPCR CL in transduction of signals from variousmembers of the calcitonin family of peptide hormones.

The unregulated desensitisation of endothelial GPCRs cancontribute to the aetiology of GPCR-based vascular diseases,since these receptors play important roles in vascular biologyin mediating signal transduction from survival and angiogenicstimuli (Liu et al., 2005; LeCouter et al., 2003; Nikitenko etal., 2003; Maguire and Davenport, 2005). For example, injury-induced alterations in GPCR-mediated signalling in sinusoidalendothelium result in endothelial dysfunction and leads tointrahepatic portal hypertension (Liu et al., 2005). In humanheart failure, desensitisation of GPCR-mediated beta-adrenergic signal transduction has been reported to be one ofthe main pathophysiological alterations (Schotten et al., 2000).Additionally, hypertension is accompanied by increased levelsof endothelin-1 (ET-1) and decreased arterial contraction inresponse to ET-1 owing to desensitisation of its cognate GPCR(ETA) (Thakali et al., 2004). Similarly, altered signallingthrough AM and CGRP receptors has been observed inconditions where ligand levels are upregulated but theunderlying mechanisms remain unclear. For example, inchronic heart failure, a cardiovascular disease where plasmalevels of AM are raised, the potent and long-lastingvasodilatory effect of AM on skeletal muscle arteries issignificantly attenuated, in part because of impaired productionof nitric oxide (NO) in the forearm resistance vessels (Kato etal., 1996; Nakamura et al., 1997). Our findings suggest thatthese effects could be a result of the EndoCL internalisation

and desensitisation in endothelial cells, since studies by othergroups have demonstrated a role for AM in the regulation ofblood pressure by stimulating NO production and have shownthat the NO pathway is essential for AM-induced endothelium-dependent vasodilatation in particular (Shindo et al., 2001;Zhang and Hintze, 2001).

The precise mechanisms underlying the response ofendogenous AM and CGRP receptors to constitutive/chronicagonist stimulation in conditions such as cardiovasculardisease, neoplasia and inflammation are yet to be investigated.Detailed knowledge of mechanisms regulating the function(sensitisation/desensitisation) of EndoCL in endotheliummight enable better understanding of alterations in vascularresponses in these conditions. Therefore, the molecularmechanisms of trafficking, function and desensitisation ofendogenous GPCR CL, as a key AM/CGRP receptor, requirefurther investigation. Detailed knowledge of these mechanismsmight provide insights for the development of novelcompounds and antibodies that can modulate the activity ofendogenous AM and CGRP receptors. Such drugs might havepotential in cardiovascular disease and in cancer (Doods et al.,2000; Martinez et al., 2004).

In summary, our data demonstrate for the first time that thefunction of EndoCL is tightly regulated in endothelial cells.The finding that endogenous endothelial CL can interact withboth AM and CGRP, as well as that either of these interactionscan result in complete desensitisation of the receptor, suggeststhat the mechanisms regulating expression and function of thisGPCR in endothelium have the potential to lead to translationaladvances in vascular therapeutic methods and techniques forconditions where levels of AM or CGRP are elevated.

Materials and MethodsConstruction of human CL cDNA vector and expression offunctional receptorThe full open reading frame (ORF) of the reported human (h)CL, modified toprovide a consensus Kozak sequence (Aiyar et al., 1996), was cloned intopcDNA3.1+ vector (Invitrogen) at its multicloning site, using RNA isolated fromMVECs as a template (Nikitenko et al., 2003). 1 �g of total RNA was reverse-transcribed (RT) with an oligo (dT) adaptor primer to generate the first-strand DNAusing the Reverse-iTcDNA synthesis Kit (Advanced Biotechnologies), as previouslydescribed (Nikitenko et al., 2001; Nikitenko et al., 2003). The full ORF wasamplified using appropriate primers containing XbaI and KpnI restriction sites atthe 5� and 3� ends, respectively. The 5� and 3� adaptor sequence ends of hCL wereligated into the XbaI and KpnI sites of pcDNA 3.1+ (Invitrogen Life Technologies)at its multicloning site. The fused product was termed hCLpcDNA and sequencedusing an Applied Biosystems 377 Genetic analyser. Myc-RAMP1 and Myc-RAMP2cloned into pcDNA3.1 were a gift from D. Poyner (Aston University, Birmingham,UK), and were originally prepared by Dr S. Foord and co-workers(GlaxoSmithKline, UK).

1 �g DNA per plasmid was transfected for each T 75 cm2 flask containing 50-60% confluent HEK293T cells. Cells were transfected with hCLpcDNA alone orcotransfected with RAMP1 and RAMP2 cDNAs. Cells were left for 24 hours beforeeither plating into 96-well plates for intracellular cAMP assay (as described below,and to validate expression of functional specific AM and CGRP receptor subtypesupon hCL and RAMP cotransfection in HEK293T cells). Protein lysates fromHEK293T transfectants expressing functional CL-RAMP heterodimers were usedin the present study for immunoblotting analysis solely for the purpose of validationof specificity of our own polyclonal antibody (see below).

Antibody production and characterisationRabbit polyclonal antibody LN-1436 was raised against synthetic peptidecorresponding to residues 427-461 (HDIENVLLKPENLYN) at the extreme C-terminus of hCL protein (Accession numbers AAC41994 and AAA62158) (SigmaGenosys). The peptide antigen was selected based on high sequence specificity(CLASTP analysis; http://www.ncbi.nlm.nih.gov). The hCL peptide was conjugatedto keyhole limpet haemocyanin using 1-ethyl-3-(3-dimethylamino-propyl)carbomide hydrochloride before being used for immunisation of NewZealand White rabbits. The titres of the resulting antisera were accessed by enzyme-

Jour

nal o

f Cel

l Sci

ence

920 Journal of Cell Science 119 (5)

linked immunosorbent assay (ELISA) using synthetic peptide. Finally, antibody wascharacterised using ELISA and immunoblot analysis of transiently expressed(transfected with hCLpcDNA) HEK293T cells or endogenous MVEC-expressedhCL as indicated below.

Peptides and growth factorsSynthetic human AM, �-CGRP, AM22-52, CGRP8-37 and amylin were from Bachem.Human VEGF was from R&D Systems.

Cell culturePrimary MVEC linesPrimary human dermal, lung, endometrial and myometrial MVECs and non-endothelial cell lines were obtained and cultured as previously described (Nikitenkoet al., 2000; Nikitenko et al., 2003). Cells were seeded onto plastic culture dishes(Falcon; �5�105 cells/cm2) precoated with Attachment Factor (TCS Cellworks)according to the manufacturer’s instructions, and supplemented with EGM-2MVBullet Kit medium (Bio Whittaker). Cultures were incubated at 37°C in a 5% CO2

humidified atmosphere and the medium was replaced every 1-2 days. Cells werepassaged 1:3 at confluence by release with trypsin/EDTA (Sigma).

Non-endothelial cell linesSeveral cell lines were chosen to serve as controls by representing cells from thenon-endothelial cell lineage. These included epithelial, vascular and non-vascularsmooth muscle cells. Primary myometrial smooth muscle cells were isolated bycollagenase digestion from uteri obtained at hysterectomies in the NuffieldDepartment of Obstetrics and Gynaecology, John Radcliffe Hospital, Oxford. MDA231 (a breast epithelial carcinoma cell line) and HEK293T (human renal epithelialcells) were obtained from the American Tissue Culture Collection (ATCC) (Tokeset al., 1981; Wierland et al., 1996). Primary coronary artery vascular smooth musclecells (VSMCs) were obtained from Cambrex and maintained according to thesupplier’s protocols. All other non-endothelial cell lines were maintained in DulbeccoModified Eagles Medium (DMEM) supplemented with 10% fetal calf serum (FCS),maintained at 37°C and 95% humidity, and passaged 1:3 at confluence.

Cell migration assaysPrimary MVECs were used to study the effect of human AM, CGRP and amylinon migration. Human dermal MVECs were seeded into the upper chamber ofTranswell inserts (BD Biosciences). Test peptides (human AM, CGRP and amylin)were added to the lower chamber. Full microvascular endothelial growth mediumEGM-2MV and vascular endothelial growth factor were used as a positive control.After 24 hours incubation, the cells were labelled with a fluorescent dye (calcein–AM;Molecular Probes). The number of cells migrated to the lower surface of the insertwas determined by measurement of fluorescence of invaded cells in a fluorescenceplate reader Elx 800 with bottom reading capabilities at excitation/emissionwavelengths of 485/530 nm. Data were analysed with KC4v30 PowerReportsSoftware and photographs taken with a Nikon CoolPix 990 camera. Only thoselabelled cells that have migrated through the pores of the FluoroBlok membranecan be detected.

Deglycosylation experimentsGlycosidases F and H were used to differentiate between N-linked glycoproteins(McLatchie et al., 1998). Cell lysates were treated with glycosidases F (whichdeglycosylates both the mature and core-glycosylated forms of the receptor) and H(which deglycosylates only the core-glycosylated form) according to the supplier’sprotocols (Roche) before SDS-PAGE and immunoblotting with the antibody LN-1436.

SDS-PAGE and immunoblottingProtein lysates from cell lines and tissues were subjected to SDS-PAGE andimmunoblotting as previously described (Nikitenko et al., 2001). In brief, uponreaching confluence, cells were collected into the lysis buffer and homogenised. Celllysates were subjected to SDS-PAGE and immunoblotting with primary anti-hCLpolyclonal antibody LN-1436, followed by incubation with goat anti-rabbithorseradish peroxidase (HRP)-conjugated secondary antibody (Dako). The signalwas developed using enhanced chemiluminescence ECL Plus (Amersham-Pharmacia) or Supersignal (Pierce & Warriner) systems. The membranes were thenreprobed with anti-human �-actin antibody (Clone AC-15; Abcam) to monitor andconfirm equal loading of total protein in samples. Anti-human CD31 antibody wasused to demonstrate the purity of endothelial cell lines. The density of bands wasanalysed on an Alpha ImagerTM 1220 Documentation & Analysis System v.5.5using linked background subtraction.

Intracellular cAMP assaysFor cAMP determinations, the transfected HEK293T cells were cultured for afurther day, washed with PBS and preincubated in DMEM medium containing 300�M IBMX (Sigma) for 30 minutes at 37°C. Forskolin (300 �M), CGRP or AMwere added for a further 15 minutes.

MVECs at 3rd-5th passage were seeded into the 96-well microtitre plates (tissue-

culture grade) precoated with Attachment Factor (TCS Cellworks) with cellconcentrations of 8000 per well and grown for three days in EGM-2MV mediumat 37°C in a 5% CO2 humidified atmosphere and the medium was replaced daily.Cells were supplemented with medium containing 0.5% FCS (0.5%FCS-EGM-2)for 16 hours and then pretreated with the same medium containingphosphodiesterase inhibitor IBMX (300 �M; Sigma) for 30 minutes at 37°C. Theeffect of forskolin (300 �M) or human AM, CGRP, amylin, VEGF (at a range ofconcentrations from 1 pM to 1 �M) (four wells for each dilution), or full medium(EGM-2-MV) on cAMP formation was examined. Plates were incubated for 15minutes at 37°C.

After exposure to stimuli, and to terminate the reaction, HEK293T cells andMVECs were washed with ice-cold PBS. Intracellular cAMP measurement wasperformed using the non-acetylation EIA assay (Amersham Pharmacia Biotech)according to the manufacturer’s instructions and essentially as described previously(Nikitenko et al., 2000).

FACS analysisDermal MVECs were grown as described, collected by incubation with accutase(PAA Laboratories) for 10-20 minutes, fixed in 4% paraformaldehyde, then pelletedand washed three times in PGB Buffer (PBS, 20 mM glucose, 0.1% BSA). The cellswere then resuspended in PGN Buffer (PBS, 20 mM glucose, 5% normal goatserum) and preincubated for 20 minutes, followed by the incubation for 25 minuteswith the primary polyclonal anti-hCL LN-1436 antibody diluted 1:1000 in PGN.Following three further washes in PGB Buffer, cells were incubated with thesecondary antibody (goat anti-rabbit FITC-conjugated IgG; diluted 1:100) for 30minutes in the dark. Finally, cells were washed in PGB Buffer and resuspended inPBS, 2% fetal calf serum, 0.1% sodium azide. For permeabilisation, cells wereincubated with rotation in PBS, 0.1% BSA, 0.1% saponin for 20 minutes at 4°C,prior to labelling with primary antibodies. FACS analysis was performed on aBeckman Coulter EPICS Altra Elite (Coulter) using 488 nm line of the UV laser.Lasers were aligned using Flow Check Beads (Coulter) prior to experiment.Photomultiplier tubes (PMTs) collected fluorescent light at 525 nm (FITC). 20,000cells were sorted in each experiment.

RNA isolation and RT-PCRRNA isolation and reverse-transcription polymerase chain reaction (RT-PCR) wereperformed as previously described (Nikitenko et al., 2001; Nikitenko et al., 2003).Amplifications were routinely performed using 25-30 cycles in the Perkin ElmerGene Amp PCR System 2400 for �-actin control, AM, CL and RAMPs usingprimers designed for their specificity and spanning neighbouring exons to enablethe detection of any genomic DNA contamination (supplementary material, TableS2).

Immunohistochemistry and immunofluorescenceCultured MVECs and cryostat sections were prepared and processed forimmunohistochemistry and immunofluorescence as previously described (Nikitenkoet al., 2000; Nikitenko et al., 2001).

Cultured cellsMVECs were grown to confluence on attachment-factor-precoated 8-well Permanoxslides (Nunc). Prior to immunohistochemistry, cultures were fixed in ice-coldacetone/methanol for 3-5 minutes and air dried for a minimum of 2 hours. Slideswere stored at –20°C prior to staining. In parallel experiments, cells also were fixedin 4% paraformaldehyde, washed and stored in PBS at 4°C.

Cryostat sectionsEndometrial tissues were snap frozen in liquid nitrogen and stored at –70°C. Frozensections 8-10 �m were cut on a Leica Kryostat 1720 Digital cryostat, allowed todry at room temperature and then fixed for 5 minutes in 4% ice-coldparaformaldehyde solution in PBS. Sections were then rinsed three times for 2minutes in PBS, dehydrated sequentially in 70% and absolute ethanol and stored at4°C in absolute ethanol (Nikitenko et al., 2001). Before immunohistochemistry,frozen sections were dried, washed for 5 minutes in distilled water and rinsed inBuffer 1 (100 mM Tris-HCl, 150 mM NaCl pH 7.5) for 5 minutes.

Multiple tissue microarrayFormalin-fixed, paraffin-embedded specimens (n=74) of 20 normal human tissueswere selected from archival files of the Department of Cellular Pathology, JohnRadcliffe Hospital, University of Oxford, Oxford, UK. Multiple tissue microarrays(TMAs) were produced by acquiring cylindrical cores (1.0 mm diameter) for eachspecimen arrayed at high density into a recipient TMA block (Kononen et al., 1998;Fox et al., 2004). This approach allowed immunohistochemical analysis of CLexpression in all tissue samples under standardised conditions. Consecutive 4 �msections were cut and dewaxed in two changes of xylene, rehydrated sequentiallyin absolute, 95% and 70% ethanol and distilled water, and rinsed in Buffer 1 priorto immunohistochemistry. The antigen-retrieval procedure (microwaving in citratebuffer) was carried out for anti-hCL polyclonal antibody LN-1436. Consecutivesections were used for controls or for analysis with other antibodies as indicatedbelow.

Jour

nal o

f Cel

l Sci

ence


AntibodiesIdentification of individual cell types and intracellular structures was performedaccording to the staining pattern with primary monoclonal and polyclonal antibodiesas indicated below.

Immunocytophenotyping was performed on frozen sections or adjacent paraffinsections according to the staining pattern with monoclonal anti-CD31 (endothelialcell marker), anti-smooth muscle actin (SMA; smooth muscle cell marker), anti-CD45 (peripheral blood mononuclear cell marker) and anti-CD68 (macrophagemarker) antibodies, and polyclonal antibody against von Willebrand Factor (vWF;endothelial cell marker) (all from Dako).

Intracellular structures were identified using monoclonal anti-GM130 (Golgimatrix protein of 130 kDa), anti-EEA1 (early endosome associated antigen 1), anti-calnexin (endoplasmic reticulum marker), anti-CD107a (LAMP-1; lysosome-associated membrane protein 1) antibodies (all from BD Biosciences Pharmingen);anti-CD71 (late endosome marker; transferrin receptor; Sigma), anti-mannose-6-phosphate receptor (late endosome marker; Abcam) and anti-CD31 (cell surfacemarker; Dako) antibodies.

Secondary antibodies were Texas Red- (TR) and fluorescein isothyocyanite(FITC)-conjugated horse anti-mouse and horse anti-rabbit IgG, biotinylated horseanti-mouse, anti-rabbit and anti-goat antibodies (all Vector Labs), and HRP-conjugated swine anti-rabbit antibody (Dako).

Immunohistochemistry and immunofluorescenceImmunohistochemistry and immunocytochemistry were performed as previouslydescribed (Nikitenko et al., 2000; Nikitenko et al., 2001). For hCL detection in vivoon TMA, a two-stage anti-rabbit HRP with the EnVision (Dako) system, which isnot affected by the endogenous biotin that is present in some human tissues, wasused. All immunohistochemistry or immunocytochemistry reactions were followedby final washes and detection. When biotinylated secondary antibodies were used,final washes were followed by application of streptavidin-alkaline phosphatasecomplex using Vectastain ABC-AP Kit (Vector). For slides processed with theEnVision system, HRP visualisation was performed using diaminobenzidine (DAB)(Vector). Frozen sections and cultures immunostained with the use of fluorescentmarkers were embedded in Vectashield mounting medium containing 4�,6-diamidino-2-phenylindole (DAPI) (Vector) to detect nuclei. Controls includedomission of primary antibodies, substitution of primary antibody with isotype-matched IgG or preimmune serum from corresponding species (Vector), all used atappropriate concentrations. These served as negative controls to indicate thespecificity of the antibodies used. Visualisation was performed with the use of aLeitz Diaplan or Leitz DMRBE microscopes (Leica), a Hamamatsu Orca C4742-95digital camera and Openlab software (both Improvision).

Desensitisation and internalisation assaysDesensitisation assayHuman dermal MVECs were supplemented with 0.5%FCS-EGM-2 medium for 16hours. Cells were then treated with or without 100 nM AM or CGRP in 0.5%FCS-EGM-2 medium for 15 minutes. After agonist exposure, cells were washed twicein 0.5%FCS-EGM-2, left for 30 minutes and then re-exposed to 100 nM AM orCGRP in 0.5%FCS-EGM-2 for indicated time periods (up to 15 minutes).Desensitisation of endogenous AM and CGRP receptors was accessed using Aktphosphorylation, since activation of Akt in endothelial cells plays an important rolein AM- and VEGF-induced angiogenesis (Ilan et al., 1998; Kim et al., 2003). Aktphosphorylation levels were analysed by immunoblotting using anti-Akt (SantaCruz) and anti-phospho-specific Akt (Ser473; Cell Signalling Technology)antibodies. VEGF (10 ng/ml) in 0.5%FCS-EGM-2 medium was used as a positivecontrol. After stimulation with AM, CGRP or VEGF, cells were processed for SDS-PAGE and immunoblotting as described above.

Internalisation assayHuman dermal MVECs were supplemented with 0.5%FCS-EGM-2 medium for 16hours. Cells were then treated with or without 100 nM AM in the presence orabsence of antagonists (AM22-52 or CGRP8-37) in 0.5%FCS-EGM-2 medium for 15minutes. Internalisation was accessed by immunofluorescence as described aboveusing anti-hCL antibody and cell-surface, early endosome and lysosome markers.

Statistical methodsData were analysed by Student’s t test to determine the significance of differencesbetween the control and means corresponding to each sample. Values P<0.05 wereconsidered significant.

Ethical approvalUse of hysterectomy samples was approved by the Oxford Clinical Research EthicsCommittee (COO.147).

We are grateful to Robin Roberts-Gant for assistance withformatting figures; Veronica Carroll and Ulrike Knies-Bamforth fortheir critical reading of the manuscript; and Jelena Jovanoviç,

Alexandre Akoulitchev and Johann Riedemann for very stimulatingdiscussions. We also thank Kingsley Micklem, Angela Borzychowskiand Celine Jones for their technical advice. This work was supportedin part by The Wellcome Trust (L.L.N. and M.C.P.R.), CancerResearch UK (R.B.), Medical Research Fund (University of Oxford,UK) (L.L.N.) and Royal Society UK (L.L.N., travel award). N.B. isa Final Honour School student from Queen’s College, Oxford.

ReferencesAiyar, N., Rand, K., Elshourbagy, N. A., Zeng, Z., Adamou, J. E., Bergsma, D. J. and

Li, Y. (1996). A cDNA encoding the calcitonin gene-related peptide type 1 receptor.J. Biol. Chem. 271, 11325-11329.

Aiyar, N., Disa, J., Dang, K., Pronin, A. N., Benovic, J. L. and Nambi, P. (2000).Involvement of G protein-coupled receptor kinase-6 in desensitization of CGRPreceptors. Eur. J. Pharmacol. 403, 1-7.

Bomberger, J. M., Spielman, W. S., Hall, C. S., Weinmann, E. J. andParamesarwaran, N. (2005). Receptor activity-modifying protein (RAMP) isoform-specific regulation of adrenomedullin receptor trafficking by NHERF-1. J. Biol. Chem.280, 23926-23935.

Brain, S. D. and Cambridge, H. (1996). Calcitonin gene-related peptide: vasoactiveeffects and potential therapeutic role. Gen. Pharmacol. 27, 607-611.

Brain, S. D. and Grant, A. D. (2004). Vascular actions of calcitonin gene-related peptideand adrenomedullin. Physiol. Rev. 84, 903-934.

Brell, B., Temmesfeld-Wollbruck, B., Altzschner, I., Frisch, E., Schmeck, B., Hocke,A. C., Suttorp, N. and Hippenstiel, S. (2005). Adrenomedullin reducesStaphylococcus aureus alpha-toxin-induced rat ileum microcirculatory damage. Crit.Care Med. 33, 819-826.

Doods, H., Hallermayer, G., Wu, D., Entzeroth, M., Rudolf, K., Engel, W. andEberlein, W. (2000). Pharmacological profile of BIBN4096BS, the first selective smallmolecule CGRP antagonist. Br. J. Pharmacol. 129, 420-423.

Evans, B. N., Rosenblatt, M. I., Mnayer, L. O., Oliver, K. A. and Dickerson, I. M.(2000). CGRP-RCP, a novel protein required for signal transduction at calcitonin gene-related peptide and adrenomedullin receptors. J. Biol. Chem. 275, 31438-31443.

Fox, S. B., Bragança, J., Turley, H., Campo, L., Han, C., Gatter, K. C., Bhattacharya,S. and Harris, A. L. (2004). CITED4 inhibits hypoxia-activated transcription in cancercells, and its cytoplasmic location in breast cancer is associated with elevatedexpression of tumor cell hypoxia-inducible factor 1alpha. Cancer Res. 64, 6075-6081.

Haegerstrand, A., Dalsgaard, C. J., Jonzon, B., Larsson, O. and Nilsson, J. (1990).Calcitonin gene-related peptide stimulates proliferation of human endothelial cells.Proc. Natl. Acad. Sci. USA 87, 3299-3303.

Hagner, S., Stahl, U., Knoblauch, B., McGregor, G. P. and Lang, R. E. (2002).Calcitonin receptor-like receptor: identification and distribution in human peripheraltissues. Cell Tissue Res. 310, 41-50.

Hay, D. L., Poyner, D. R. and Smith, D. M. (2003). Desensitisation of adrenomedullinand CGRP receptors. Regul. Pept. 112, 139-145.

Hay, D. L., Conner, A. C., Howitt, S. G., Takhshid, M. A., Simms, J., Mahmoud, K.and Poyner, D. R. (2004). The pharmacology of CGRP-responsive receptors incultured and transfected cells. Peptides 25, 2019-2026.

Hay, D. L., Christopoulos, G., Christopoulos, A., Poyner, D. R. and Sexton, P. M.(2005). Pharmacological discrimination of calcitonin receptor: receptor activity-modifying protein complexes. Mol. Pharmacol. 67, 1655-1665.

Hilairet, S., Foord, S. M., Marshall, F. H. and Bouvier, M. (2001a). Protein-proteininteraction and not glycosylation determines the binding selectivity of heterodimersbetween the calcitonin receptor-like receptor and the receptor activity-modifyingproteins. J. Biol. Chem. 276, 29575-29581.

Hilairet, S., Belanger, C., Bertrand, J., Laperriere, A., Foord, S. M. and Bouvier, M.(2001b). Agonist-promoted internalization of a ternary complex between calcitoninreceptor-like receptor, receptor activity-modifying protein 1 (RAMP1), and beta-arrestin. J. Biol. Chem. 276, 42182-42190.

Hinson, J. P., Kapas, S. and Smith, D. M. (2000). Adrenomedullin, a multifunctionalregulatory peptide. Endocr. Rev. 21, 138-167.

Hippenstiel, S., Witzenrath, M., Schmeck, B., Hocke, A., Krisp, M., Krull, M.,Seybold, J., Seeger, W., Rascher, W., Schutte, H. et al. (2002). Adrenomedullinreduces endothelial hyperpermeability. Circ. Res. 91, 618-625.

Ilan, N., Mahooti, S. and Madri, J. A. (1998). Distinct signal transduction pathways areutilized during the tube formation and survival phases of in vitro angiogenesis. J. CellSci. 111, 3621-3631.

Kamitani, S., Asakawa, M., Shimekake, Y., Kuwasako, K., Nakahara, K. and Sakata,T. (1999). The RAMP2/CRLR complex is a functional adrenomedullin receptor inhuman endothelial and vascular smooth muscle cells. FEBS Lett. 448, 111-114.

Kato, H., Shichiri, M., Marumo, F. and Hirata, Y. (1997). Adrenomedullin as anautocrine/paracrine apoptosis survival factor for rat endothelial cells. Endocrinology138, 2615-2620.

Kawai, J., Ando, K., Tojo, A., Shimosawa, T., Takahashi, K., Onozato, M. L., Yamasaki,M., Ogita, T., Nakaoka, T. and Fujita, T. (2004). Endogenous adrenomedullin protectsagainst vascular response to injury in mice. Circulation 109, 1147-1153.

Kim, W., Moon, S. O., Sung, M. J., Kim, S. H., Lee, S., So, J. N. and Park, S. K.(2003). Angiogenic role of adrenomedullin through activation of Akt, mitogen-activated protein kinase, and focal adhesion kinase in endothelial cells. FASEB J. 17,1937-1939.

Koenig, J. A. and Edwardson, J. M. (1997). Endocytosis and recycling of G protein-coupled receptors. Trends Pharmacol. Sci. 18, 276-287.

Jour

nal o

f Cel

l Sci

ence

922 Journal of Cell Science 119 (5)

Kononen, J., Bubendorf, L., Kallioniemi, A., Barlund, M., Schraml, P., Leighton, S.,Torhorst, J., Mihatsch, M. J., Sauter, G. and Kallioniemi, O. P. (1998). Tissuemicroarrays for high-throughput molecular profiling of tumor specimens. Nat. Med. 4,844-847.

Krupnick, J. G. and Benovic, J. L. (1998). The role of receptor kinases and arrestinsin G protein-coupled receptor regulation. Annu. Rev. Pharmacol. Toxicol. 38, 289-319.

Kuwasako, K., Shimekake, Y., Masuda, M., Nakahara, K., Yoshida, T., Kitaura, M.,Kitamura, K., Eto, T. and Sakata, T. (2000). Visualization of the calcitonin receptor-like receptor and its receptor activity-modifying proteins during internalization andrecycling. J. Biol. Chem. 275, 29602-29609.

Kuwasako, K., Cao, Y. N., Nagoshi, Y., Kitamura, K. and Eto, T. (2004).Adrenomedullin receptors: pharmacological features and possible pathophysiologicalroles. Peptides 25, 2003-2012.

LeCouter, J. and Ferrara, N. (2003). EG-VEGF and Bv8. a novel family of tissue-selective mediators of angiogenesis, endothelial phenotype, and function. TrendsCardiovasc. Med. 13, 276-282.

Liu, S., Premont, R. T., Kontos, C. D., Zhu, S. and Rockey, D. C. (2005). A crucialrole for GRK2 in regulation of endothelial cell nitric oxide synthase function in portalhypertension. Nat. Med. 11, 952-958.

Maguire, J. J. and Davenport, A. P. (2005). Regulation of vascular reactivity byestablished and emerging GPCRs. Trends Pharmacol. Sci. 26, 448-454.

Martinez, A., Vos, M., Guedez, L., Kaur, G., Chen, Z., Garayoa, M., Pio, R., Moody,T., Stetler-Stevenson, W. G., Kleinman, H. K. et al. (2002). The effects ofadrenomedullin overexpression in breast tumor cells. J. Natl. Cancer Inst. 94, 1226-1237.

Martinez, A., Julian, M., Bregonzio, C., Notari, L., Moody, T. W. and Cuttitta, F.(2004). Identification of vasoactive nonpeptidic positive and negative modulators ofadrenomedullin using a neutralizing antibody-based screening strategy. Endocrinology145, 3858-3865.

Matsui, H., Shimosawa, T., Itakura, K., Guanqun, X., Ando, K. and Fujita, T. (2004).Adrenomedullin can protect against pulmonary vascular remodeling induced byhypoxia. Circulation 109, 2246-2251.

McLatchie, L. M., Fraser, N. J., Main, M. J., Wise, A., Brown, J., Thompson, N.,Solari, R., Lee, M. G. and Foord, S. M. (1998). RAMPs regulate the transport andligand specificity of the calcitonin-receptor-like receptor. Nature 393, 333-339.

Muff, R., Leuthauser, K., Buhlmann, N., Foord, S. M., Fischer, J. A. and Born, W.(1998). Receptor activity modifying proteins regulate the activity of a calcitonin gene-related peptide receptor in rabbit aortic endothelial cells. FEBS Lett. 441, 366-368.

Nakamura, M., Yoshida, H., Makita, S., Arakawa, N., Niinuma, H. and Hiramori,K. (1997). Potent and long-lasting vasodilatory effects of adrenomedullin in humans.Comparisons between normal subjects and patients with chronic heart failure.Circulation 95, 1214-1221.

Nikitenko, L. L., MacKenzie, I. Z., Rees, M. C. P. and Bicknell, R. (2000).Adrenomedullin is an autocrine regulator of endothelial growth in humanendometrium. Mol. Hum. Reprod. 6, 811-819.

Nikitenko, L. L., Brown, N. S., Smith, D. M., MacKenzie, I. Z., Bicknell, R. and Rees,M. C. P. (2001). Differential and cell-specific expression of calcitonin receptor-likereceptor and receptor activity modifying proteins in the human uterus. Mol. Hum.Reprod. 7, 655-664.

Nikitenko, L. L., Smith, D. M., Bicknell, R. and Rees, M. C. P. (2003). Transcriptional

regulation of the CRLR gene in human microvascular endothelial cells by hypoxia.FASEB J. 17, 1499-1501.

Njuki, F., Nicholl, C. G., Howard, A., Mak, J. C., Barnes, P. J., Girgis, S. I. and Legon,S. (1993). A new calcitonin-receptor-like sequence in rat pulmonary blood vessels.Clin. Sci. 85, 385-388.

Oehler, M. K., Hague, S., Rees, M. C. and Bicknell, R. (2002). Adrenomedullinpromotes formation of xenografted endometrial tumors by stimulation of autocrinegrowth and angiogenesis. Oncogene 21, 2815-2821.

Penn, R. B., Pronin, A. N. and Benovic, J. L. (2000). Regulation of G-protein coupledrece3ptor kinases. Trends Cardiovasc. Med. 10, 81-89.

Poyner, D. R., Sexton, P. M., Marshall, I., Smith, D. M., Quirion, R., Born, W., Muff,R., Fischer, J. A. and Foord, S. M. (2002). International Union of Pharmacology.XXXII. The mammalian calcitonin gene-related peptides, adrenomedullin, amylin, andcalcitonin receptors. Pharmacol. Rev. 54, 233-246.

Roh, J., Chang, C. L., Bhalla, A., Klein, C. and Hsu, S. Y. (2004). Intermedin is acalcitonin/calcitonin gene-related peptide family peptide acting through the calcitoninreceptor-like receptor/receptor activity-modifying protein receptor complexes. J. Biol.Chem. 279, 7264-7274.

Satoh, F., Takahashi, K., Murakami, O., Totsune, K., Sone, M., Ohneda, M., Abe,K., Miura, Y., Hayashi, Y., Sasano, H. et al. (1995). Adrenomedullin in human brain,adrenal glands and tumor tissues of pheochromocytoma, ganglioneuroblastoma andneuroblastoma. J. Clin. Endocrinol. Metab. 80, 1750-1752.

Schotten, U., Filzmaier, K., Borghardt, B., Kulka, S., Schoendube, F., Schumacher,C. and Hanrath, P. (2000). Changes of beta-adrenergic signaling in compensatedhuman cardiac hypertrophy depend on the underlying disease. Am. J. Physiol. HeartCirc. Physiol. 278, H2076-H2083.

Sexton, P. M., Albiston, A., Morfis, M. and Tilakaratne, M. (2001). Receptor activitymodifying proteins. Cell. Signal. 13, 73-83.

Shichiri, M., Kato, H., Doi, M., Marumo, F. and Hirata, Y. Induction of max byadrenomedullin and calcitonin gene-regulated peptide antagonizes endothelialapoptosis. Mol. Endocrinol. 13, 1353-1363.

Shindo, T., Kurihara, H., Maemura, K., Kurihara, Y., Kuwaki, T., Izumida, T.,Minamino, N., Ju, K. H., Morita, H., Oh-hashi, Y. et al. (2000). Hypotension andresistance to lipopolysaccharide-induced shock in transgenic mice overexpressingadrenomedullin in their vasculature. Circulation 101, 2309-2316.

Shindo, T., Kurihara, Y., Nishimatsu, H., Moriyama, N., Kakoki, M., Wang, Y., Imai,Y., Ebihara, A., Kuwaki, T. and Ju, K. H. et al. (2001). Vascular abnormalities andelevated blood pressure in mice lacking adrenomedullin gene. Circulation 104, 1964-1971.

Thakali, K., Fink, G. D. and Watts, S. W. (2000). Arteries and veins desensitizedifferently to endothelin. J. Cardiovasc. Pharmacol. 43, 387-393.

Tokes, Z. A., Gendler, S. J. and Dermer, G. B. (1981). The synthesis of a proteinaseinhibitor, alpha-1-antichymotrypsin, by human breast epithelial cells. J. Supramol.Struct. Cell Biochem. 17, 69-77.

Wierland, S. J., Gong, Q., Poblete, H., Fletcher, J. E., Chen, L. Q., Kallen, R. G.(1996). Modulation of human muscle sodium channels by intracellular fatty acids isdependent on the channel isoform. J. Biol. Chem. 271, 19037-19041.

Zhang, X. and Hintze, T. H. (2001). cAMP signal transduction cascade, a novel pathwayfor the regulation of endothelial nitric oxide production in coronary blood vessels.Arterioscler. Thromb. Vasc. Biol. 21, 797-803.

Jour

nal o

f Cel

l Sci

ence

Adrenomedullin and CGRP interact with endogenous ... · Adrenomedullin (AM) and calcitonin gene-related peptide (CGRP) are related peptides with distinct pharmacological proﬁles.

Documents