European Journal oj Pharmacology 0

-

Molecular

Pharmacology

Sectivn, 227 (1Y92) 257-265

1992 Elsevier Science Publishers B.V. All rights reserved 0922-41Oh/92/$M.(lO

EJPMOL

YO3bll

K-Opioid binding sites on a murine lymphoma cell line Jean M. Bidlack, Lalitha D. Saripalli and Diane M.P. Lawrence Department

oj Pharnumk~gy,

The Unisersity of Rochester, School

of Medicine

and Dentistv?

Rochester, NY 14642, USA

Received26 May 1992, revised MS received 15 June 1992, accepted 30 June 1992

As a first step in dctcrmining whelhcr any subset of lymphocytes expresses opioid receptors, membranes prcparcd from mouse lymphoma cell lines were screened for [‘Hlnaloxone binding sites. Membranes from the R1.1 cell line specifically bound [‘Hlnaloxone. The Hill coefficient for [“Hlnaloxone binding was 0.93 & 0.18, and nonlinear regression analysis indicated that a one-site model was the best fit of the 13H]naloxone saturation binding data. Low concentrations of K-selective opioids, but neither p nor 6 opioids, inhibited [‘Hlnaloxone binding. Saturation binding studies with the K-SdCCtiVc compound [“H]Uh9,593 revealed a single binding site with a K D value of 0.204 + 0.039 nh4 and a B,,, value of 31.7 + 3. I fmol/mg of mcmbranc protein. The Hill coefficient for [JH]U69.593 binding was 1.03 f 0.11, indicativs of a single site, Time courses for the association p:nd dissociation of [‘H]U69,593 binding at 25°C exhibited properties consistent with a single class of binding sites. bw concentrations of K-SdeCtiVe opioids, including dynorphin peptides, inhibited [“H]U69,593 binding, while high concentrations of /L opioids were needed to inhibit binding, and the li-selective ligdnds were ineffective at concentrations up to 10 PM. Stereosclectivity of the binding site was demonstrated by the finding that the Ki value for ( - I-pentazocine in inhibiting [‘H]U69.593 binding was 25 times less than for the !+)-isovr. Based on its high affinity for U69,593, cu-nco-cndorphin, and dynorphin B. the K opioid binding site on R1.l cell membranes belongs Lo the K,,, subtype. As observed with brain K opioid binding sites, sodium inhibited [“H]U69,593 binding to RI.1 cell membrane: in a concentration-dependent manner. These data demonstrate that the murk lymphoma cell line R1.1 exprcsscs K opioid binding sites that are very similar to brain K opioid binding sites. Opioid receptors; R 1.I lymphoma cell lint; K-Opioid binding sites; [ “H]U69,593

1. Intrudustion

Pharmacologic evidence suggests that some cells of the immune system possess opisid receptors. In clinical studies, heroin addicts have heen shown to have elevated rates of infection (Cherubin and Millian, 1968; Sapira et al., 1980). After this initial observation, it was determined that lymphocyte mitogenenis was depressed in heroin addicts (Brown et al., 19741, and that opiate addiction produced a significant depression in the absolute number of total T lymphocytes in peripheral blood as measured by the ability of the lymphocytes to rosette sheep red blood cells (McDonough et al., 1980). Dpioids and opioid peptidrs a’ltcr many parameters of immunocompetence. The K-selective peptide dynorphin, as well as P-endorphin, increased superoxide produiiion in human polymorphonuclear leukocytes and peritoneal macrophages at peptide concentrations

CorrespondenceIO: Jean M. Bidlack, PII.D., Ekpaiiiiikiii trf Flikllilwcoloyy,The Universiryof Rochester.School of Medicine and Dcntistry, 601 Elmwood Ave., Rochcslcr, NY 14042, USA. Tel. i716) ZWWO; Fax (716) 244.Y?H3.

as low as lo-‘4 M (Sharp et al., 1985). p-Endorphin, dynorphin, and met-enkephalin have been reported to both enhance and suppress T-cell proliferation (Gilman et al., 1982; McCain et al., 1982; Fontana et al., 1987; Kusnecov et al., 1987; Hemmick and Bidlack, 1990) and natural killer cell activity (Mathews et al., 1983; Mandler et a)., 1986). In vitro antibody production in murine splenocytes was inhibited by p- and K-selective opioid agonists at concentrations as low as IO-“’ M, suggesting that cc.and K opioid receptors are involved in regulation of lymphoid cell production of antibodies (Taub et al., 1991). This latter study is one of the first to carefully examine the type of opioid receptor involved in the pharmacologic respon:,es to opioids. Opioid binding studies, using lymphocytes, have generally not addressed the type of opioid receptor on a certain population of lymphocytes. The binding of at least one radiolabeled opioid ligand to erythrocytes (Abood et al., 1976), mcnocytes (Lopker et al., 19801, granulocytes (Abood et al., 1976: Lopker et al., I980!. platelets (Mehrishi and Mills. I983), thymocytcs (ROY et al., 1991), and purified human T cells (Madden cl al., 19X7) has been reported. Opioid binding to these

mixed cc opioid alkaloids may be different than braintype opioid binding sites. Cell lines have been useful in the initial characfefization of ~-~nd~)~hin (Hazum et al., 1979) and opioid binding sites (Schweigerer. I98Sa,b; Fiorica and Spector, 108X: Parr et al., 1989). Cell lines offer the advantags of being a h~rn~~~en~)us~~pulati~~n, while even purified T-cell populations contain many different subsets of T cells at varying stages of maturity. Hazum et al. (I9791 described a noun-opioid ~-endorphin binding site on cultured B-cell-derived human lymphoid cells. A non-opioid P-endorphin binding site has also been found on EL4 murinc ~hyrn~~rnacells (Schwci~erer et al.. 19XSa.b). In addition to p-endorphin binding sites on EL4 cells, ( - I-[ “Hjbremazocine binding sites were described on this cell line (Fiorica and Spector, l988). A K,, value of 60 nM for ( - )-[‘H]bremazocinc binding to EL4 cells was reported in assays Fhat used whole cclis (Fiorica and Spector, 1988). The h--sclcctivc compound U-SO.488 was the most potent compound in inhibiting SO nM ( - 1-1‘H]hremazocine binding with an IC,,, vatuc of 0.57 PM. Both dcxtrorphan and Ievorphanol inhibited ( - )-[ ‘HJbremazocine binding equally well with IC,, values in the range of 1.9-2.9 PM, demonstrating a tack of stcreoselcctivity (Fiorica and Spector. l’:W. Using lysed cells from the macrophagc cell line, P388d,, Carr et al. (19X9) reported [ZHJU69,?93 binding to thcsc lyscd cells, with a K,, value of I7 nM. USO,488 inhibited 30 nM [‘H]U69,593 binding with an ISIS,, value of 8 nM. However, dynorphin and nidtrexone at final concentrations of IO FM inhibited less than 60% of the 30 nM (‘HjU69,SC)J binding to lysed F388d, cct!s fCarr et al., 1989). Using an isothiocyanatc derivative of USt),48X at a final concentration of 2W nM, Carr and colleagues re&!rted that the K binding site on the P388d, cells had an apparent molecular weight of 42,0(10, when separated on sodium dodecyl sulfate-polyacrylamidc gels under reducing conditions (Carr et al.. 19~11). ‘hilt a number of studies arc suggcstivc 01 opioid receptors On lymphocytes, opioid binding studies have not provided convincing data that opioid receptors do in fact exist on lymphocytes iSibinga and Goldstein,

I9XtO. The present study was undertaken to determine if brain-type opioid binding sites exist on any type of lymph~~te. Most previous studies have used whole cells when examining opioid binding. Due to the fact that sodium ions inhibit agonist binding to opioid rcceptors, examining agonist binding to whole cells is difficult. Aiso, with whole cells, care must be taken to avoid trapping of opioids intraccllularly (Maloteaux et al., l9S3, particularly with alkaloids, such as naloxone, that are quite lipophilic (Abood and Hess, 197%. In the present study, membranes prepared from murinc Iymphoma cell lines were screened for the presence of opioid binding sites, using [3H]naloxone as a nonselertivc opioid ligand to detect all types of opioid binding sites. Opioids selective for the different types of opioid receptors were then tested for their ability to inhibit [‘H]nal:txonc binding to membranes. The murine lymphoma cell line, RI .I, possesses K opioid binding sites, which were characterized using the K-selective compound 1‘HlU69,593.

2. Mate~als and methods 2. I. Marerids The murinc lymphoma cell line, RI.1 (TIB 42), was obtained from the American Type Culture Collection (Rockvillc, MD, USA). RPM1 1640 media was purchased from Gibco Laboratories (Grand Island, NY, USA). Iron-supplemented bovine calf serum was obtained from Hyclonc Laboratories (Logan, UT, USA). [‘HjNaloxone (60 Ci/rnmolJ and [%I]~~69593 (54 Ci/mmol) wcrc purchased from Amersham (Arlington Hcighls, IL, USA). Naloxonc and U50.488 were gifts from Endo Labs (Garden City. NY, USA) and Upjohn ~Kalarnaz~~~, Ml, USA), respectively. ICI 174,864 was obtained from Cambridge Research Biochemicals (Atlantic Beach, NY, USA). Norbinaltrophimine (norBNI) was purchased from Research B~ochemicals (Natick, MA, USA). The stcreoisomcrs of pcntazocine were obtained from the National lnstitute on Drug Abuse (RockviIle, MD, USA). Opioid peptides were obiaincd from cithcr Bachcm (Torrance, CA, USA) or Peninsula Laboratories (Belmont, CA, USA). Protease inhibitors and p~~lyethyleneimine were purchased from Sigma Chemical Co. (St. Louis, MO, USA). Glass fiber filter sheets No. 32 were purchased from Schleichcr & Schuell (Keene, NH, USA). Ecoscint A scintillalion !lclid was purchased from National Diagnostics (Manville, NJ, USA).

R 1.1 cells were culturctt to confluence in 5% CO, at 37°C in RPM1 l64iI mcditim. buffcrcd with I23 mM

HEPES, pH 7.2, and containing 300 pg/mi I-glutamine, 100 units/ml penicillin, 100 Fg/ml streptomycin, 60 PM 2-ctharlolamine, and SO PM 2mercaptoethanol. The medium was augmented with 10% bovine calf serum, iron-supplemented. Cells were collected and counted in a hemacytometer in the prcsence of trypan blue, to determine viability. RI.1 cells were centrifuged at 200 X g for IO min at 4°C. The supernatant was removed, and the cells were resuspended in 50 mM Tris-HCI, pH 7.5, followed by homogenization of the cells with a Polytron homogenizer at a setting of 4 for I5 s. The membranes were centrifuged at 39,000 X g for 20 min at 4°C. The membrane pelt&s were resuspended in 50 mM Tris-HCI, pH 7.5. and the ccntrifugation step at 39,000 X g for 20 min was repeated. The membranes were finally resuspended in 50 mM Tris-HCi, pH 7.5, at a protein concentration of 7-13 mg/ml. Membrane protein concentration was determined by the method of Bradford (1976) with bovine serum albumin as the standard, The membranes were tither used immediately or were frozen at - 80°C. Freezing membranes at -80°C did not alter the properties of the [~H]U69,593 binding !AC.

In a final volume of 0.5 ml of 50 mM Tris-HCI, pH 7.5, 0.3 mg of RI.1 membrane protein was incubated with 0.30-20 nM [3H]naloxonc at 25°C for 60 min. Nonspecific binding was measured by the inclusion of either 10 PM naloxone or 10 FM U50,488. Membranes were filtered through Schieichcr & Schuell No. 32 filters, followed by washing 3 times with 4 ml of ice-cold 50 mM Tris-HCI, pH 7.5. The filters were counted in 2 ml of Ecoscint A scintillation fluid. The IC& values for the inhibition of [‘Hjnaloxonc binding to RI.1 membranes were determined with the h-selective peptidc [D-AlaZ,0vle)Phe4,Gly(ol~s]enkcphal~n ~DAMG~j~ the &preferring pcptides ICi 174,864 and fD-Ser2,Leu”,Thr”]enkcphalin (DSLET), the K-setective campounds U50,488 and U69,593, and ( -t I- and ( - ~-pc~tazocjile. R1.1 cell memb~ncs were incubated with I2 different concentrations of each l&and in the prcscncc of 2 nM [‘H]naloxone at 25°C for 60 min before f~ltcring.

at 25°C for 60 min. Nonspecific binding was measured by the addition of 10 I.LM naloxone. Protein linearity of binding was determined by varying the protein conLentration from 0.012 to 0.5 mg of membrane protein in the presence of 0.4 nM [“H]U69,SY3. The association of [~H]U69,593 with nicmbranes was dctcrmined by incubating membranes with 0.4 nM [‘H]U69,593 for l-120 min at 25°C before filtering. The dissociation rate for 0.4 nM [~H]U69,S93 from membranes was determined by incubating membranes with [ ‘H]U69,593 at 25°C for 60 min prior to the addition of 1 PM U50,488. Membranes were then incubated for 2-120 min before filtering. IC,,, values for the inhibition of 0.4 nM I’H]U69,593 binding to 0.3 mg of R1.1 cell membrane protein were determined as described above for the [“Hlnaloxone experiments. When the dynorphin peptides were tested, membranes were incubated in polypropylene tubes with 30 PM bestatin, 250 FM L-leucyl-L-leucine and 300 FM captopril, in addition to the dynorphin peptides, in order to minimize peptide degradation. When human p-cndorphin-(l-31) was tested, 100 pg/ml bacitracin was included in the incubation. To determine if sodium would aiter [%$.X9,593 binding, NaC! at final concentrations ranging from 0 to 260 mM was incubated with membranes and 0.4 nM [~H]U69,~93. 2.5. Analy.s& of binditrg data Saturation data were analyzed by nonlinear regression analysis using the LIGANR program (Munson and Rodbard, 19800). IC,,, values were calculated by least squares fit to a logarithm-probit analysis. Ki valucs of unlabeled compounds were calculated from the equation Ki = (IC&)/l + S where S = (concentration of radioligand)/(K, of radioligand) (Cheng and Prusoff, 1973). The association rate was analyzed by plotting in[Bc/G& - B,ij versus iiiiiC, whcrc B, is the amount of [‘H]U69,593 specifically bound at cquilibrium, atld B, is ihc amount of [“H]U69,593 specifically bound at time t. Dissociation rate constants were detcrmincd according to the equation B, = B,,c’-~ I” and k_ , determined as the slope of the plot of In(B,/B,,) versus time. Statistical comparisons were made using Student’s t-test.

3. Results [“H]W69,593 binding to membranes was measured as described for [‘H]naloxotic binding with the cxception that the glass fiber filters were soaked in 0.25% polycthylcnciminc for a Icast 60 min before USC. To dcterrn~~~e the K,, and B,,,,,, values for [‘H]ki69,593 binding to 1111.1 membranes, 0.3 mg of mcmbranc protein. was incubated with 0.025-3.2 nM 6’H]lJ69,593

As shown in fig, I, RI.1 cell membranes specifically bound [ “HJnaloxonc. 0ther ccl1 lines that WCexamined bound some but Icss [‘H]naloxonc than the RI .l cell mcmbrancs. Low levels of ~~H]nalox~~ne l?indin~, displaced by III PM naloxone, were seen with most cells

PM naioxone. nonlinear regression analysis of the data indicated binding to a singic site with an average K, value of 3.2X + 0.32 nM and a B,;,, vaiuc of 64.” f 11.4 fmoi/mg of jjrotein. While both the K,, and B,;,, values wcrc slightly greater than those obtained when naioxcilc was used to mcasurc nonspecific binding, the difference was not statistically significant. The Hill CcJefficientwas 1.08 + 0.007 when U50,488 was used as the displacer.

or0 .

B (fmol/mg ’ 2

*

’ 4

.

protein) 8

8

FREE [‘H]NALOXONP

(nM1

Fig. I. [ ‘HjNdoxonc hiding IO RI.1 cell membranes. RI.1 cdl membranes. II.3 mg of memhranr pwtein. were incubated with Il..W-20 nM [ ‘~ijnaloxone for ho min at 25°C in il final volume of 0.5 ml of 50 mM Tris-HCI. pH 7.5. Nonspecific hindinp, which was always less than 2OC of toId binding, was mcasurcd by inclusion of IO PM nlrloxone. Points reprcsenl the mean fmol of [ ‘~ljnaloxone hau-d/my. of membrane protein +_S.E.M. from three cxpcrimcnts pcrbrrncd in Iriplicatc. The inset is a ScaIchard pltr( of ;I reprcsemstivr cxperimcnt. Nonlinear regression analysis indicated that ;I onesiIr model WB~Ihe best fit of Ihe &da. having an average K,, value of I.40 f O.IH nM and an average B,,.,, value of 47.6 f 6.7 fmoI/mg of protein.

tested as has been observed by other investigators (Madden et al.. 1987; Ovadia et al., 1989). Nonlinear regression analysis of [ZH]naloxonc saturation binding data resulted in a one-site model being the best fit of the data. and showing an average K, value of 1.40 + 0.90 nM and an average B,;,, value of 47.6 + 6.7 fmoi/mg of protein. The Hill coefficient for [ ‘Hlnaioxone binding was 0.93 f 0.18. indicative of a single binding site. I2.

Sekridy

of I -‘Hlnaloxone

3.3. l.‘H]U69,593

binding to AI. I cell membranes

The specific binding of [ ‘HlU69.593 to Rl.1 ceil qlcmbranes, shown in fig. 2. rcachcd saturation at 1.5 nM [-‘H]U69,593. Scatchard analysis of [>H]U69,593 binding revealed a linear plot, with an average K, value of 0.204 &-0.039 nM and a B,;,, value of 31.7 f 3.1 fmol/mg of membrane protein (fig. 2). The Hill cocfficicnt for [ “H]U69,593 binding was I .03 + 0.11, suggesting that [JH]U69,593 binding was to a single site. In the preparation of membranes from brain, membranes are usually incubated at 37°C for 30 min to dissociate cndogcnous opioid peptidcs from the membranes (Pastcrnak et al.. 1975). Incubating RI.1 ccl1 mcmbranes at 37°C for 30 min, after the first centrifugation ot 39,000 x g did not increase [ “H]U69,593 binding (data not shown). Conscqucntiy, a 3PC-incubation step was not included in the membrane preparation protocol. [‘H]U69,593 binding to RI.1 membranes increased iincarly with increasing protein, at protein levels ranging from 12 to 500 pg. as demonstrated in fig. 3. The time course for the association of 0.4 nM I’HlU69.593 with 0.3 mg of Rl.1 ceil mcmbranc protein at 25°C showed rhat equilibrium was reached by I5 min, as depicted in fig. 4. The binding rcmaincd con-

binding TAB!_!Z !

To determine if any of the [‘Hlnaioxone binding sites on Rl.1 ceil membranes resembled brain-type opioid binding sites, opioids specific for CL, S. and .*’ receptors were tested for their ability to inhibit [‘Hlnaioxone binding. Table 1 shows that the CL-seiective pcptidc DAMGO had a K, value of almost 700 nM. while the 8-seiective peptides ICI 174,864 and DSLET were virtually ineffective at inhibiting [‘HInaioxone binding to RI.1 ceil membranes. In contrast, the K-S&CtiVC compounds USO,488 and U69,Ci)3 had K, values of less than 1 nM. Stereoselectivity of the [‘Hlnaloxonc binding site was demonstrated by the finding that ( - I-pentazocine was 14 times more potent in inhibiting binding than the ( +)-isomer. These resuits suggested that the [‘Hlnaioxonc binding was to a ~-type opioid binding site. When the [“Hlnaioxone saturation experiments were repeated using 10 PM USO.4X8 lo mcasurc non!;pccific hinding instead of 10

K, vdues and I lill coefficients of opioids against hiding to RI.1 cell membranes.

[“I I]naloxone

Rl. I cdl mcmhranes. 0.3 mg of mumbr;Inc protein. wcrc indated with I? different conccnIraticrns of each opioid and 2 nM [“lljnaloxone in ;1 final volume of II.5 ml iIf SO mM Tris-l1Cl. ptl 7.5, a~2S”C for 60 min hcforc filkring Ihc s;lmpics w glass fihrr filters. 1(‘5,, ~ahes wc’rc oht;Gncd from line;lr regression analysis assuming il single de and were converted to K, vduex, using the K,, value of I.4 nM for I ‘Hjn;Iloxone hiding. D;IIiI iIre presenkd as the mean Ki value&S.E.M. and the mean Hill coefficienI f S.E.M. ohtaincd from three experiments, performed in trip!icale. Opioid

K, (nM)

Hill coefficient

GO.4XH

0.45 i- O.Oh 0.72t O.Ob I I.3 t 4.0 15X i- 24 1,‘)11 i 104 > 4,IIIHI > 4.IWHl

O.H4f 0. IO O.HO + u.i:7 I .05 f O.ilH 0.00 f O.WI o.Y2 %0.07

UbY.SY.1 ( - Wenl;~zocinc ( f )-PClltilU~inc DAMtiC) DSLEI’ ICI I74,XO4

6

2 kii q j

6

4

-I

40 B (fmol/mg I 1

0 0

2

FREE [aH]UB9,5133

protein) 3

2

0

(nld)

0

Fig. 2. [‘H]U69,593 binding to R1.1 cell membranes. In a final volume of 0.5 ml of SO mM Tris-HCI, pH 7.5, 0.3 mg of R1.l cell memhrane protein was incubated with 0.025-3.2 nM [‘H]UbY,SY3 for 60 min at 25°C before filtering samples rhrough glass fiber fibers that had been soaked in 0.25% polyethyleneimine. Nonspecific binding, which was always less than 30% of total hinding, was measured by inclusion of 10 /IM naioxone. Points represenl the mean fmol oi [ ZH]UW.SY3 bound/nig of membrane protein f S.E.M. from four experiments performed in triplicate. The inset is a Scatchard plot of a represenlativz experiment. A one-site model best fit the saturation data, giving an average K n value of 0.204 f 0.039 nM and an average I&,, value of 31.7 k 3. I fmol /mg of membrane protein.

stant for at least 120 min. To determine the dissociation rate of [‘H]U69,593, membranes were first incubated with 0.4 nM [-‘H]U69,593 for 60 min at 25°C. U50,488, at a final concentration of 1 PM, was then added, and incubated with the membranes at times ranging from 2 to 120 min. Fig, 5 shows that between 10 and 15 min after the addition of U50,488, half of the specifically bound [ “H]U69,593 had dissociated from the binding site. An incubation time of 100 min was necessary to obtain complete dissociation of [“HI-

I

I

I

20

40

60

TIME (min) I I 80

100

I 120

TIME (min) Fig. 4. Association time course for [ ‘H]Uu9.593 binding to RI.1 cell membranes. R1.l cell membranes were incubated with 0.4 nM [.‘H]Uh9.593 for time intervals ranging from I to I20 min before filtering the samples. as described in the legend to fig. 2. Points represent the mean fmol of [‘H]UhY.S93 bound/O.3 mg of membrane protein f S.E.M. from three experiments performed in triplicate. The inset is a representative experiment.

U69S93 from the binding site. Kinetic analysis of the dissociation data resulted in a linear plot of InfB,/B,,) versus time, suggesting that [ ‘H]U69,593 was dissociating from a single binding site. 3.4. Selecririty

of l~7HIU69,593 bindirtg

Ki values, shown in table 2, for the inhibition of [‘H]U69,593 binding to RI.1 membranes by opioids selective for the different types of opioid receptors

0

TIME

(min)

0 0

20

46

60

60

100

126

TIME (mia) 0

0

100

200 PROTEIN

300

400

600

(~6)

Fig. 3. 1’H]UhY,SY3 binding to varying amounts of RI.1 cell mcmhrane protein, RI.1 cell membrane protein. at levels ranging from I2 to SO0 gg. wcrc i~~ctba!cd wirh 0.4 nM [’ H]Uf)Y,SY.l as doscribcd in the lcgcnd to fig. 2. Points rcprcsrnt the mean fmol of [ ‘H]lJW).SY.l hound t S.E.M. from five experiments pcrformcd in triplicate.

Fig. S. Dissociation time course lor [.‘H]UhY.SY3 binding to RI.1 cell membranes. RI.1 cell membranes were incubated with 0.4 nM [“H]Uh9,SYd for h0 min ;II 2S”C before the addition of I MM USO,48X. Samples were filtered at time inlcrvals ranging from 2 to I20 min after the addition of USK48H. as described in the legend 10 fig. 2. [ ‘H]UhY.SY3 binding to membranes in the iIhsenWof USO.4XX was rcgardcd as control hinding. Poinls rcprcscnt tbc mean pcrccnle EM. from l!?rre expcrilWnl\ ilgC (11 COllllUl 1’1 ijiif+I.5’13 liifiiiiilg I Y, performed in Iril)licalc. The inbcl is a rcpresenlative expcrimcnt.

\vepe d~t~~~i~~d in o&r to further characterize the u~Y.SY_3binding site. The K-sclectivc compounds .W8 and nor-RN1 inhibited [ ‘HlUhY.SY3 binding K, values of kss than 1 nM. The Hil! coefficient for the i~l~~bit~~~~ of ~-~H]~~6Y,SY3 binding by USO.488 was (1.91 k 0.10. suggesting binding to a single site. The Cc-sckctive pcptide DAMGO had a K, value greater than 200 nM. while the &selective antagonist ICI 17.t.X64.at onccntrations up to 10 PM. did not inhibit of [ ~H]U6Y.SY~ birding. Since high concentrations DAMGO inhibited ~~H]U6Y,SY3,the membranes were examined for their ability to bind [‘HIDAMGO over the concentration range of 0.25-32 nM [“I-!]DAMGO. No specific [“H]DAMGO binding was detected (data not shown). Therefore. the inhibition seen with high DAMGO ~o~ccntrations prob~bIy represents DAMGO binding to the K opioid binding site. The stereoselectivity of the binding site was addressed by determining the K, values for the inhibition of [-‘H]UhY.SY3 binding by the stereoisomers of pentazocine. The ( - I-isomer of pentazocinc had a 25-fold lower K, value than the (+ tisomer, demonstrating that like brain opioid receptors, this K opioid binding site also was stcreroselectivc for the ( - J-isomer (table 2).The Hill coefficient for the inhibition of [‘H]tJ6Y,SY3

TABLE

2

K, values for the inhibition of [‘llJlJf~Y,SY3 hinuing to RI.1 cell membranes. In a final vrllumr of 0.5 ml of 50 mM Tris~Hfl. pH 7.5. RI.1 ccl1 membranes. 0.3 mp of memhranc pratein. wcrc incubated for Ml min at ‘SC with 0.4 nM ~‘H]UhY.S93 and It c(~ncentrations of the opkids. as dekhed in Materials ;md methods. If?,, values were obtained from linear regression analysis assuming it single site and were ctrnverted IO K, values using the K,, value of 0.204 nM for [ ‘HjlJhY.5Y3 hindinp. Data ar+’ pre\ented as the me;m K, value f S.E.M. obtained from at least three experiments. performed in trig+ztttt. Opiold

K, (nMl

K

uswnx t - I-Pentazocine t + I-Pentazclcinr p and ri DA MGO Morphine DSLET An!qqmistr nor-l&N1 Naloxttnr ICI 174.KM Endogrnous peptides cr-Nco-endorphin Dynorphin A-( I - 13) Dyntrrphin B PI,-EndorphW I-31 f Dynorphin A-(2- I?)

0.35 +0.13 2.91 +o.o.i 73.3 +4.7 ?I4 24.‘) > 3.lHHl

+4 f 1.2

0.07fl + o.tll4 2.4 I + 0.42 > K.lXD 0.23 I.25 0.23 30.0 > 300

f WIN ~0.1s f ll.IlX t 2.3

100

-

0 0.1

10

1 NaCl

100

(mM)

Fig. h. Effect ttf NaCl on [~H]U~Y.S93 binding. RI.1 cell m~mbrunes were incubated with 0.4 nM ~~H~U~~.S~~ and concentraticms of NaCI. varying from 0 to 260 mM. as described in the legend to fig. 2. Points arc the mcun percentage b0undkS.E.M. from three cxperiments performed in triplicate.

binding by aif opioids was near unity (range 0.86-1.21, suggesting a single binding site, Dynorphin peptides also inhibited [-‘H]U69,593 binding to R1.l cell membranes as shown in table 2. Of the dynorphin peptides tested, a-neo-endorphin, with a K, value of less than (1.3 nM, had the highest affinity for the binding site. Dynor~hin A-(1- 131 and dynorphin B had equal affinities for the site, while &,-endorphin-(I -31) had a considerably lower affinity for this [“H]U69,593 binding site. Concentrations of dynorphin A-(2-171 up to 1 PM were ineffective at inhibiting [~H]U69,S93 binding. As with brain K opioid hinding sites, the N-terminus of the peptide is critical for the binding of the dynorphin pcptides to RI .I cell membranes.

Since sodium ions inhibit K opioid binding to brain membranes (Mack et al., 1985; Clark et al., 1988: Smith et al., 1989), the ability of NaCl to inhibit 0.4 nM [‘H]U69,593 binding to RI.1 cell membranes was investigated. As demonstrated in fig. 6, NaCL inhibited ~‘H]U6Y,593 binding in a conccntratjon-dependent manner. Half-maximal inhibition of binding was obtained with I9 f 0.5 mM NaCI, and complete inhibition was reached with 260 mM.

4. Discussion The study reported here has shown that the murine lymphoma cell line Rl. 1, expresses K opioid binding sites, similar to tc binding sites observed in brain (Koster~itz et al., 1981; Van Vo~~tlander et al., 1983;

Robson et al., 1984; Lahti et al., 1985; Smith et al., 1989). No~l~inear regression analysis of both [.‘H]naJoxone and [‘H]U69,593 saturation data revealed a one-site model was the best fit. In the [XH]naloxone saturation studies, neither the K, nor Blttilrvalues were significant!y different regardless of whether 10 PM naloxone or 10 PM 1J50,488was used to measurz nonspecific binding. The B,,, value for [‘Hjnaloxone binding was approximately 50 fmol/mg of protein, while [3H]U69,.593binding studies resulted in an average B,,, value of 32 fmol/mg of protein. These two B,,, values were not statistically different, suggesting that both ligands were labeling the same site. This K opioid binding site had approximately a 7-fold greater affinity for [“H]U69,593 than for [“Hlnaloxone, The [?HfU69,593 binding site exhibited properties that were essentially identical to those observed with brain K binding sites, namely, saturabihty, protein linearity, stereoselectivity, and displacement by low concentrations of K-selective opioids. The density of the K: opioid binding sites, 30-50 fmolfmg of protein, is low in comparison to guinea pig brain, where B,,, values greater than 100 fmol/mg of protein have been reported ~Robson et al., 1984; Rothman et al., 1990). However, the affinity of the it site for (“H]U69,593 was very high and with a Hill coefficient near unity, [“H]U69,593 appears to be binding to a single site on the RI.1 cells, while multiple IC opioid binding sites have been postulated to occur in brain membranes (Mock et al., 1988; Zukin et al., 1988; Clark et al., 1989; Price et al., 1989; Rothman et al., 1990; Kinouchi and Pasternak, 1991). [~H]U69,593 binding to RI.1 membranes was only inhibited by low concentrations of K opioids. The F-selective peptide DAMCO inhibited [“H]U49,593 b’m6’mg with a Ki value of greater than 200 nM, while the &-selective peptide ICI 174,864was ineffective at concentrations up to 10 NM (table 2). The p-prefer&g alkaloid morphine exhibited a Kr value of 25 nM in inhibiting [~H]U69,593 binding. Thus, while morphine exhibits preference for the @ opioid receptor, it crosses over to K binding sites. Naloxane had good affinity for the [‘H]U69,593 binding site, having a Ki value that was IO-fold less than morphine and only 6-fold greater than U50,488. Thus, while naloxone has some preference for the p binding site, it also recognized the [‘H]U69,593 binding site. Of all the compounds tested, nor-BNl had the highest affinity for the [“H]U69,593 binding site. Nor-BNI also binds with high affinity to brain K opioid binding sites (Fortoghese et al., f987). The stereosel~ct~vi~ of the binding site for the (-)-isomer was demonstrated bY the finding that (- )-pentazocine had a 25fold lower Ki value than ( -I- )-pentazocine for the inhibition of [3H]U69,593 binding, There is increasing evidence for multiple K opioid binding sites in brain (Neck et al., 1988; Zukin et al.,

1988;Clark et al., 1989; Rothman et al., 1990;Kjnouchi and Pasternak, 19%). The existence of K, (U69,593sensitive) and K~ tU69,593-insensitive) sites in brain has been postulated (Neck et al., 1988; &kin et al., 1988). The ‘(I subtype of opioid receptor has been divided into K,~ and K,~, based on their affinity for dynorphin B and a-neo-endorphin (Clark et at., 1989). Both K~. and K,,, sites bind dynorphin A with high affinity. However, only K,~ sites have high affinity for a-neo-endorphin (Clark et al., 1989).The ic2site, which is less sensitive to U50,488, has also been further been subdivided into ~~~ (~-neo-endorphin-insensitive) and ~~~~~-neo-endorphin-sensitive) ~Rothman et al., 1990). Additional K opioid binding subtypes have been pro.. posed (Clark et al., 1989; Price et al., 1989; Rothman et al., 1990). The Rl.1 K binding site exhibited high affinity for U50,488, which had a Ki value of 0.35 nM for the inhibition of [“HJU69,593 binding to membranes (table 2). U50,488 completely inhibited all of the [“HjU69,593 binding. Taken together, these data suggest that the K opioid binding site on R1.l cells belongs to the K, subtype. Based on its high affinity for ru-neo-endorphin, this site appears to belong to the I(,,., subiype of K) opioid binding sites, based on the nomenclature of Clark et al. (1989). While opioids have been reported to alter a number of parameters of immunocompetcnce, K opioids have been shown to suppress in vitro antibody production (Taub et al., 1991) and to stimulate superoxide production in human polymorphonuclear leukocytes and peritoneal macrophages (Sharp et al., 1985). K-Selective opioids at concentrations as Iow as lo- *” M inhibited antibody production, while superoxide production was stimulated by K-seicctive opioids at concentrations as low as W-t4 M. It has been difficult to understand how te opioids could exert effects at concentrations below 1 nM, but binding studies have reported needing IO-# M and greater concentrations in order to observe any specific endings and even at these concentrations, binding has not been displaced by all opioids (Fiorica and Spector, 1988; Sibinga and Goldstein, 1988: Carr et al., 1989, 1991). The study reported here is the first to show high affinity binding of K-selective opioids, including ~ndogenous opioid peptides, to membranes prepared from a lymphoma cell line. Preliminary data suggest that this K opioid binding site is negatively coupled to adenylyl cyclase. Additional studies arc needed to determine if this binding site can be detected in freshly isolated lymphocytes. The R1.i lymphoma was derived from a thymoma which arose spontaneously in a C58/J mouse (Ralph, 1973; Wyman and Stallings, 1976). This lYmphoma expresses antigens characteristic of thymocytes, suggesting that the K opioid binding site, described in this study, may be expressed on tl~ym~lcYtes.Pools of thymoeytes, splenocytcs, and even purified T cells,

c\Intilin many JiffCrcnt subsets of T CCllS at varying stages of maturity. Probably only a small subset of Cells expresses K vpiuid binding sites. Consequently, they will be difficult to detect with~~ut extensive fractionation of lymphoqtic populations. On RI.1 ceils, the K opioid binding site was present at a concentration of M-50 frn~~~~rngof membrane protein. This cell line is ~omoge~ous. if this cell line was diluted onty d-fold by cells that did not express K opioid binding sites, the specific binding of [3H]Uh9,S93 to this mixed population of cells wouId be difficult to detect. Thus, extcnsive ce~~u~arpurification protocols will undoubtedly be necessary in order to characterize K opioid binding sites on fresh lymphocytes. However, lymphoma ceils. such ;1s the RI.1 cells, offer the advantage of studying K opioid binding sites from a homogenous source. Like the neuroblastoma x glioma hybrid cell line, NG IWlS, that only expresses 6 opioid receptors, this cell line will also he valuable to in~~estjgatorsinterested in studying K binding sites in cells that do not express other types of opioid receptors.

Acknowledgements This work was supported hy I!.% Puhiic Health Service Cirants DA%?SS and DAII?X! from the National fnstitutc on Druy Abuse and hy ihe Margo Clcvciand Fund.

References ~hood. L.G. and W. Hess. lY75. Sfervospecific morphine adsorption Io phosphafidylserine and other memhrani~us et~mponents of brain. Eur. J. Fharmid. 32. hh. Ahood. L.C.. H.G. Atkinson ::nd M. MacNeil, 1976. Stereospecific opiate binding in human erythrocyte membranes changes in heroin addicts. J. Neurosci. Res. 2. 427. Bradford. MM.. 1976, A rapid and se&live method for Ihe quantitation 0f microgram quantifies oi protein utilizing the principle of proIein.Jye binding. Anal. Biochem 72. 248. Brown. SM.. B. Stimmel. R.N. faub. S. Kockwa and R.E. Roscnfield. 197-t. lmmun~~i~~~icdys~uncfi~)n in heroin addicts, Arch. Intern. Med. 135. 1001. Carr. D.J.J.. B.R. DeCosta. C.-H. Kim, A.E. Jacobson, V. Guarccllo. K.C. Rice and J.E. Blalock. IYtfY. Opioid receptors on cells of the immune system: evidence for & ;Ind n-classes. J. Endocrinol. 122. 161. Cbrr. D.J.J.. B.R. DeCosfa. A.E. Jacobson, KC. Rice and J.E. Blalock. 1991. Enantioselecfive kappa opioid binding sites on the macrophage crll line, P3X8d,, Life Sci. 4Yi 45. Cheng. Y.C. and W.H. Prusoff, 1973, Relationship between the inhibific~n constant IK, f and the concentration of inhibitor which causes 50 per cent inhihiIion (I