Multiple angiotensin receptor subtypes in normal and tumor astrocytes in vitro

Multiple Angiotensin ReceptorSubtypes in Normal and Tumor

Astrocytes In VitroDAVID J. FOGARTY, M. VICTORIA SANCHEZ-GOMEZ, AND CARLOS MATUTE*Department of Neurosciences, Faculty of Medicine and Dentistry, University of the Basque

Country, Leioa, Vizcaya, Spain

KEY WORDS AT(1–7) receptor; mitogenesis; brain tumor; tumorigenesis; astrocy-toma; proliferation; development

ABSTRACT A role for neuropeptide receptors in glial tumorigenesis has recently beenproposed. Although angiotensin receptors are known to mediate proliferative effects inmany cell types, including brain astrocytes, the possible participation of these receptors inglial tumorigenesis remains unknown. In the present study, we have examined the expres-sion of the molecularly defined angiotensin receptor subtypes AT1a, AT1b, and AT2 in normalperinatal rat astrocytes and in a panel of tumor adult astrocytoma cells, using the reversetranscriptase-polymerase chain reaction (RT-PCR). Subsequently, we compared the mito-genic effect of the angiotensins A(1–8), A(2–8), A(3–8) and the heptapeptide “metabolite”A(1–7), on both normal and tumor astrocytes, measured in terms of the incorporation oftritiated thymidine. Our results indicate that AT1a, AT1b, and AT2 angiotensin receptormRNA is commonly expressed by many of these cells. Of notable exception is the astrocy-toma U373 which was not found to express AT1 or AT2 mRNA. Chronic (24-h) incubation ofcells with A(1–8) and A(1–7) lead to the induction of mitogenesis, even in the AT1 and AT2mRNA negative astrocytoma cell line U373. Moreover, pharmacological analysis indicatedthat the observed mitogenic effects are not mediated by the AT1 or AT2 type receptors, butrather by a novel, specific A(1–7) angiotensin receptor, since mitogenesis was shown to bepartially blocked by the A(1–7) analogue D-Ala7A(1–7) and by the protease inhibitor ortho-phenanthroline (100 �M). Using Fura-2 spectrophotometry, we found that activation of thisreceptor does not alter intracellular calcium levels; however, preincubation with the proteinkinase kinase inhibitor U0126 (10 �M) was found to inhibit these mitogenic effects partially.Overall, these results which demonstrate that normal and tumor astrocytes express agreater variety of angiotensin receptor subtypes than previously thought, support the ideathat A(1–7) and its receptor signaling system may play an important role in shaping theastrocyte population during development. Moreover, the untimely expression of this A(1–7)receptor may represent an important etiological component in the development of brainastrocytomas. GLIA 39:304–313, 2002. © 2002 Wiley-Liss, Inc.

INTRODUCTION

Pharmacological, electrophysiological, and molecularbiological techniques have revealed the presence of atleast two principal subtypes of angiotensin receptorstermed AT1 and AT2. Both receptors contain 7-trans-membrane (7-TM) domains and are typically coupled toG proteins. The AT1 receptor is antagonized by theimidazole derivative losartan, while the AT2 subtype isantagonized by PD123,319, a tetrahydroimidazopyri-

dine (reviewed in Gallinat et al., 2000). The octapeptideangiotensin II or A(1–8), which is generated from the

*Correspondence to: Carlos Matute, Department of Neurosciences, Faculty ofMedicine and Dentistry, University of the Basque Country, E-48940, Leioa,Vizcaya, Spain. E-mail: [email protected]

Received 21 August 2001; Accepted 8 May 2002

DOI 10.1002/glia.10117

Published online 19 July 2002 in Wiley InterScience (www.interscience.wiley.com).

GLIA 39:304–313 (2002)

© 2002 Wiley-Liss, Inc.

precursor angiotensinogen through the sequential ac-tions of renin and angiotensin converting enzyme(ACE), has been considered the principal endogenousagonist for both receptor subtypes, although it wasrecently proposed that A(2–8) may be the principalactive angiotensin, at least in some tissues (see reviewby Wright and Harding, 1997).

In the light of the abundant and diverse actions ofangiotensins, it has long been anticipated that variousreceptor subtypes should exist (Catt and Abbott, 1991;Lucius et al., 1999). Evidence has accumulated in sup-port of the existence of an AT3 receptor that is insen-sitive to losartan and PD123,319 (Chaki and Inagami,1992) and of an AT4 receptor that is specifically acti-vated by the hexapeptide A(3–8) (Wyse et al., 1995).Moreover, a specific receptor for A(1–7), an A(1–8) me-tabolite, has also been proposed. This A(1–7) receptorhas been found in endothelial cells and vascularsmooth muscle cells (Ferrario et al., 1997), as well inthe brain (Santos et al., 1994). However, the molecularidentity of these putative receptors remains unknown,as they have not yet been cloned.

Angiotensin receptors on brain astrocytes have beenstudied in detail both in situ (Fogarty and Matute,2001) and in in vitro cell culture systems. Most of theevidence suggests that these receptors in vitro are prin-cipally of the AT1 type, although some functional AT2

and AT4 type receptors have also been reported onthese cells (Tallant et al., 1996; Greenland et al., 1996).Astrocyte AT1-type receptors have been found to becoupled to inositol phospholipid hydrolysis throughphospholipase C (PLC) and to the induction of theproto-oncogene c-fos (Sumners et al., 1991; Matute etal., 1994). In tumor astrocytoma cells, AT1, AT2 andAT4 type receptors have also been reported and theactivation of tumor astrocyte angiotensin receptors hasbeen found to be coupled to PLC activation, calciummobilization and prostaglandin release (see Jaiswal etal., 1991; Wyse et al., 1995).

In recent years, the possibility that aberrant signal-ing associated with 7 TM receptors may contribute topathologies including cancer has been raised. Thus,muscarinic receptors were found to be potentially on-cogenic (Gutkind et al., 1991); indeed, it has been spec-ulated that receptors for neuropeptides such as sub-stance P and neurotensin may also participate incellular transformation (reviewed in Sharif, 1998; Ry-der et al., 2001). In the present study, we show thatnormal perinatal astrocytes and adult tumor astrocy-tomas express mRNA encoding a variety of angiotensinreceptor subtypes including AT1a, AT1b, and AT2. Wealso present evidence in favor of the existence of anovel, specific, functional AT(1–7) angiotensin receptorsubtype in both normal and tumor astrocytes whoseactivation leads to the induction of cellular mitogene-sis. These results support the idea that the centralrenin-angiotensin system may play an important rolein shaping the astrocyte population during brain devel-opment. Moreover, alterations in this system may con-

tribute to the development of brain astrocytoma tu-mors.

MATERIALS AND METHODSCell Cultures

Purified cultures of normal rat astrocytes were ob-tained from cerebral cortices of newborn or 1-day-oldrats as described in detail elsewhere (McCarthy andDeVellis, 1980). To remove potentially contaminatingmicroglia, upon first tripsinization, cells were left in-cubating on Sterilin™ Petri dishes for 30 min at 37°C,during which time microglia preferentially adhere tothe plastic substrate. The cell suspension was thenrecovered by gently pouring off the medium, leavingbehind adhering microglia. The purity of these cultureswas assayed by standard immunocytochemical tech-niques (Domercq et al., 1999). Greater than 99% astro-cyte purity was obtained by this method. A panel ofastrocytoma cell lines was purchased from the Ameri-can Type Culture Collection (ATCC) and stored at�140°C until required.

RNA Extraction, RT-PCR, andRestriction Digestion

Cultures were allowed to grow to confluence for10–14 days in Iscove’s modified Dulbecco’s medium(IMDM; Sigma) containing 10% fetal calf serum (FCS;Life Technologies). Cells were suspended using trypsinand sewn in 96-well plates, coated with poly-D-lysine(10 �g/ml) for proliferation assays; 200,000 cells wereresewn in 25-cm2 poly-D-lysine-coated flasks and grownunder conditions similar to those used for cells beingprepared for proliferation assays. In parallel with theinitiation of proliferation assays, cells in the 25-cm2

flasks were trypsinized and RNA was extracted fromthe associated cell pellets using a gDNA-free RNA ex-traction kit (Bioline). RNA with an OD260/280 value of�1.8 was used for subsequent reverse transcription-polymerase chain reaction (RT-PCR).

In this study,1 �g of RNA was reverse transcribedwith AMV-reverse transcriptase (USB) and randomprimers (Promega, Madison, WI) as indicated by thesupplier, with the following modification: to facilitateefficient annealing of the random primers before re-verse transcription, the reverse transcriptase mix, in-cluding the random primers, was incubated at roomtemperature (RT) for 10 min, before the addition of thereverse transcriptase and subsequent incubation at42°C for 1 h.

Enhanced sensitive PCR was achieved by using a hotstart, touchdown protocol. PCR was carried out with a0.4 �l reverse transcriptase product in 40 �l PCR mixfor a total of 40 cycles. After an initial denaturationstep at 94°C for 2 min, Taq polymerase was added tothe PCR mix at 80°C (hot-start). PCR cycles were sub-sequently initiated; they initially consisted of anneal-

305ANGIOTENSIN SUBTYPES IN NORMAL AND TUMOR ASTROCYTES

ing temperatures of 65°C, which dropped by 0.5°C eachcycle until an annealing temperature of 56°C wasreached (touchdown). Subsequent cycles were per-formed with an annealing temperature of 56°C. PCRproducts were resolved on 1.8% agarose gels (Hispana-gar) and visualized using ethidium bromide under ul-traviolet (UV) radiation.

In this study, 10 �l of AT1 PCR product was incu-bated with 250 mU of PvuII restriction endonucleaseand incubated at 37°C for 2 h. Restriction digestionproducts were resolved on 1.8% low melting agarosegels (Hispanagar) and visualized as before. AT2 ampli-cons were similarly digested with the RsaI restrictionenzyme (not shown). Specificity controls: 0.2 �l of RNAextracted from cell cultures was subjected to similarPCR conditions, using 45 cycles and 0.25 �MHUMTH01 primers that detect the short tandem re-peat locus HUMTH01 (Lareu et al., 1994). DNA ex-tracted from the same cultures was used as a positivecontrol. AT1, and AT2 amplicons derived from RT-PCRwere sequenced using the dideoxy chain terminatormethod described by Sanger et al. (1977). The se-quences had 95–98% homology with the correspondingAT1 and AT2 mRNA transcripts (GeneBank Accessionnumbers: m74054 (AT1) and d16480 (AT2)).

Mitogenesis Assays

Both normal and tumor astrocytes were sewn in 96-well plates at a density of 1,500 cells � cm�2. Cells wereallowed to recover for 48 h and were then washed twicewith serum-free Hank’s balanced saline solution(HBSS). They were then synchronized for 24 h inIscove’s modified Dulbecco’s medium (IMDM) with G5supplement (Life Technologies), hereafter referred toas control medium. Synchronized cells were exposed tovarious concentrations of congener angiotensin pep-tides (Sigma) and incubated for 18 h. Medium wasreplaced with fresh medium containing the same con-centration of peptide in addition to 1 �Ci/ml tritiatedthymidine (Amersham) and left for �6 h. In parallelexperiments, cells were incubated with angiotensinsfor 1 h, washed and incubated with control medium for18 h and then incubated with thymidine as before.Incorporated thymidine was retrieved on glass fiberfilter papers (Millipore) and quantified using liquidscintillation counting.

For antagonism experiments, a similar protocol wasfollowed, except that before induction with angiotensinpeptides, a 2-�l dose of antagonist (losartan, a gener-ous gift from Merck Sharpe and Dohme; PD123,319,Sigma; and D-Ala7 A(1–7), Bachem Chemicals) wasadded to the corresponding wells, to achieve the re-quired final concentration of antagonist and incubationproceeded for �15 min before the removal of mediumand the addition of fresh medium containing both ag-onist and antagonist at the required final concentra-tions.

Dose-response curves were constructed from at leastthree independent experiments. In each experiment,eight replicate data points were obtained. The means ofthese octuplicate data points were calculated and per-centage proliferation was expressed as cpm (angioten-sin-exposed cells) � cpm (control cells) � 100/cpm (con-trol cells). The percentage Emax was calculated bynormalizing the dose-response curves from each inde-pendent experiment to 100 (maximum percentage pro-liferation). The standard error of the mean (SEM) ofthe dose-response curves was calculated as the SEM ofthe data from at least three independent experiments(each performed in octuplicate). Sigmoid-type curveswere fitted to the dose-response data, using the “Prism”program, which also calculated the EC50 values.

Fura-2 Spectrophotometry

Alterations in the levels of intracellular calcium innormal and tumor cells were measured, using Fura-2spectrophotometry; 2,500 cells � cm�2 were plated onpoly-D-lysine coated 25-mm discs and allowed to re-cover in IMDM containing 10% FCS for 48 h. Cellswere then incubated with 5 �M Fura-2 acetoxymethyl(Fura-2 AM) ester in IMDM � 10% FCS for 30 min at37°C and then washed of free Fura-2 AM for 30 min inHBSS containing 2 mM calcium before assay. Angio-tensins or other reagents were added to continuouslystirred cell medium (HBSS including 2 mM calcium).Cell fluorescence was measured at excitation wave-lengths of 340 nm and 380 nm and at an emissionwavelength of 505 nm, using a dual-excitation wave-length spectrophotometer. Data were exported inASCII format and analyzed using Excel and Prismsoftware.

Essays With Other Inhibitors

With a view to characterizing further the mecha-nisms underlying angiotensin-mediated mitogenesis,we employed the peptidase inhibitor orthophenanthro-line (Sigma), which inhibits the metabolism of A(1–8)to A(1–7). Synchronized cells were incubated with or-thophenanthroline (100 �M) in basal medium for atleast 15 min, and subsequently with orthophenanthro-line (100 �M) � 1 �M A(1–8) for �18 h. Cells were thenincubated with 1 �Ci/ml thymidine in the same me-dium for a further 6 h before harvesting. Synchronizedcells were treated in the same way with the MAPkinase kinase inhibitor U0126 (10 �M, Tocris), to in-vestigate the participation of this kinase in the A(1–7)-mediated effects. Data were plotted as the mean of atleast three independent experiments, each performedin octuplicate.

306 FOGARTY ET AL.

RESULTSNormal and Tumor Astrocytes express AT1a,AT1b, and AT2 Angiotensin Receptor mRNA

We examined the expression of mRNAs encoding AT1and AT2 angiotensin receptors in purified cultures ofperinatal rat astrocytes and in a panel of culturedhuman astrocytomas, including the cell lines SW1088,T98, U138, and U373 and the rat glioma cell line C6,using custom-designed, subtype-specific primer pairs(Table 1). RT-PCR of gDNA-free RNA extracted fromthese cultures, using these primers, showed that nor-mal and tumor astrocytes express both AT1 and AT2angiotensin receptor mRNAs (Fig. 1A). Only one ofthese cell lines, U373, was consistently found not toexpress AT1 or AT2 transcripts (Fig. 1B). However,�-actin was repeatedly detected in these preparations,confirming the integrity of U373 RNA samples. Restric-tion analysis of the resulting AT1 amplicons employingthe PvuII restriction endonuclease revealed that theAT1B isoform is also expressed by both normal andtumor rat astrocytes, the latter represented by the ratglioma cell line C6 (Fig. 1C). Moreover, this analysisindicated that the proportion of AT1a and AT1b isoformsin normal and tumor rat astrocytes is similar to thatfound in the rat whole brain (AT1a is the predominantisoform; not shown). Amplicons from human cells werenot digested by PvuII, supporting the idea that theAT1b isoform does not exist in the human genome. RNAused for these expression studies was confirmed to befree of gDNA by means of analysis by PCR (Fig. 1D),using HUMTH01 primers, which have typically beenused to detect trace quantities of gDNA in fossil re-mains (de Pancorbo et al., 1995). Thus, whereas a typ-ical amplification in the region of 180–195 bp corre-sponding to the HUMTH01 short tandem repeat wasobserved from human DNA samples, no ampliconswere generated in identical PCRs containing RNA in-stead of DNA.

A(1–8) and A(1–7) Induce the Mitogenesis ofNormal and Tumor Astrocytes

Subsequently, we examined the mitogenic activity ofvarious angiotensin peptides in normal and tumor as-trocytes. Acute incubation of normal and tumor cells

for 1 h with 5 �M A(1–8), A(2–8), A(3–8), or A(1–7) didnot induce any significant mitogenic activity 18 to 24 hlater as assessed by the incorporation of tritiated thy-midine (Fig. 2). In contrast, when cells were exposed tothe angiotensins for 24 h, a strong proliferative effectwas observed with the angiotensins A(1–8) and A(1–7).The congener peptides A(2–8) and A(3–8) remainedineffective. Intriguingly, the AT1 and AT2 mRNA neg-ative cell line U373 also responded mitogenically toboth A(1–8) and A(1–7) (Fig. 2B).

We next examined the dose-response effect of A(1–8)and A(1–7) on the proliferation of normal and tumorastrocytes. As shown in Figure 3, both normal andtumor astrocytes responded mitogenically in a dose-dependent manner to the presence of A(1–8) and A(1–7), with an EC50 value of �1 �M and response satura-tion at around 30 �M. The EC50 values were calculatedto be: rCx, A(1–8) 1.6 �M, A(1–7) 0.7 �M; SW1088,A(1–8) 1.7 �M, A(1–7) 1.0 �M; T98, A(1–8) 0.6 �M,A(1–7) 0.3 �M and U373, A(1–8) 1.2 �M, A(1–7) 1.4�M. The dose-response curves for both peptides werepractically superimposable, suggesting that the pep-tides may act through the same receptor.

To evaluate the participation of AT1 and AT2 recep-tors in the observed mitogenic effects, we repeated theprevious experiments in the presence of the AT1 antag-onist losartan, or alternatively in the presence of theAT2 specific antagonist PD123,319. Neither of theseantagonists at a concentration of 30 �M significantlyaltered the mitogenic dose-response curves associatedwith A(1–8) induction in normal astrocytes (Fig. 4A).This absence of antagonism was confirmed with a sub-EC50 dose of A(1–8) (i.e., 500 nM) and increasing con-centrations of antagonist (�10 �M) (Fig. 4B). In tumorastrocytes, a similar lack of antagonism was observed.Thus, a 50-fold excess of AT1 or AT2 type antagonistswas unable to block the mitogenic effects caused by 1�M of A(1–8) or A(1–7) (Fig. 4C), indicating that nei-ther of these receptors mediates the observed prolifer-ation effects. This conclusion is consistent with the factthat U373 (Fig. 4C) expresses neither AT1 nor AT2receptor RNA. Nevertheless, losartan completely an-tagonized the currents induced by 10 �M A(1–8) inXenopus laevis oocytes injected with mRNA from theadult rat brain (not shown), confirming that this an-tagonist in our hands is capable of antagonizing AT1-mediated effects.

TABLE 1. Sequence and properties of PCR primers used in this study

Name Sequence (5� 3 3�) T melt (°C) GC content (%) Amplicon (bp)

AT1.up ATT GTC YAC CCA ATG AAG TC 60 50AT1.down AAT TTY TTC CCC AGA AAG CC 60 45 539

AT2.up TGG ACC TGT GAT GTG CAA AGT 62 48AT2.down CAC TRC GGA GCT TCT GTT GGA A 62 50 659

�-Actin.up TTG TAA CCA ACT GGG ACG ATA TGG 64 46�-Actin.down GAT CTT GAT CTT CAT GGT GCT AGG 62 46 764,1316

THO1.up ATT CAA AGG GTA TCT GGG CTC TGG 65 50THO1.down GTG GGC TGA AAA GCT CCC CGA TTA T 67 50 183, 195


A(1–7) Generated From A(1–8) Binds to aSpecific A(1–7) Receptor That Transduces a

Mitogenic Signal Via a Calcium-IndependentPathway Involving MAP Kinase Kinase

We then examined the effect of the angiotensinsA(1–8) and A(1–7) (both at 1 �M) on the mobilization ofintracellular calcium in normal and tumor astrocytesusing Fura-2 fluorescent spectrofluorometry. In normalastrocytes, neither of these peptides altered intracellu-lar calcium levels, whereas ATP (1 mM) or the puriner-gic receptor agonist benzoyl-ATP (100 �M) induced asharp increase in these levels, demonstrating that thecells were capable of mobilizing intracellular calcium(Fig. 5A). The reversal of the order of addition of thesepeptides did not alter the response profile (Fig. 5B),ruling out the possibility of heterologous desensitiza-tion. In tumor astrocytes, a similar lack of effect wasobserved. Thus, neither A(1–8) nor A(1–7) significantly

altered intracellular calcium levels in tumor astrocyteswhereas ATP (1 mM) or the neuropeptide Substance P(250 nM) induced measurable increases in the levels ofintracellular calcium (Fig. 5C,D), once again indicatingthat the system is capable of detecting calcium tran-sients associated with receptor stimulation.

To further corroborate the specificity of the observedeffects, we performed mitogenesis assays in the pres-ence of three components: (1) 10 �M D-Ala7A(1–7), anA(1–7) analogue that has been reported to act as aspecific competitive antagonist of the A(1–7) receptor;(2) 100 �M orthophenanthroline, a protease inhibitorthat inhibits the metabolism of A(1–8) to A(1–7); and(3) 10 �M U0126, an inhibitor of MAP kinase kinase,which is implicated in many mitogenic transductionpathways. The results of these experiments (Fig. 6)demonstrated that in the astrocytoma line U373, theeffects of 1 �M A(1–7) are potently blocked byD-Ala7A(1–7), whereas the effect due to 1 �M A(1–8) is

Fig. 1. Angiotensin receptor transcripts are expressed in normalastrocytes and in the majority of a panel of tumor astrocytes. A: AT1and AT2 mRNA expression in mRNA extracted from rat whole brain(rWb), from cultured rat astrocytes (rAst) and from human and ratastrocytoma cell lines (SW, abbreviation for SW1088). Both AT1 (539bp) and AT2 (659 bp), transcripts were found in all mRNA prepara-tions except for (B) U373, in which amplification of the �-actin house-keeping gene mRNA (764 bp) served as a control for the integrity ofthe RNA templates. C: Both AT1a (314 bp) and AT1b (414-bp) mRNAisoforms are expressed in normal and tumor rat astrocytes. The hu-

man AT1 amplicon was resistant to PvuII digestion, consistent withthe fact that the AT1b isoform is not present in the human genome. D:The mRNA used in these studies is free of gDNA. Polymerase chainreaction using HUMTH01 primers, typically employed to detect tracequantities of gDNA in ancient fossil remains, failed to produce anyamplicons in mRNA samples, indicating the absence of gDNA in theRNA employed. Numbers to the right (in C) represent size standardsin base pairs (bp), whereas those to the left represent the size of cutand uncut AT1 restriction fragments.

308 FOGARTY ET AL.

partially inhibited by orthophenanthroline, indicatingthat A(1–8) exerts its effects following its metabolismto A(1–7) and subsequent activation of a specific A(1–7)receptor. The fact that the MAP kinase kinase inhibitorpartially reduced the effects due to A(1–7) indicate thatMAP kinase participates in mediating the observedmitogenic effects.

DISCUSSIONAT1a, AT1b, and AT2 Expression in Normal and

Tumor Astrocytes

Using custom-designed, subtype-specific primers forRT-PCR, we have detected mRNA encoding AT1 andAT2 angiotensin receptor subtypes in both normal andtumor astrocytes in vitro. The specificity of RT-PCR isbased on the following criteria: (1) the size of the am-plicons corresponds to that predicted from a theoretical

analysis of the AT1 and AT2 cloned sequences; (2) AT1

and AT2 amplicons were digested by restriction en-zymes in accordance with theoretical predictions; and(3) sequencing of the amplicons revealed �95% homol-ogy with the corresponding AT1 or AT2 transcript. Inaddition, these AT1 and AT2 amplicons are derivedfrom expressed RNA and not from potentially contam-inating genomic DNA, as we employed a gDNA-free,RNA extraction procedure involving the removal ofgDNA by means of adsorption of the nucleic acid phasewith a silica based Adsorbin™ compound. Moreover,HUMTHO1 primers which have been employed to de-tect trace quantities of gDNA in ancient human fossilmaterial (de Pancorbo et al., 1995) yielded negativeresults in our RNA preparations, corroborating the ab-sence of gDNA in these preparations. AT1 and AT2

mRNA could also possibly be ascribed to the presenceof other cell types in our astrocyte cultures, such asfibroblasts or microglia. However, only those cell cul-tures that expressed �95% GFAP� cells, as assayed byglial fibrillary acidic protein (GFAP) immunocytochem-istry, were employed for RNA extraction or functionalassays. In addition, RT-PCR of mRNA extracted frompurified microglial cell cultures showed that this celltype does not express either AT1 or AT2 transcripts (notshown).

AT1 receptors have already been reported to be ex-pressed by neonatal rat astrocytes using both bindingassays (Bottari et al., 1992) and functional analysis(Sumners et al., 1991; Raizada et al., 1993; Tallant andHigson, 1997). Similarly, the presence of AT1-type re-ceptors in tumor astrocytoma cells has also been re-ported (Jaiswal et al., 1991). Our results with molecu-lar biological techniques confirm and extend thesefindings. By means of restriction digestion of the AT1

amplicons, we analyzed the presence of AT1a and AT1b

transcripts expressed by rat astrocytes. Thus, we re-port for the first time the presence of AT1b receptortranscripts in rat astrocytes. Our results indicate that,as in the whole brain, the AT1a transcript is the pre-dominant form expressed in neonatal rat astrocytes.The PvuII restriction enzyme failed to cut human AT1

amplicons, consistent with the idea that there is onlyone isoform of the AT1 receptor in the human genome(Yoshida et al., 1992). The fact that the proportion ofAT1a and AT1b transcripts was similar in the rat gliomacell line C6 is consistent with the idea that malignancydoes not involve an alteration in the relative levels ofexpression of these isoforms.

The enigmatic AT2 receptor has also been suggestedto be present in cultured astrocytes on the basis ofcompetitive binding studies (Sumners et al., 1991).Likewise, this receptor has also been reported in hu-man tumor astrocytoma cells on the basis of the antag-onism of prostacyclin release by a type-2 antagonistCGP42112A (Jaiswal et al., 1992). Our results clearlyindicate that cultured neonatal rat astrocytes and as-trocytoma cells express AT2 receptor mRNA. Neverthe-less, the functional significance of AT2 expression in

Fig. 2. Mitogenic effect of congener angiotensin peptides in normaland tumor astrocytes. A: Normal rat astrocytes were incubated with5 �M of various congener angiotensin peptides for 1 h or 24 h. A 1-hexposure to any of the peptides was insufficient to induce any mito-genic effect measured 18–24 h later. In contrast, when cells wereincubated for 24 h with A(1–8) or A(1–7), a potent mitogenic effectwas observed. The congener peptides A(2–8) and A(3–8) remainedineffective. B: Similar effects were observed in tumor astrocytes,including the cell line U373 illustrated here. Data are presented asmean SEM, n 4. Values on the ordinate represent percentageincrease in proliferation with respect to control cultures, not exposedto angiotensins.


normal astrocytes continues to elude definition (seealso Gallinat et al., 2000).

Angiotensin-Induced Mitogenesis in Normaland Tumor Astrocytes

Previous studies have indicated that A(1–8) can in-duce the mitogenesis of a variety of cell types includingastrocytes (Matute et al., 1994; Sumners et al., 1994),although the angiotensin receptor subtype that medi-ated these effects was not clearly established. In thelight of our previous results indicating the expressionof both AT1 and AT2 in normal and tumor astrocytes,we aimed to identify the degree of participation of bothreceptor subtypes in the mitogenic effects due to angio-tensins and to identify potential differences betweennormal and tumor astrocytes at this level. Unexpect-edly, we found that neither the AT1 type antagonistlosartan, nor the AT2 type antagonist PD123,319 al-tered the dose-response curves associated with the mi-togenic angiotensins A(1–8) and A(1–7). However, theabsence of mitogenic activity of A(2–8), a congenerpeptide postulated to be the principal active angioten-sin acting at AT1 and AT2 receptors (Wright and Har-ding, 1997), further supported the idea that neitherAT1 nor AT2 receptors participated in the observedmitogenic effects. Furthermore, both A(1–8) andA(1–7) induced mitogenesis in the astrocytoma cell line

U373, which we found did not express AT1 or AT2receptors.

It was also unexpected to find that the heptapeptidecongener A(1–7), typically considered an inactive an-giotensin metabolite (Ferrario et al., 1991), mimickedthe potent mitogenic effect of A(1–8) in normal andtumor astrocytes, with a similar dose-response profile,EC50 (�1 �M) and insensitivity to AT1 or AT2 antago-nists. Heretofore, A(1–7) has consistently been shownto oppose the actions of A(1–8), particularly in thevasculature system (reviewed in Tallant et al., 1999).Our EC50 data are consistent with the idea that bothpeptides may ultimately be acting through the samereceptor. This receptor is unlikely to be the AT3 type,as it has only been reported in mouse neuroblastomacells (Chaki and Inagami, 1992). Participation by theAT4 subtype, which has been reported to be expressedon glial cells, can also be ruled out, since the preferredagonist for this receptor is A(3–8), which we found to bewithout any mitogenic capacity in astrocytes (Fig. 2).The mitogenic effect of A(1–7) in astrocytes is thusmost likely to be mediated by a novel angiotensin re-ceptor, termed the AT(1–7) receptor, in keeping withthe guidelines established by the International Unionof Pharmacology Nomenclature Subcommittee for An-giotensin Receptors (De Gasparo et al., 1995). A specificreceptor for A(1–7), which is insensitive to both AT1and AT2 type antagonists, but can be partially antag-onized by the A(1–7) analogue D-Ala7A(1–7) has re-

Fig. 3. Dose-response curves forthe mitogenic effect of A(1–8) andA(1–7) on (A) normal and (B–D)tumor astrocytes. A: AngiotensinsA(1–8) and A(1–7) induced theproliferation of normal rat astro-cytes, with an EC50 of around 1�M. The effect saturated at �30�M. Similar dose-response curveswere observed in tumor astrocytecell lines, including SW1088(B),T98(C), and U373(D). In the lattercase, the dose-response curveswere superimposable. Data arepresented as mean SEM from atleast three independent experi-ments, each performed in octu-plicate.

310 FOGARTY ET AL.

cently been reported to be expressed in the brain andarguments in favor of the existence of such a receptorhave been proposed in detail elsewhere (see Ferrario etal., 1997; Tallant et al., 1999). The fact that thisAT(1–7) receptor has low affinity for A(1–8) (Rowe etal., 1995) indicated that the mitogenic activity ofA(1–8) may be due to its prior metabolism to A(1–7) bymeans of a carboxypeptidase enzyme (Tallant et al.,1999), an idea corroborated by the finding that thepeptidase inhibitor orthophenanthroline inhibited thepotent mitogenic effect due to A(1–8).

The signal transduction pathways activated by thisAT(1–7) receptor, specifically activated by A(1–7) arecurrently unknown. Our study of the mobilization ofintracellular calcium suggests that the astrocyteAT(1–7) receptor is not coupled to intracellular path-ways typically employed by the AT1 receptor, whichinvolve the activation of PLC, the generation of inositolphosphates and the liberation of calcium from intracel-lular stores. This lack of mobilization of intracellularcalcium by A(1–7) appears to contradict the report thatA(1–8) is coupled to PLC activation, IP formation, andthe release of calcium in normal astrocytes (Sumners etal., 1990, 1991). However, more recent studies reportedthat A(1–8) leads to IP formation in astrocytes in aregion-specific manner. Thus, cultured astrocytes de-rived from the rat medulla and cerebellum respondedto the application of A(1–8) and A(2–8) by increasingIP formation, but astrocytes obtained from the cerebralcortex did not (Tallant and Higson 1997), consistentwith our finding. Similarly, studies of human astrocy-toma cell lines indicated that two of three astrocytomasdid not increase intracellular calcium upon exposure toA(1–8) (Tallant et al., 1991). Furthermore, none ofthese cell lines mobilized calcium in response to A(1–7),in agreement with our results. Interestingly, this cal-cium-independent signaling is consistent with a signaltransduction pathway that uses phospholipase D in acalcium-independent manner to activate the PKC�isoenzymes, which has been reported to be highly over-expressed in many astroglial cell lines (Sharif andSharif, 1999). Our finding that the MAP kinase kinaseinhibitor U0126 partially inhibits A(1–7) mitogenesisis indicative of the participation of MAP kinase also inA(1–7) signal transduction.

Pathological and Physiological Implications

We initially examined the angiotensin receptor phe-notype of normal and tumor astrocytes with a view toidentifying possible differences in the expressionand/or functional coupling of these receptors associatedwith the transformed phenotype. However, at the levelof both expression and function, we did not observe anydifferences characteristic of tumor astrocytes, suggest-ing that these receptors do not participate in the tu-morigenic process. Nevertheless, it must be borne inmind that we compared adult tumor cells with neona-tal normal cells. The similarity in the angiotensin phe-notype of both models raises the possibility that adulttumor astrocytes may be recapitulating the angioten-sin receptor phenotype of neonatal astrocytes whenthey are in the rapidly proliferating phase of astrogen-esis (the last few days of fetal life and the first week ofpostnatal life; see Lee et al., 1987). Our results aretherefore consistent with the idea that astrocyte tu-morigenesis may involve a dedifferentiation to an ear-lier astrocyte angiotensin receptor phenotype.

Our finding that neonatal astrocytes express a mito-genically-coupled AT(1–7) receptor together with the

Fig. 4. A(1–8) and A(1–7)-mediated effects are not mediated by AT1or AT2-type angiotensin receptors. A: Dose-response curve associatedwith A(1–8) was unaltered when normal rat astrocytes were preincu-bated with losartan (30 �M) or PD123,319 (30 �M). Numbers on theordinate represent absolute CPM values. B: Effects of 500 nM A(1–8),which represents a sub-EC50 dose, in normal rat astrocytes wereunaltered by up to a 20-fold excess of losartan or PD123,319. C: Asimilar lack of antagonism by AT1 and AT2 receptor antagonists wasobserved in tumor astrocyte cell lines (U373 illustrated here). Valueson the ordinate of B and C represent percentage proliferation withrespect to control noninduced cultures.


observation that angiotensinogen mRNA appears inthe developing rat fetal brain at the same time thatrapidly dividing astroblasts appear (Lee et al., 1987)support the notion that this novel angiotensin receptorsubtype may play important roles in shaping the as-trocyte population during brain development. In keep-ing with this idea, reduced survival rates of angio-

tensinogen knockout newborns have been reported(Tanimoto et al., 1994). Moreover, the presence of thisreceptor in adult, human astrocytomas suggests thatderegulation of this subtype may contribute to the for-mation of glial tumors of the brain.

Overall, our studies point to new roles for the brainrenin-angiotensin system. Thus, a variety of angioten-sin receptors are expressed in normal and tumor astro-cytes and the novel AT(1–7) angiotensin receptor mayplay an important role in shaping of the astrocyte pop-ulation during development and tumorigenesis.

REFERENCES

Bottari SP, Obermuller N, Bogdal Y, et al. 1992. Characterization anddistribution of angiotensin II binding sites in fetal and neonatalastrocytes from different rat brain regions. Brain Res 585:372–376.

Catt K, Abbott A. 1991. Molecular cloning of angiotensin II receptorsmay presage further receptor subtypes. Trends Pharmacol Sci 12:279–281.

Chaki S, Inagami T. 1992. Identification and characterization of anew binding site for angiotensin II in mouse neuroblastomaneuro-2A cells. Biochem Biophys Res Commun 182:388–394.

De Gasparo M, Husain A, Alexander W, Catt KJ, Chiu AT, Drew M,Goodfriend T, Harding JW, Inagami T, Timmermans BMWM. 1995.Proposed update of angiotensin receptor nomenclature. Hyperten-sion 25:924–927.

de Pancorbo MM, Castro A, Alonso S, et al. 1995. Genetic typing withHUMTHO1, HUMVWA31A, and HUMFES/FPS short tandem re-peat loci, D1S80 variable number tandem repeat locus and HLA-DQalpha of recent and from XII–XIII century spongy bone. Electro-phoresis 16:1612–1616.

Domercq M, Sanchez-Gomez MV, Areso P, Matute C. 1999. Expres-sion of glutamate transporters in rat optic nerve oligodendrocytes.Eur J Neurosci 11:2226–2236.

Ferrario CM, Brosnihan KB, Diz DI, Jaiswal N, Khosla MC, MilstedA, Tallant, EA. 1991. Angiotensin (1–7): a new hormone of theangiotensin system. Hypertension 18(Suppl III):III126–III133.

Fig. 6. A(1–7) generated from the metabolism of A(1–8) activates aD-Ala7A(1–7) sensitive A(1–7) receptor that uses MAP kinase kinaseto transduce part of its mitogenic signal. Preincubation of synchro-nized U373 cells with 10 �M -Ala7A(1–7), an A(1–7) receptor antag-onist, or with 10 �M U0126, which inhibits MAP kinase kinase led toa significant inhibition of the mitogenic effect due to 1 �M A(1–7) inthe presence of these agents. The mitogenic effect of 1 �M A(1–8) wassubstantially reduced when U373 cells were incubated with this an-giotensin in the presence of the peptidase inhibitor orthophenanthro-line (100 �M). Data are expressed as means standard deviation (SD)from at least three independent experiments, each performed in oc-tuplicate. *P � 0.001; ANOVA followed by the Fisher PLSD test, withrespect to cultures treated with 1 �M A(1–7) alone.

Fig. 5. Mitogenic peptidesA(1–8) and A(1–7) do not mobilizeintracellular calcium in normal(A,B) or (C,D) tumor astrocytes. A:Addition of 1 �M A(1–8) or A(1–7)failed to induce changes in intra-cellular calcium levels as mea-sured using Fura-2 fluorometry. Incontrast, ATP (1 mM) or the puri-nergic agonist benzoyl-ATP (100�M) induced a sharp increase inintracellular calcium levels. C: Asimilar lack of mobilization of in-tracellular calcium was observedin tumor astrocytes (T98 shownhere). D: As a control, Substance P(250 nM) elicited a calcium peakindicative of the liberation of cal-cium from intracellular stores.

312 FOGARTY ET AL.

Ferrario CM, Chappell MC, Tallant EA, Brosnihan KB, Diz, DI. 1997.Counterregulatory actions of angiotensin-(1–7). Hypertension 30:535–541.

Fogarty D, Matute C. 2001. Angiotensin receptor-like immunoreac-tivity in adult brain white matter astrocytes and oligodendrocytes.Glia 35:131–146.

Gallinat S, Busche S, Raizada MK, Sumners C. 2000. The angiotensinII type 2 receptor: an enigma with multiple variations. Am J PhysiolEndocrinol Metab 278:E357–3374.

Greenland K Wyse, B, Sernia C. 1996. Identification and character-ization of angiotensin IV binding sites in rat neurone and astrocytecell cultures. J Neuroendocrinol 8:687–693.

Gutkind JS, Novotny EA, Brann MR, Robbins KC. 1991. Muscarinicacetylcholine receptor subtypes as agonist-dependent oncogenes.Proc Natl Acad Sci U S A 88:4703–4707.

Jaiswal N, Tallant EA, Diz DI, Khosla MC, Ferrario CM. 1991. Sub-type 2 angiotensin receptors mediate prostaglandin synthesis inhuman astrocytes. Hypertension 17:1115–1120.

Lareu MV, Phillips CP, Carracedo A, et al. 1994. Investigation of theSTR locus HUMTH01 using PCR and two electrophoresis formats:UK and Galician Caucasian population surveys and usefulness inpaternity investigations. Forensic Sci Int 66:41–52.

Lee H, Campbell DJ, Habener JF. 1987. Developmental expression ofthe angiotensinogen gene in rat embryos. Endocrinology 121:1335–1342.

Lucius R, Gallinat S, Busche S, Rosenstiel P, Unger T. 1999. Beyondblood pressure: new roles for angiotensin II. Cell Mol Life Sci56:1008–1019.

Matute C, Pulakat L, Rio C, et al. 1994. Properties of angiotensin IIreceptors in glial cells from the adult corpus callosum. Proc NatlAcad Sci U S A 91:3774–3778.

McCarthy KD, DeVellis J. 1980. Preparation of separate astroglialand oligodendroglial cell cultures from rat cerebral tissue. J CellBiol 85:890–902.

Raizada MK, Rydzewski B, Lu D, Sumners C. 1993. Angiotensin IItype 1 receptor-mediated stimulation of c-fos gene expression inastroglial cultures. Am J Physiol 265:C1046–C1049.

Rowe BP, Saylor DL, Speth RC, Absher DR. 1995. Angiotensin-(1–7)binding at angiotensin II receptors in the rat brain. Regul Pept56:139–146.

Ryder NM, Guha S, Hines OJ, Reber HA, Rozengurt E. 2001. Gprotein-coupled receptor signaling in human ductal pancreatic can-cer cells: neurotensin responsiveness and mitogenic stimulation.J Cell Physiol 186:53–64.

Sanger F, Nicklen S, Coulson AR. 1977. DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci U S A 74:5463–5467.

Santos RAS, Campagnole-Santos MJ, Baracho NCV, Fontes MAP,Silva LCS, Neves LAA, Oliveira DR, Caligiorne SS, Rodrigues ARV,Gropen C Jr, Carvalho WS, Simoes AC, Silva E, Khosla M. 1994.Characterization of a new angiotensin antagonist selective for an-giotensin-(1–7): evidence that the actions of angiotensin-(1–7) aremediated by specific angiotensin receptors. Brain Res Bull 35:293–298.

Sharif M. 1998. Mitogenic signaling by substance P and bombesin-likeneuropeptide receptors in astrocytic/glial brain tumor-derived celllines. Int J Oncol 12:273–286.

Sharif TR, Sharif M. 1999. Overexpression of protein kinase C epsilonin astroglial brain tumor derived cell lines and primary tumorsamples. Int J Oncol 15:237–243.

Sumners C, Myers LM, Kalberg CJ, Raizada MK. 1990. Physiologicaland pharmacological comparisons of angiotensin II receptors inneuronal and astrocyte glial cultures. Prog Neurobiol 34:355–385.

Sumners C, Tang W, Zelezna B, Raizada MK. 1991. Angiotensin IIreceptor subtypes are coupled with distinct signal transductionmechanisms in neurons and astrocytes from rat brain. Proc NatlAcad Sci U S A 88:7567–7571.

Sumners C, Tang W, Paulding W, Raizada MK. 1994. Peptide recep-tors in astroglia: focus on angiotensin II and atrial natriureticpeptide. Glia 11:110–116.

Tallant EA, Higson JT. 1997. Angiotensin II activates distinct signaltransduction pathways in astrocytes isolated from neonatal ratbrain. Glia 19:333–342.

Tallant EA, Jaiswal N, Diz DI, Ferrario CM. 1991. Human astrocytescontain two distinct angiotensin receptor subtypes. Hypertension18:32–39.

Tallant EA, Diz DI, Ferrario CM. 1996. Identification of AT1 receptorson cultured astrocytes. Adv Exp Med Biol 396:121–129.

Tallant EA, Diz DI, Ferrario CM. 1999. Antiproliferative actions ofangiotensin-(1–7) in vascular smooth muscle. Hypertension34:950–957.

Tanimoto K, Sugiyama F, Goto Y, Ishida J, Takimoto E, Yagami K,Fukamizu A, Murakami K. 1994. Angiotensinogen-deficient micewith hypotension. J. Biol. Chem. 269:31334–31337.

Wright JW, Harding JW. 1997. Important roles for angiotensin IIIand IV in the brain renin-angiotensin system. Brain Res Rev 25:96–124.

Wyse B, Greenland K, Sernia C. 1995. Specific binding sites for (3–8)angiotensin in C6 glioma cells. Brain Res 681:41–46.

Yoshida H, Kakuchi J, Guo D, et al. 1992. Analysis of the evolution ofangiotensin II type 1 receptor gene in mammals (mouse, rat, bovineand human). Biochem Biophys Res Commun 186:1042–1049.


Multiple angiotensin receptor subtypes in normal and tumor astrocytes in vitro

Documents