Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The alpha-2 adrenergic receptors constitute a group of receptors that mediate many of the physiological effects of the endogenous catecholamines epinephrine and norepinephrine, through activation of G-proteins of the G_i or G_o class. Extensive radioligand binding studies and more limited functional investigations have demonstrated the existence of three pharmacological receptor subtypes (alpha-2A, alpha-2B and alpha-2C), which can be distinguished by their affinity for subtype-selective drugs such as oxymetazoline, prazosin, chlorpromazine and WB4101 (Bylund, 1988). The genes that encode alpha-2A, alpha-2B and alpha-2C adrenergic receptors have been cloned and were termed ₂C10, ₂C2 and ₂C4, respectively, on the basis of their localization on human chromosomes (Kobilka et al., 1987; Regan et al., 1988; Lomasney et al., 1990). The physiological significance of this diversity of receptors is not fully understood. According to binding data (DeVos et al., 1992) and RNase protection experiments (Perala et al., 1992; Berkowitz et al., 1994), the different receptor subtypes and their RNAs exhibit distinct tissue distribution, suggesting that they may be endowed with discrete functions in vivo. However, with the exception of a few situations where a given receptor has been assigned to a particular function (Trendelenburg et al., 1994; MacMillan et al., 1996; Link et al., 1996), the specific roles of each subtype are far from clear. Recent results obtained in differentiated MDCK II cells transfected with constructs allowing the expression of epitope-tagged alpha-2A, alpha-2B or alpha-2C adrenergic receptors also pointed out subtype differences in subcellular distribution and membrane targeting (Wozniak and Limbird, 1996). Finally, another major difference between receptor subtypes may result from the fact that they are differentially regulated. In this respect, experiments on transfected Chinese hamster ovary cells demonstrated that short-term exposure to epinephrine resulted in a subsequent attenuation of alpha-2 agonist-induced inhibition of adenylyl cyclase in cells expressing the alpha-2A or alpha-2B subtype, whereas no desensitization was observed with the alpha-2C subtype (Eason and Liggett, 1992; Kurose and Lefkowitz, 1994). Moreover, alpha-2A and alpha-2B subtypes undergo down-regulation after long-term exposure to the agonist, but alpha-2C does not (Eason and Liggett, 1992). Other experiments carried out on cell lines endogenously expressing alpha-2 adrenergic receptors, however, suggest that the situation could be much more complex. Treatment of OK cells, a model that natively expresses the alpha-2C subtype, with norepinephrine indeed results in a rapid decrease in the potency of alpha-2 agonists to inhibit cAMP production (Jones et al., 1990). Moreover, long-term exposure to the neurotransmitter induces a significant reduction in receptor number (Shreve et al., 1991; Pleus et al., 1993), demonstrating that both desensitization and down-regulation of the alpha-2C adrenergic receptor subtype occurs in these cells. It is thus likely that the regulation of alpha-2 adrenergic receptors is not only subtype specific but also cell type dependent.

In contrast to the study of the human alpha-2A adrenergic receptor, which greatly benefited from the availability of the colon adenocarcinoma cell line HT29 (Jones et al., 1990; Sakaue and Hoffman, 1991; Devedjian et al., 1991), the study of other human receptor subtypes has been hampered by the lack of cell lines constitutively expressing alpha-2B and/or alpha-2C subtypes. The aim of the present work was therefore to find such models. The analysis of a panel of human cell lines from various origins, using RT-PCR, RPA and radioligand binding, unequivocally demonstrated that the hepatocarcinoma cell line HepG2 and the neuroblastoma cell line SK-N-MC express alpha-2 adrenergic receptors of the alpha-2C subtype.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Drugs and reagents. [³H]RX821002 (59 Ci/mmol) was from Amersham (Amersham, UK), [³H]MK912 (80.9 Ci/mmol) from New England Nuclear (Boston, MA) and [-³²P]UTP from ICN (Costa Mesa, CA). Phentolamine was donated by Ciba-Geigy (Basel, Switzerland) and prazosin hydrochloride and UK14304 tartrate by Pfizer (Sandwich, UK). Oxymetazoline, forskolin, pertussis toxin, GppNHp and all other chemicals were from Sigma. Fetal calf serum was purchased from Gibco-BRL (Cergy Pontoise, France). Radioimmunoassay kits for cAMP determination were from Immunotech (Luminy, France). The antibodies generated against the common carboxyl-terminal decapeptide of _i1 and _i2 (anti-_i1/_i2) or against the carboxyl-terminal decapeptide of _o (anti-_o/_i3) were generously provided by Dr. B. Rouot (INSERM U.431, Université Montpellier II, Montpellier, France). The human alpha-2 adrenergic receptor genes (₂C2, ₂C4 and ₂C10) were kindly provided by Dr. R. J. Lefkowitz (Duke University, Durham, NC).

Cell culture. The human cell lines HT29 and CaCo2 (colon adenocarcinoma), HepG2 and SK-Hep (hepatocarcinoma), HeLa (cervix), IMR-90 (lung), A-431 (epidermoid carcinoma) and SK-N-MC (neuroblastoma) were from the American Type Culture Collection (Rockville, MD). HT29, HepG2, SK-Hep, HeLa, A-431 and IMR-90 cells were grown in DMEM containing 25 mM glucose, 100 µg/ml streptomycin and 100 IU/ml penicillin, supplemented with 5% (HT29) or 10% (CaCo2, HepG2, SK-Hep, HeLa and IMR-90) fetal calf serum. SK-N-MC cells were cultured in minimal essential medium containing the same concentrations of antibiotics and supplemented with 5% fetal calf serum and 2 mM glutamine.

Expression of the human alpha-2 adrenergic receptor subtypes in COS-7 cells. The plasmids that were used to express the different alpha-2 adrenergic receptor subtypes in COS-7 cells were pDP₂C2, pDP₂C4 and pDP₂C10. These expression vectors were constructed in our laboratory and contained the cytomegalovirus promoter, the entire coding region of the ₂C2, ₂C4 or ₂C10 gene and the BamHI-XhoI fragment of rabbit -globin genomic sequence (IVS2-), to increase the stability of the transcripts. COS-7 cells were transfected using the DEAE-dextran method (Cullen, 1987) and were collected 48 hr later.

Preparation of cellular RNAs and RT-PCR experiments. Total cellular RNAs were isolated using the guanidium isothiocyanate/phenol-chloroform extraction method (Chomczynski and Sacchi, 1987). The integrity of the preparations was assessed by agarose gel electrophoresis, and the RNA concentrations were measured by UV spectrophotometry. Poly(A)⁺ RNAs used in RT-PCR experiments were prepared from 10 µg of cellular RNAs using the Dynabeads-oligo(dT)₂₅ kit (Dynal, Oslo, Norway). RT was performed for 1 hr at 37°C in a 20-µl reaction volume containing 5 ng/µl oligo(dT)_12-18 (Pharmacia Biotech, Uppsala, Sweden), 0.5 mM deoxynucleotide triphosphates and 200 U of SuperscriptII reverse transcriptase (Gibco-BRL, Cergy Pontoise, France) in supplied buffer. Reaction was stopped by a 5-min treatment at 95°C. Negative controls were treated identically, except that reverse transcriptase was omitted from the reaction. PCRs were carried out with 20 µl of cDNAs in a 100-µl reaction volume containing 10% dimethylsulfoxide, 2.5% deionized formamide, 2.5 mM MgCl₂, 50 µM deoxynucleotide triphosphates, 125 nM levels of each primer and 2.5 U of Taq DNA polymerase (Promega, Madison, WI). The reaction mixture was covered with two drops of mineral oil and subjected to incubation for 3 min at 92°C and 30 cycles consisting of 1 min at 92°C, 1.5 min at 55°C and 1.5 min at 72°C, followed by a final elongation for 7 min at 72°C, in a thermal cycler (Biometra, Goetingen, Germany).

Oligonucleotides. Three sets of primers were used in PCR experiments. The sense and antisense primer pairs for ₂C2 and ₂C4 gene were identical to those used by Eason and Liggett (1993). The ₂C2 primers 5-CCTGGCCTCCAGCATCGGAT-3 (sense) and 5-CAGAGCACAAAAACGCCAAT-3 (antisense) amplified a 630-bp fragment corresponding to nucleotides 519 to 1148 of the ₂C2 ORF. The digestion of this product by PstI gives two fragments of 465 and 165 bp. The ₂C4 primers 5-GTGGTGATCGCCGTGCTGAC-3 (sense) and 5-CGTTTTCGGTAGTCGGGGAC-3 (antisense) amplified a 574-bp fragment corresponding to nucleotides 214 to 787 of the ₂C4 ORF. BstXI digestion of this product gives three fragments, of 271, 225 and 78 bp. The ₂C4/10 primers 5-AAACCTCTTCCTGGTGTCTCT-3 (sense) and 5-GTGCGCTTCAGGTTGTACTC-3 (antisense) allow amplification of either a 233-bp fragment corresponding to nucleotides 259 to 491 of the ₂C4 ORF or a 234-bp fragment corresponding to nucleotides 204 to 437 of the ₂C10 ORF. The product from ₂C10 amplification but not that from ₂C4 amplification contains a BglII restriction site, generating two fragments of 117 bp each. Conversely, the product from ₂C4 amplification but not that from ₂C10 amplification contains a SacI restriction site, generating two fragments of 153 and 80 bp.

Preparation and synthesis of the subtype-specific riboprobes. The antisense riboprobes for detection of ₂C2, ₂C4 and ₂C10 mRNAs were obtained by subcloning regions of the three genes into pBluescriptII KS+ (pKS+; Stratagene, La Jolla, CA). The plasmid pKSC2-221 contained a 221-bp fragment (BamHI-HindIII) corresponding to nucleotides 1311 to 1531 of the ₂C2 sequence. The plasmid pKSC4-370 contained a 370-bp fragment (SmaI-MaeIII) corresponding to nucleotides 1014 to 1382 of the ₂C4 sequence. The plasmid pKSC10-352 contained a 352-bp fragment (PstI-PstI) corresponding to nucleotides 1041 to 1392 of the ₂C10 gene. For synthesis of the radiolabeled probes, the plasmids were linearized with the appropriated restriction enzyme and antisense RNAs were synthesized in the presence of [³²P]UTP, using T3 RNA polymerase (Promega).

RPA. RPAs were performed as previously described but with slight modifications (Devedjian et al., 1991). Two hundred micrograms of lyophilized cellular RNAs were taken up in 30 µl of hybridization buffer [80% deionized formamide, 0.4 M NaCl, 1 mM EDTA, 40 mM piperazine-N,N-bis(2-ethanesulfonic acid), pH 6.7] containing an excess of ³²P-labeled riboprobe. The samples were heated to 95°C for 5 min and then immediately placed at 55°C for 14 hr. Nonhybridized probe was eliminated by the addition of 0.3 ml of RNase A (40 µg/ml) and RNase T₁ (2 µg/ml), in 300 mM NaCl, 5 mM EDTA, 10 mM Tris-HCl (pH 7.5). After 2 hr at 37°C, digestion was stopped by addition of 5 µl of proteinase K (10 mg/ml) and the samples were further incubated for 15 min at 37°C. Carrier tRNA (10 µg) and 0.3 ml of solution D (4 M guanidinium thiocyanate, 25 mM sodium citrate, pH 7.0, 0.1 M 2-mercaptoethanol, and 0.5% sarkosyl) were added to each tube, and protected hybrids were precipitated with isopropyl alcohol. After washing with 70% ethanol, RNA pellets were dissolved in 10 µl of sample buffer (97% deionized formamide, 0.1% sodium dodecyl sulfate, 10 mM Tris-HCl, pH 7.0) and loaded onto a 5% acrylamide/7 M urea gel. The gels were exposed for 48 hr, at 80°C, to X-ray film (Hyperfilm; Amersham), with intensifying screens. The quantification of the amounts of radiolabeled antisense probe protected by mRNA was performed using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA).

Receptor quantification. The quantification of alpha-2 adrenergic receptors was performed on crude membrane preparations using the selective alpha-2 adrenergic antagonists [³H]RX821002 (Devedjian et al., 1994) and [³H]MK912 (Pettiborne et al., 1989). Frozen cells were harvested in 5 ml of TE buffer (50 mM Tris-HCl, 5 mM EDTA, pH 7.5), disrupted using a Dounce homogenizer and centrifuged at 39,000 × g for 10 min. The particulate fraction was washed in TE buffer, and the final crude membrane pellet was taken up in the appropriate volume of TM buffer (50 mM Tris-HCl, 0.5 mM MgCl₂, pH 7.5) for immediate use. The protein concentration was determined using the Coomassie blue method (Bradford, 1976). Total binding was measured by incubating 100 µl of cell membrane with the radioligand in a total volume of 400 µl of TM buffer. After a 45-min incubation at 25°C, bound radioactivity was separated from free by filtration through GF/C Whatman filters, using a Millipore manifold sampling unit. Filters were rapidly washed with ice-cold TM buffer, and membrane-bound radioactivity was determined by liquid scintillation counting. Specific binding was defined as the difference between total and nonspecific binding measured in the presence of 10 µM phentolamine. For saturation studies, the final concentrations of radioligand ranged from 0.25 to 22 nM for [³H]RX821002 and from 0.04 to 8 nM for [³H]MK912. For inhibition studies, the indicated concentrations of competitor were added to the incubation mixture before addition of the membrane suspension. Saturation isotherms and inhibition curves were analyzed using the EBDA-LIGAND computer programs (McPherson, 1985).

Immunoblotting of G_i-protein -subunits. Immunoblotting was carried out as described previously (Homburger et al., 1987). After electrophoresis on 10% sodium dodecyl sulfate-polyacrylamide gels, proteins were electrotransferred to nitrocellulose membranes (16 hr at 150 mA). To minimize nonspecific protein binding, the nitrocellulose sheet was treated with Tris-saline buffer (10 mM Tris-HCl, 500 mM NaCl, pH 7.5) containing 2% gelatin. The blots were incubated overnight at room temperature in the same buffer supplemented with 1% gelatin and containing anti-_i1/_i2 or anti-_o/_i3 (1/250 dilution). After three washes in 50 mM Tris-HCl, 500 mM NaCl, 0.05% Tween 20, the blots were incubated for 1 hr with ¹²⁵I-protein A (0.14 µCi/ml) in 50 mM Tris-HCl, 500 mM NaCl, 0.02% NaN₃. After extensive washing in Tris-HCl buffer containing 500 mM NaCl and 0.05% Tween 20, blots were dried and autoradiographed as indicated above.

Determination of cAMP content. Cells were detached in phosphate-buffered saline containing 0.6 mM EDTA and were collected by gentle centrifugation (400 × g for 5 min at 37°C). The pellet was suspended in DMEM buffered with 25 mM Hepes (pH 7.4). Aliquots of the cell suspension (180 µl, corresponding to 0.4 mg of total protein) were incubated in a 200-µl final volume of Hepes-buffered DMEM containing 0.2 mM 3-isobutyl-1-methylxanthine and the indicated concentration of the drug to be tested. After 15 min at 37°C, the reaction was stopped by adding 1.8 ml of methanol/formic acid (95:5, v/v). The cell lysate was centrifuged (3000 × g for 10 min at 4°C), and an aliquot of supernatant was evaporated. The dry samples were taken up in acetate buffer containing 0.1% NaN₃, and their cAMP content was determined by radioimmunological assay (Steiner et al., 1972).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

RT-PCR experiments. The aim of the present work was to search for human cell lines expressing alpha-2 adrenergic receptors of subtypes other than alpha-2A. The RT-PCR approach was chosen as a primary screen because it allowed us to rapidly test a large panel of cell types and because of its high sensitivity. As a first step, the specificity of each primer pair was assessed on DNase-treated RNAs prepared from COS-7 cells transfected with pDP₂C2, pDP₂C4 and pDP₂C10. As can be seen in figure 1, the ₂C2 and the ₂C4 primer pairs allowed amplification of a single fragment of the expected molecular size (630 and 574 bp, respectively) only in cells transfected with the corresponding gene. On the other hand, the ₂C4/C10 primers worked as well on RNA from cells transfected with pDP₂C4 as on that from cells transfected with pDP₂C10 and allowed amplification of two fragments of indistinguishable size (233 and 234 bp, respectively; see "Materials and Methods" for additional comments).

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Specificity of the oligonucleotide primers used in RT-PCR experiments. RT-PCRs were performed as described in "Materials and Methods," using 2C2 primers (left), 2C4 primers (middle) or 2C4/C10 primers (right). The specificity of the different oligonucleotide pairs was validated on DNase-treated RNAs prepared from COS-7 cells transfected with pDP2C2 (C2), pDP2C4 (C4) or pDP2C10 (C10). Fragments of the expected molecular size were amplified only in reactions where the primer pair corresponding to the transfected gene was used. Lane M, 100-bp ladder.

View larger version (53K): [in this window] [in a new window]   Fig. 2.   RT-PCR experiments on RNAs from HepG2, SK-N-MC and HT29 cells. Upper, poly(A)+ RNAs were prepared from HepG2 (left), SK-N-MC (middle) and HT29 (right) cells. RT-PCRs were performed as described in "Materials and Methods," using 2C2 primers (2C2), 2C4 primers (2C4) or 2C4/C10 primers (2C4/C10). Lanes , reactions where reverse transcriptase was omitted. Lower, restriction analysis of RT-PCR products. The fragments resulting from the amplification with 2C2 primers (2C2) and 2C4 primers (2C4) were digested with PstI (P) and BstXI (X), respectively. The fragments resulting from the amplification with 2C4/C10 primers (2C4/C10) were digested with either BglII (B) (which cuts 2C10 product) or SacI (S) (which cuts 2C4 product). Lanes , undigested products. Lanes M, 100-bp ladder.

RPAs. Because our RT-PCR conditions were not quantitative, RPAs were performed, to yield a better estimation of the respective amounts of the three RNA species in the three cell lines that gave positive results in RT-PCR experiments. The subtype specificity of the riboprobes used in these experiments has been demonstrated in previous work (Valet et al., 1993), and assays were performed with 200 µg of cellular RNAs, to detect weak signals. Qualitatively, the results from RPA (fig. 3) agreed fully with those from RT-PCR experiments; however, the amounts of the different receptor RNAs appeared to vary widely, according to the cells considered. HepG2 was found to express exclusively ₂C4 and contains large amounts of this gene transcript. In agreement with binding data (Devedjian et al., 1994) and with previous results from RPAs (Devedjian et al., 1991), the predominant alpha-2 adrenergic receptor mRNA species in HT29 is the ₂C10 mRNA. A very weak ₂C2 mRNA signal was also detected, and its quantification indicated that it represented <1% of the amount of ₂C10 in this cell line. SK-N-MC cells contained fairly similar amounts of ₂C4 and ₂C10 mRNA but, in contrast to conclusions from RT-PCR, the amounts of ₂C4 and ₂C10 transcripts were much lower than that of ₂C4 in HepG2 cells and than that of ₂C10 in HT29 cells. Together, the results from RT-PCR and RPA definitively show that HepG2 and SK-N-MC contain ₂C4 mRNAs. Previous studies carried out with SK-N-MC and HepG2 have demonstrated that both cell lines also contain alpha-1 adrenergic receptor mRNAs. However, although the presence of transcripts was followed by alpha-1 adrenergic receptor expression in SK-N-MC (Esbenshade et al., 1995), no receptor was detected in HepG2 (Kost et al., 1992). Binding experiments were thus conducted to verify that the presence of ₂C4 transcripts was accompanied by alpha-2C adrenergic receptor expression.

View larger version (30K): [in this window] [in a new window]   Fig. 3.   RPA of RNAs from HepG2, SK-N-MC and HT29 cells. Two hundred micrograms of RNAs prepared from HepG2, SK-N-MC and HT29 cells were hybridized with an excess of 2C2-221 (left), 2C4-370 (middle) or 2C10-352 (right) 32P-labeled riboprobe. Samples were digested with a mixture of RNases A and T1. The resistant hybrids were electrophoresed and gels were autoradiographed as described in "Materials and Methods." Lanes P, undigested probes; lanes , hybridization with 5 µg of RNA from untransfected COS-7 cells; lanes C2, C4 and C10, hybridization with 5 µg of RNA from COS-7 cells transfected with pDP2C2, pDP2C4 and pDP2C10, respectively.

Binding experiments. Alpha-2 adrenergic receptors were quantified using the radiolabeled antagonists [³H]RX821002 and [³H]MK912. The Scatchard plots obtained by transformation of saturation binding data for HT29, SK-N-MC and HepG2 cell membranes using [³H]RX821002 are depicted in figure 4. Under the conditions used, [³H]RX821002 labeled a single class of high-affinity binding sites in HT29 and HepG2 membrane preparations, and statistical analysis of the data indicated that a two-site model did not give a better fit than the one-site model. On the other hand, Scatchard transformation of [³H]RX821002 binding on SK-N-MC yielded curvilinear plots, and in this case data were better fitted by a two-component model than by a one-component model (P < .002). The results obtained under the same conditions but with [³H]MK912 as radioligand are reported in figure 5. Like [³H]RX821002, this molecule appears to label a single class of binding sites in HT29 and HepG2 membrane preparations, whereas it reveals an heterogeneous population of binding sites in SK-N-MC. Again, in contrast to observations for HT29 and HepG2, binding of [³H]MK912 on SK-N-MC membranes was significantly better fitted by a two-site model (P < .001). Table 1 summarizes the results from the computer-assisted analysis of the data obtained from several experiments. It also allows us to compare the binding parameters of [³H]RX821002 and [³H]MK912 with membranes from the three cell lines with those obtained with membranes from COS-7 cells transfected with the ₂C2, ₂C4 or ₂C10 gene. As expected, HT29 cells expressed only one class of binding sites, with K_d values for [³H]RX821002 (1.01 ± 0.24 nM) and [³H]MK912 (0.64 ± 0.27 nM) that match those found in COS-7 cells transfected with the ₂C10 gene (1.37 ± 0.33 and 0.71 ± 0.31 nM, respectively). The sites labeled in HepG2 cells displayed a moderate affinity for [³H]RX821002 (K_d = 3.5 ± 0.4 nM) and a remarkably high affinity for [³H]MK912 (K_d = 0.08 ± 0.02 nM), which is characteristic of the ₂C4-adrenergic receptor subtype. SK-N-MC exhibited fairly similar amounts of two binding site populations with different affinities for [³H]RX821002 and [³H]MK912. Given the relative selectivity of the two radioligands for the different receptor subtypes, it is obvious that the receptor subpopulation displaying high affinity for [³H]RX821002 corresponds to alpha-2A, whereas that having high affinity for [³H]MK912 corresponds to the alpha-2C subtype. According to the radioligand used, the alpha-2A receptor fraction represented 32 to 46% of the whole receptor population. The pharmacological properties of these receptors were further studied by comparing the ability of the subtype-selective drugs oxymetazoline and prazosin to inhibit ³H-labeled antagonist binding to membrane preparations from HT29, HepG2 and SK-N-MC cells (table 2). Whereas oxymetazoline was several orders of magnitude more potent than prazosin in preventing [³H]RX821002 binding to HT29 receptors (K_i ratio = 470), these two compounds exhibited similar potencies for inhibiting [³H]MK912 binding at HepG2 receptors (K_i ratio = 0.55). In agreement with the conclusion from saturation data, inhibition curves for [³H]RX821002 binding to SK-N-MC receptors were clearly biphasic and allowed two sites, with distinct affinities for oxymetazoline or prazosin, to be distinguished. Based on the known selectivity of the two compounds, sites having high affinity for oxymetazoline and low affinity for prazosin correspond to the alpha-2A subtype receptor population, whereas those having fairly similar affinities for the two compounds represent alpha-2C receptors.

View larger version (13K): [in this window] [in a new window]   Fig. 4.   Scatchard plots of [3H]RX821002 binding to membranes of HepG2, SK-N-MC and HT29 cells. Membranes prepared from HepG2 (left), SK-N-MC (middle) and HT29 (right) cells were incubated in the presence of various concentrations of radioligand. The amount of specifically bound [3H]RX821002 was determined using 105 M phentolamine to estimate nonspecific binding. The presented data are from a typical experiment, and each point represents the mean of duplicates.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Scatchard plots of [3H]MK912 binding to membrane of HepG2, SK-N-MC and HT29 cells. Membranes prepared from HepG2 (left), SK-N-MC (middle) and HT29 (right) cells were incubated in the presence of various concentrations of radioligand. The amount of specifically bound [3H]MK912 was determined using 105 M phentolamine to estimate nonspecific binding. The presented data are from a typical experiment, and each point represents the mean of duplicates.

View this table: [in this window] [in a new window]   TABLE 1 Binding parameters of [3H]RX821002 and [3H]MK912 Membranes prepared from HepG2, SK-N-MC, HT29 and COS-7 cells transfected with pDP2C2, pDP2C4 or pDP2C10 were incubated in the presence of increasing concentrations of [3H]RX821002 or [3H]MK912, and specific binding was determined as described in "Materials and Methods." Computer analysis of the binding data indicated that the two radioligands labeled a single class of binding sites in HepG2, HT29 and transfected COS cells. In contrast, two receptor populations were distinguishable in SK-N-MC cells. The maximum number of binding sites (Bmax) and the dissociation constant value (Kd) were calculated by nonlinear regression analysis of the data according to a one- or two-component model. Reported values are the means ± S.E.M. from n (in parentheses) determinations.

View this table: [in this window] [in a new window]   TABLE 2 Inhibition of radioligand binding by oxymetazoline and prazosin Inhibition studies were performed as described in "Materials and Methods." The concentrations of radioligand used in these experiments were 4 nM [3H]RX821002 for HT29, 0.7 nM [3H]MK912 for HepG2 and 12 nM [3H]RX821002 for SK-N-MC. Data were analyzed using computer programs allowing curve fitting to a one-site (HT29 and HepG2) or two-site (SK-N-MC) inhibition model (McPherson, 1985). Reported values are means ± S.E. of three determinations.

Receptor coupling, identification of G_i-proteins and inhibitory effect of UK14304 on cAMP production. The degree of receptor coupling to G-proteins was estimated by studying the inhibition of ³H-antagonist binding by the alpha-2 agonist UK14304, in the absence or presence of GppNHp/Na. For all cell lines considered (fig. 6), agonist-inhibition curves obtained under control conditions exhibited a Hill coefficient value significantly different from 1 and were better fitted by a multicomponent model. The addition of GppNHp/Na to the binding medium resulted in a rightward shift and in a significant increase in the slope of the curves, reflecting the conversion of the whole receptor population into a low-affinity state for the agonist. From computer-assisted analysis of the data, it was calculated that the percentage of receptors in the high-affinity state for UK14304 was 43 ± 11% in HepG2 and 29 ± 7% in SK-N-MC. Similar observations were made when membranes were treated with pertussis toxin and the G_i/G_o-proteins expressed by the cells were identified by immunoblotting using either anti-_o/_i3 or anti-_i1/_i2 antibody. As shown in figure 7, each of these antibodies recognized a single band in HepG2 and SK-N-MC. Comparison of the relative mobility of these proteins with those labeled in rat brain membranes indicated that the two cell lines express G_i2 and G_i3. In a final series of experiments, the biological efficacy of the receptor was tested by measuring the extent of the inhibitory effect of UK14304 on the intracellular cAMP accumulation induced by forskolin. The alpha-2 agonist caused a significant reduction of forskolin-induced cAMP accumulation in both cell lines (table 3). On the basis of 18 determinations in three independent experiments, it was calculated that the extent of the inhibitory effect of UK14304 reached 47% in HepG2 and 23% in SK-N-MC. This effect was dose-dependent and was abolished by addition of an excess of yohimbine or by prior treatment of the cells with pertussis toxin (data not shown).

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Inhibition of 3H-labeled antagonist binding by UK14304. Binding studies were performed as described in "Materials and Methods." Concentrations of [3H]MK912 and [3H]RX821002 were 0.7 nM (HepG2) and 8 nM (SK-N-MC), respectively. Inhibition of radioligand binding by UK14304 were measured in the absence () or presence () of GppNHp (100 µM) plus NaCl (100 mM). The presented data are from a typical experiment, and each point represents the mean of duplicates.

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Identification of Gi-protein -subunits by immunoblotting. Membrane proteins from rat brain (RB) and HepG2 and SK-N-MC cells were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. Immunoblottings were performed as described in "Materials and Methods," with either anti-i1/i2 antibody (left) or anti-o/i3 antibody (right).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The existence of three subtypes of alpha-2 adrenergic receptors (namely, alpha-2A, alpha-2B and alpha-2C) was initially proposed on the basis of the pharmacological profiles of receptors expressed in various species tissues and was then confirmed by the cloning of ₂C10, ₂C2 and ₂C4 genes in humans. The prototypical cell lines expressing each of these subtypes are so far considered to be the colon adenocarcinoma cell line HT29 for alpha-2A, the neuroblastoma-glioma hybrid cell line NG108-15 for alpha-2B and the epithelial kidney cell line OK for alpha-2C (Bylund, 1988). One of the major limitations to the use of NG108-15 and OK cells for the study of the regulation of alpha-2B and alpha-2C adrenergic receptors results from the fact that they are not of human origin. NG108-15 is a mouse-rat hybrid cell line and expresses a receptor population that may be a mixture of the products from two different genes, RNG and RG10 (Hu et al., 1993), which are the rodent homologs of human ₂C2 and ₂C4, respectively. On the other hand, OK cells are derived from American opossum and express a receptor for which the amino acid sequence exhibits large divergences from human ₂C4 (Blaxall et al., 1994). Major differences between the two polypeptides are located in the third intracellular loop, a region particularly important in receptor regulation. On the basis of binding data, the retinoblastoma Y79 line was proposed to represent a model of human cells expressing alpha-2C adrenergic receptors (Gleason and Hieble, 1992). However, another study assigned the Y79 receptor to the alpha-2A subtype (Kazmi and Mishra, 1989), and definitive demonstration that the receptor expressed in this cell line is encoded by ₂C4 is missing. One of the consequences of the lack of human models clearly established as expressing alpha-2B and/or alpha-2C is that the study of the regulation of these two human receptor subtypes has thus far been restricted to homologous regulation in transfected cells (Eason and Liggett, 1992; Kurose and Lefkowitz, 1994).

Based on data from mRNA identification by RT-PCR and RPA and on the results from binding experiments with [³H]RX821002 and [³H]MK912, the present work unequivocally demonstrates that the human hepatoma cell line HepG2 and the human neuroblastoma cell line SK-N-MC both contain alpha-2 adrenergic receptors of the alpha-2C subtype. HepG2 expresses exclusively this receptor subtype, at a density of 55 fmol/mg of protein. The receptor is coupled to pertussis toxin-sensitive G-proteins (G_i2 and/or G_i3), and its stimulation by the alpha-2 agonist UK14304 efficiently reduces cAMP production, indicating that inhibition of adenylyl cyclase is one of the primary mechanisms of signal transduction in these cells. The presence of alpha-2C adrenergic receptors in HepG2 was rather unexpected. This cell line, which is commonly used in many laboratories, was apparently never examined in this respect. It would be interesting to determine whether the presence of alpha-2C adrenergic receptors corresponds to ectopic expression in a given transformed cell or whether it reflects the situation in normal human hepatocytes. The lack of alpha-2 adrenergic receptors in the other human hepatoma cell line, SK-Hep, and the failure to identify noradrenaline-displaceable [³H]idazoxan binding sites in membranes from human liver (Tesson et al., 1991) support to the first alternative. However, the fact that ₂C4 transcripts were found in human liver (Berkowitz et al., 1994; H. Paris and S. Schaak, personal observation), together with the fact that alpha-2 adrenergic receptors negatively coupled to adenylyl cyclase were identified in rat liver membranes (Hoffman et al., 1981; Jard et al., 1981), supports the second possibility. The application of the techniques used in the current work to freshly isolated human hepatocytes may bring a definitive answer to this question. According to previous observations made with cells transfected with the ₂C4 gene (Wozniak and Limbird, 1996) or with its mouse homolog M₂-4H (Von Zastrow et al., 1993), intracellular localization is a peculiar feature of the alpha-2C adrenergic receptor subtype. Whether this is true in HepG2 would also merit examination. Whatever the answers to these questions, HepG2 can now be considered as the first human cell line unequivocally demonstrated as expressing the alpha-2C adrenergic receptor. The future use of this model may yield valuable information on the mechanisms of regulation of alpha-2C receptor expression and ₂C4 gene transcription.

The presence of alpha-2 adrenergic receptors in SK-N-MC is less surprising, because other human neuroblastoma cell lines, such as SH-SY5Y (Kazmi and Mishra, 1989) and SK-N-SH (Baron and Siegel, 1989), were previously shown to exhibit alpha-2 adrenergic receptors. In addition to being a second model expressing the alpha-2C subtype, SK-N-MC has the very interesting characteristic of coexpressing the alpha-2A subtype, making this cell line a suitable system to investigate the parallel regulation of these two receptor subtypes in a single cell line of neural origin. According to estimations of the number of [³H]MK912 binding sites, the two subtypes are expressed in fairly similar amounts (20 ± 8 fmol/mg of protein for alpha-2A and 23 ± 3 fmol/mg of protein for alpha-2C). The results from the measurement of cAMP did not allow estimation of the respective contributions of the two receptor populations to inhibition of cAMP production, but it is likely that both subtypes participate in this process. In comparisons of the results obtained in the three cell lines, it is also obvious that there is no direct correlation between the amounts of mRNA for the different receptors and their respective levels of expression. Quantitative analysis of the data from RPA experiments indicates, for example, that the amount of ₂C4 mRNA in SK-N-MC is about 100-fold lower than in HepG2, whereas binding data show that the alpha-2C receptor number in SK-N-MC is only half that in HepG2. In the same manner, the amount of ₂C10 transcripts in HT29 is approximately 200 times higher than that in SK-N-MC, whereas the number of alpha-2A receptors is 10 to 15 times higher. The reasons for these differences were not investigated, but differences in the translation efficacy and/or in the degradation rate of the receptor proteins may be at the origin of this apparent discrepancy.

With the exception of HT29 cells, which contain a small amount of ₂C2 gene transcripts but express none or undetectable traces of the protein, our search for a human cell line expressing alpha-2B adrenergic receptors has been unsuccessful. In spite of this failure, surely due to the relatively limited number of cell types screened in the current study, our work provides two new human models, which will certainly be valuable systems for the study of homologous and heterologous regulation of the alpha-2C adrenergic receptor subtype in vitro.