Carcinogenesis vol.15 no.9 pp.1807-1813, 1994

Enhancement of liver cell gap junction protein expression byglucocorticoids

Ping Ren, Adriaan W.de Feyter1, David L.Paul2 andRandall J.Ruch3

Department of Pathology, Medical College of Ohio, Toledo, OH 43699,'Meridian Instruments, Inc., Okemos, MI 48864 and ^Department ofNeurobiology, Harvard Medical School, Boston, MA 02115, USA

'To whom correspondence should be addressed

Gap junctional intercellular communication (GJIC) andthe expression of gap junction proteins (connexins) may beinvolved in growth regulation and neoplastic transforma-tion. The mechanisms of connexin gene regulation in normaland neoplastic tissues are poorly understood. In this study,the glucocorticoids, dexamethasone and hydrocortisone,enhanced fluorescent dye-coupling in primary cultured rathepatocytes and MH,Ci rat hepatoma cells. Other typesof steroids (fj-estradiol, testosterone, aldosterone and pro-gesterone) had no effect Northern blot, Western blot,nuclear run-on and immunohistochemical analyses showedthat glucocorticoids enhanced the expression of connexin32in these cells in a dose- and time-dependent fashion. Con-nexin26 expression was also enhanced slightly by dexa-methasone in hepatocytes, but not MHiCx cells. Connexin43expression in these cells was not affected by steroids. InWB-F344 rat liver epithelial cells, which were highlycoupled and expressed high levels of connexin43 and nodetectable connexin32 or connexin26, dexamethasone hadno effect on coupling or connexin expression. These resultsindicate that dye-coupling and the expression of connexin32and connexin26, but not connexin43, were upregulated byglucocorticoids in a cell-specific manner. These effects onGJIC and connexin expression may be involved in theinduction of hepatic differentiation and inhibition ofgrowth.

IntroductionGap junctional intercellular communication (GJIC*) has beenimplicated in the regulation of cellular growth and expressionof the neoplastic phenotype (reviewed in 1-4). Gap junctionalchannels may serve as pathways for the direct cell-to-cellexchange of growth and phenotypic regulatory factors.

Studies have shown that most neoplastic tissues have fewergap junctions compared to their non-transformed counterparts(1). This defect appears to be due to reduced expression ofgap junctional proteins (connexins) as well as altered post-translational phosphorylation and assembly of junctionalchannels (5-8). For example, in many rodent and human liverneoplasms, the number of immunohistochemically detectablegap junctions and the levels of the predominant hepato-cyte connexin, connexin32 (9), were reduced compared tosurrounding parenchyma (10-17).

•Abbreviations: GJIC, gap junctional intercellular communication; GRE,glucocorticoid response element.

Little information exists about the mechanisms regulatingconnexin expression in normal and neoplastic cells. Twelverodent connexins have been cloned and they are expressed ina cell-, tissue- and development-specific manner (reviewed in18). Several agents or treatments have been reported toaffect connexin expression. Cyclic AMP agonists increasedconnexin43 expression in flbroblasts (19) and connexin32expression in primary cultured hepatocytes (20,21). Retinoidsand carotenoids enhanced connexin43 expression in fibroblastsand fibrosarcoma cells (22-24). Phorbol myristate acetateincreased connexin26 expression in primary cultures ofmammary epithelial cells (8). Primary culture of hepatocyteson plastic dishes led to the loss of GJIC and down-regulationof connexin32 and connexin26 (25,26). Co-culture of thesecells with non-parenchymal cells (27,28) or the addition ofextracellular matrix components to the culture medium (25,29)enhanced GJIC and connexin32 expression. The steroid hor-mones estrogen and progesterone, along with prostaglandinsand cyclic AMP, appear to regulate connexin43 expression inuterine myometrium (30-32). Testosterone has been reportedto enhance connexin32 expression in spinal cord motor neurons(33). As these examples illustrate, many factors modulatethe expression of connexin26, connexin32 and connexin43,although the molecular mechanisms are not known. Littleinformation exists, however, about the regulation of otherconnexin genes.

We have reported that GJIC of primary cultured rat andmouse hepatocytes was enhanced by culture in the presenceof dexamethasone and hydrocortisone but not other types ofsteroids (34). The mechanism(s) for this effect were notidentified, however. In the present study, we have examinedwhether these actions of glucocorticoids were due to enhancedconnexin32, connexin26 and/or connexin43 gene expressionand if the response was limited to hepatocytes. Given thelikely role of down-regulated connexin expression in neoplastictransformation (1—4), it is important to define the mechanismsregulating the expression of these channel-forming proteins.

Materials and methodsReagents

Leibovitz's L-15 medium, RPMI 1640 medium, Dulbecco's MEM mediumand HBSS were purchased from GIBCO/BRL (Grand Island, NY). Richter'sImproved MEM medium was supplied by Irvine Scientific (Santa Ana, CA).FBS was obtained from Hyclone Laboratories (Logan, UT). TRI-REAGENT™was supplied by Molecular Research Center, Inc. (Cincinnati, OH). Connexin43and connexin26 full-length cDNAs (1.6 and 1.1 kb respectively) were providedby Dr E.Beyer (Washington University) (35) and Dr B.Nicholson (StateUniversity of New York at Buffalo) (36) respectively. A full-length connexin32cDNA and polyclonal rabbit antibody to connexin32 amino acid residues98-124 (LOLA antibody) have been described previously (9). Other reagents,including steroid hormones, were purchased from Sigma Chemical Co. (StLouis, MO) unless otherwise indicated.

Rat hepatocyle isolation and culture of liver cellsHepatocytes were isolated from male Fischer F344 rats (150-200 g) by two-stage portal vein perfusion with collagenase D (Boehringer Mannheim,Indianapolis, IN) as described previously (34). Primary adult rat hepatocyteswere plated at a density of -5X104 viable cells/cm2 into 100 mm plastic

© Oxford University Press 1807

at University of C

alifornia, Santa Cruz on D

ecember 2, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

P.Ren et aL

tissure culture dishes in L-15 or RPMI 1640 medium supplemented with 10%FBS and 50 (ig/ml gentamicin (complete media). L-15 medium was alsosupplemented with D-glucose (1 g/1). Culture dishes were not pretreated withcollagen or otheT extracellular matrix components. The cultures were refedwith fresh medium 2 h after plating and every 24 h thereafter and wereincubated in humidified, 100* air (L-15) or 95% air/5% CO2 incubatorsat 37°C.

The MHjC] cell line was originally established in vitro from a rathepatocellular carcinoma (Morris hepatoma no. 7795) (37). MH|C| cells werepurchased from the American Type Culture Collection (Rockville, MD) andwere cultured in Dulbecco's MEM supplemented with 10% FBS and 50 ng/ml gentamicin at 37°C.

Non-transformed rat liver epithelial cells (WB-F344 cells) (38) wereobtained from Dr J.Gnsham (University of North Carolina). The cells wereutilized between passages 20 and 27 and were cultured in Richter's ImprovedMEM supplemented with 5% FBS and 50 Jlg/ml gentamicin at 37°C asdescribed previously (39).

Treatment of cells with steroid hormonesAfter 1 day in culture, hepatocytes were refed with supplemented L-15or RPMI 1640 medium and treated with one of the following steroids:dexamethasone, hydrocortisone, [5-estradiol, testosterone, progesterone oraldosterone. The steroids were prepared in absolute ethanol as 10 000-foldstock solutions and applied to the cells in volumes of 0.1 nl/ml of culturemedium. Control cultures received equivalent volumes of ethanol. MH|C|and WB-F344 cells were treated in a similar manner at culture densities of75-80%.

Dye-coupling microinjection assay of GJICDye-coupling was assayed in hepatocytes and other liver cells as a quantitativeindicator of the relative levels of GJIC as described previously (39). Thefluorescent dye, Lucifer Yellow CH, was loaded into one cell by microinjectionand iontophoresis and the frequency of dye spread to directly adjacentneighboring cells was determined 5 min later. Ten cells per dish weremicroinjected and the percentage of dye-coupled adjacent cells was determined.Three or four dishes were assayed per treatment condition.

RNA isolaton and Northern blot analysis of connexin mRNAsLevels of connexin32, connexin26 and connexin43 transcripts in hepatocytes,MH,C| cells and WB-F344 cells were analyzed by Northern blotting asdescribed previously (39). Hybridization buffers contained 1.5-2X10* c.p.mJml of ^P-labeled connexin26, connexin32 or connexin43 cDNA (sp. act0.5-1 XI09 c.p.m./|lg DNA). The cDNAs were excised from their respectiveplasmids by digestion with £coRI and labeled by random primer extension(40). Autoradiographic films were developed after exposure for 1-2 days(connexin32 and connexin43 blots) or 14 days (connexin26 blots) at — 80°C.

Western blot analysis of connexin32Connexin32 protein levels in cultured liver cells were analyzed by Westernblotting essentially as we have described (39). The LOLA anti-connexin32antibody was used as the primary antibody at a 1:1000 dilution.

Nuclear run-on assayThe effect of dexamethasone on the transcriptional activity of the connexin32gene was analyzed by nuclear run-on assay (41). MH,C| cells were treatedwith dexamethasone (10 |iM) for 48 h or not treated, then nuclei were isolatedon a 1.8 M sucrose cushion. The isolated nuclei were resuspended in storagebuffer (25% glycerol, 5 mM magnesium acetate, 0.1 mM EDTA, 5 mMdithiothreitol, 50 mM Tris-HCl, pH 8.0) at a density of 1X108 nuclei/ml andstored at —80°C until use. For the transcription reactions, 0.1 ml aliquots ofnuclei (107) were added to 50 |xl of solution A (2 mM ATP, 2 mM CTP,2 mM GTP, 0.2 mM UTP, 0.1 mM 5-adenosylmethionine), 40 \i\ of solutionB (0.6 mM KC1, 12.5 mM magnesium acetate), 10 \l\ deionized, autoclavedwater and 50 u\Ci of [a-32P]UTP (3000 Ci/mmol) (DuPont/NEN, Boston,MA). The reactions were incubated for 2 h at 30°C. The RNA was isolatedusing 1 ml of TRI-REAGENT™ per reaction. The RNA was ethanolprecipitated and washed with 70% ethanol to remove non-incorporatednucleotides. The RNA (1 X 106 c.p.mVml hybridization buffer) was hybridizedto connexin32 cDNA or pGEM3 plasmid DNA (control for non-specifichybridization) which had previously been bound to Hybond N+ membraneby dot blot (3 |J.g DNA/dot). Prehybridization, hybridization and washingconditions were the same as those described above for Northern blotting.

Gap junction immunoslaining and visualization

Connexin32 gap junctions in cultured liver cells were detected by indirectimmunofiuorescence histochemistry as described previously (39). The LOLAanti-connexin32 antibody was used as the primary antibody at a 1:1000 dilution.

The cells were viewed by conventional epifluorescencc microscopy usinga Nikon Diaphot inverted microscope or by confocal microscopy using anInSIGHT PLUS Laser Scanning Confocal Microscope (Meridian Instruments,

i

plin

gco

uU

ye-

100•0to70eoBO

40

30

zo100<

Treatment Duration (h)

— $

5£

Dexamethoaone

Fig. 1. (A) Effects of dexamethasone and culture medium on dye-couplingin primary cultured rat hepatocytes. Hepatocytes were isolated and culturedfor 24 h without added steroid (0 h time point), refed with L-15 medium(triangles) or RPMI 1640 medium (circles) containing 1 |iM dexamethasone(filled symbols) or no steroid (open symbols), and assayed for dye-coupling48 h later (n = three or four dishes per group, mean ± SD). (B) Dose-responsive induction of dye-coupling by dexamethasone in primary culturedrat hepatocytes. Hepatocytes were isolated and cultured for 24 h withoutadded steroid, refed with L-15 medium (filled circles) or RPMI 1640medium (open circles) containing dexamethasone (0-10 nM), and assayedfor dye-coupling 48 h later (n = three or four dishes per group;mean ± SD).

DOC EJT TEST PTOC US COHT

Steroid

8

nn nNora OCX EST TEST PROS *LD COHT

SUroid

Fig. 2. Steroid-specific enhancement of dye-coupling in primary cultured rathepatocytes. Hepatocytes were isolated and cultured for 24 h without addedsteroid, refed with L-15 medium (A) or RPMI 1640 medium (B) containingone of the following steroids (all 1 |iM): dexamethasone (DEX), [J-estradiol(EST), testosterone (TEST), progesterone (PROG), aldesterone (ALD) orhydrocortisone (CORT), and assayed for dye-coupling 48 h later (n = threeor four dishes per group; mean ± SD). Dye-coupling was significantlyincreased (P < 0.05) by dexamethasone and hydrocortisone compared to thecontrol group (None).

Inc., Okemos, MI). The scan unit was mounted on an Olympus IMT2 invertedmicroscope equipped with a 100X oil immersion objective (NA 1.40). Thefluorescence excitation of FTTC was provided by the 488 nm line of a 100

1808

at University of C

alifornia, Santa Cruz on D

ecember 2, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

WB-F344

o~cu©Q.

ing

O-

ou1O

o

10090

80

70

SO

50

40

30

20

10

0-0EX +D€X -DEX +OEX

Treatments

Fig. 3. Effects of dexamethasone on dye-coupling in MH,C| and WB-F344cells. The cells were grown to -75% confiuency in complete mediumwithout added steroid, refed with Dulbecco's MEM medium containing10 nM dexamethasone (+DEX) or no dexamethasone (—DEX), andassayed for dye-coupling 48 h later (n = three or four dishes per group;mean ± SD). Dye-coupling was significantly increased (P < 0.05) bydexamethasone in MH|C| cultures.

mW argon ion laser, and the fluorescence was captured with a cooled CCDthrough a 530/30 nm bandpass filter. Using the InSIGHT-IQ computer system(Meridian Instruments, Inc.), the images were rescaled and processed witha directional smoothing filter to reduce the contribution of non-specificfluorescence and background noise respectively.

ResultsEffects of steroid hormones on rat liver cell dye-couplingThe effects of dexamethasone and other steroids on GJIC(dye-coupling) in rat hepatocyte primary cell cultures wereevaluated. After isolation, hepatocytes were cultured for 24 hin complete media (L-15 or RPMI 1640) and subsequentlytreated with steroid. Figure 1A demonstrates that after 24 hculture in the absence of steroid (0 h treatment time), hepato-cytes exhibited a low percentage of dye-coupling (1.6-3.9%).Addition of 1 (iM dexamethasone to the cultures resulted in atime-dependent induction of dye-coupling that occurred overthe next 48 h of culture. Control cultures showed no significantchanges in dye-coupling over this period. The enhancementof dye-coupling by dexamethasone was greater in L-15 mediumthan in RPMI 1640 medium (Figure 1A) and was dosedependent (Figure IB). Furthermore, the response was specificto glucocorticoid steroids. Both dexamethasone and hydrocor-tisone significantly enhanced hepatocyte dye-coupling in L-15 or RPMI 1640 media, whereas |}-estradiol, testosterone,progesterone and aldosterone had no effect (Figure 2A,B).

The effects of dexamethasone on MH,C, and WB-F344 celldye-coupling were also assessed (Figure 3). MH|C] cells hada low dye-coupling percentage in the absence of dexamethasonewhich was dramatically increased by culture for 48 h in thepresence of 10 |iM dexamethasone. WB-F344 cells exhibiteda high level of coupling in the absence of steroid which wasnot affected by dexamethasone.Steroid effects on liver cell connexin expression and gapjunction immunostainingIn order to determine if the enhancement of liver cell GJICby dexamethasone was related to changes in connexin geneexpression, we analyzed connexin mRNA levels by Northernblot. Connexin32 (1.6 kb) and connexin26 (2.5 kb) transcripts(9,36) were readily detected in freshly isolated hepatocytes(Figure 4A-C, lane 1), but were barely detectable (con-nexin32) or undetectable (connexin26) after the cells had beencultured for 24 h in the absence of dexamethasone (lane 2).When dexamethasone was added to these cultures, dose-

Glncocorticoid effects on connexin expression

1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8

1 2 3

5 6

Fig. 4. Northern blot analyses of dose-responsive increases of connexin32(A3) and connexin26 (C) mRNA levels by dexamethasone in primarycultured rat hepatocytes. Hepatocytes were isolated and cultured for 24 hwithout added steroid, refed with RPMI 1640 medium (A,C) or L-15medium (B) containing dexamethasone (0—10 |iM), and cultured for anadditional 48 h prior to isolation of total RNA. Lane 1 (A—C): freshlyisolated hepatocytes; lanes 2-8 (A3): 0, 0.0001, 0.001, 0.01, 0.1, 1 and 10uJvl dexamethasone respectively; lanes 2-7 (C): 0, 0.001, 0.01, 0.1, 1 and10 (iM dexamethasone respectively. The sizes of the connexin32 andconnexin26 bands were 1.6 and 2.5 kb respectively. Left panels: Northernblot autoradiographs; right panels: 28S and 18S ribosomal RNA bands inethidium bromide-stained gels.

A 1 2 3 4 5 6 7 1 2 3 4 5

1 2

C 1 2 3 4 5 6 7 1 2 3 4 5 6 7

m - ill i LCDFig. 5. Northern blot analyses of time-dependent (A3) and steroid-specific(C) induction of connexin32 mRNA levels by steroid hormones in primarycultured rat hepatocytes. (A) Lanes 1-7: hepatocytes cultured for 24 hwithout added steroid, refed with RPMI 1640 medium containing 1 |iMdexamethasone, and cultured for 0, 4, 12, 16, 24, 36 and 48 h respectively.(B) Lanes 1-7: hepatocytes cultured for 24 h without added steroid, refedwith RPMI 1640 medium containing no dexamethasone, and cultured for 0,4, 12, 16, 24, 36 and 48 h respectively. (C) Hepatocytes cultured for 24 hwithout added steroid, refed with RPMI 1640 medium containing no steroid(lane 1) or 1 |iM steroid [dexamethasone (lane 2), (J-estradiol (lane 3),testosterone (lane 4), aldostcrone (lane 5), hydrocortisone (lane 6) orprogesterone (lane 7)], and cultured 48 h before isolation of total RNA. Theconnexin32 band was -1.6 kb. Left panels: Northern blot autoradiographs;right panels: 28S and 18S ribosomal RNA bands in ethidium bromide-stained gels.

responsive increases in connexin32 and connexin26 transcriptlevels were evident after 48 h treatment (lanes 3-8). Theenhancement of connexin32 gene expression by dexamethasonewas time dependent. Elevations of connexin32 mRNA wereevident 4 h after treatment of hepatocytes with dexamethasoneand were maximal at 48 h (Figure 5A). In the absence ofdexamethasone, connexin32 mRNA levels declined duringthese times (Figure 5B). The enhancement of connexin32transcripts was specific to glucocorticoid steroids since bothdexamethasone and hydrocortisone, but not other types ofsteroids, increased connexin32 mRNA levels (Figure 5C).The apparent elevation of connexin32 mRNA induced byaldosterone (lane 5) was not reproducible.

The cell-specific effects of dexamethasone on connexin32,connexin26 and connexin43 steady-state mRNA levels wereexamined. Connexin32 transcripts were increased in hepato-

1809

at University of C

alifornia, Santa Cruz on D

ecember 2, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

P.Ren rtai

1 2 3

B 1 2 3 4 5 6 7 1 2 3 4 5 6 7

C 1 2 3 4 5 6 7 1 2 3 4 5 6 7

Fig. 6. Northern blot analysis of cell-specific effects of dexamethasone onconnexin32 (A), connexin26 (B) and connexin43 (C) mRNA levels inhepatic cells. Lane 1: freshly isolated rat hepatocytes; lane 2 hepatocytescultured without added steroid for 72 h; lane 3: hepatocytes cultured for 24h without added steroid followed by 48 h with 1 |iM dexamethasone; lane4: MH|C| cells cultured without added steroid; lane 5: MHjC| cellscultured in the presence of 10 (iM dexamethasone for 48 h; lane 6: WB-F344 cells cultured without added steroid; lane 7: WB-F344 cells culturedin the presence of 10 |J.M dexamethasone for 48 h. The connexin32,connexin26 and connexm43 bands were ~1.6, ~2.5 and ~3.0 kb respectively.Left panels: Northern blot autoradiographs; right panels: 28S and 18Sribosomal RNA bands in ethidium bromide-stained gels.

CX32 pGEM3cDNA plasmid

- DEX

+ DEX

Fig. 7. Nuclear run-on assay of connexin32 gene transcription in controland dexamethasone-treated MH,C| cells. Nuclei were isolated from MH|C|cells that had been treated with no dexamethasone (—DEX) or 10 jiMdexamethasone (+DEX) for 48 h and incubated with ATP, CTP, GTP and[a-32P]UTP (1 X107 nuclei per reaction). Total RNA was isolated from thereactions and hybridized to connexin32 cDNA and pGEM3 plasmid DNA(3 u.g DNA per dot; 1 X10* c.p.m. total RNA per ml of hybridizationbuffer).

cytes and MH,C| cells by dexamethasone (Figure 6A, lanes2-5). Connexin26 mRNA was not detectable in untreatedhepatocytes but was increased by dexamethasone (Figure 6B,lanes 2 and 3). This transcript was not detected in controlM H ^ i cells (lane 4) and was not increased by dexamethasone(lane 5). Connexin43 transcripts (3.0 kb; ref. 35) were notdetected in control or dexamethasone-treated primary culturedhepatocytes or MK^C, cells (Figure 6C, lanes 2-5). A weaksignal was evident, however, in freshly isolated hepatocytesamples (lane 1) and might have been due to contaminatingnon-parenchymal cells. In contrast to primary cultured hepato-cytes and MHiC] cells, connexin43 mRNA was detected inWB-F344 cells (Figure 6C, lanes 6 and 7). Connexin26 andconnexin32 transcripts were not evident in these cells (Figure6A,B, lanes 6 and 7). Dexamethasone and other steroids atconcentrations of 0.1, 1 and 10 (iM did not alter the levels of

CX32 -

Fig. 8. Effect of dexamethasone on connexin32 protein levels in MH|C|and WB-F344 cells determined by Western blot. The MH,C| cells (lanes1,2) and WB-F344 cells (lanes 3,4) were cultured without dexamethasone(lanes 1,3) or with 10 U.M dexamethasone (lanes 2,4) for 48 h prior toisolation of gap junction protein by alkaline extraction. The blots wereanalyzed using rabbit polyclonal anti-connexin32 antibody (9).Approximately 20 p.g of protein were loaded per lane.

Fig. 9. Effect of dexamethasone on gap junctions in MH|C| cellsimmunofluorescently labeled for connexin32. The cells were culturedwithout dexamethasone (A,C) or in the presence of 10 |iM dexamethasone(B,D) for 48 h, fixed and immunostained using anti-connexin32 antibody (9)as described in Materials and methods. The gap junctions appear as brightspots between adjacent cells (A,B: conventional epifluorescence microscopicimage; C,D: confocal microscopic section).

these three mRNAs in WB-F344 cells (Figure 6A-C, lane 7;data not shown).

Nuclear run-on assays using nuclei isolated from controland dexamethasone-treated M H , ^ cells were performed todetermine if transcription of the connexin32 gene was increasedby the steroid. A weak hybridization signal to the blottedconnexin32 cDNA was evident using RNA isolated fromcontrol cell nuclei (Figure 7). A much stronger signal wasobserved with RNA isolated from nuclei of dexamethasone-treated cells. Non-specific hybridization to the pGEM3 plasmid

1810

at University of C

alifornia, Santa Cruz on D

ecember 2, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

Glucocorticoid effects on connexin expression

DNA was not evident. These data indicate that connexin32gene transcription was occurring in control cells and wasgreatly increased by treatment of cells with dexamethasone.

Levels of connexin32 protein in control and dexamethasone-treated MHiC] cells were analyzed by Western blot in orderto determine if the enhanced connexin32 transcript levels wereaccompanied by increases in connexin32 protein (Figure 8).A 27 kDa band corresponding to the position of connexin32on SDS-PAGE gels (12) was present in control MH,C, cells(lane 1). The intensity of this band was markedly increasedafter the cells had been cultured with 10 |iM dexamethasonefor 48 h (lane 2). Similar results were seen with primaryhepatocytes cultured in the presence of 10 uM dexamethasonefor 48 h (data not shown). No 27 kDa band was apparent inextracts of WB-F344 cells cultured with or without dexametha-sone (lanes 3 and 4).

Indirect immunofluorescence staining and analysis of gapjunctions in MH[Ci cells were performed using anti-con-nexin32 antibody. Figure 9 displays conventional epifluores-cence and confocal microscopic images of cells cultured for48 h in the presence and absence of 10 |iM dexamethasone.The connexin32 gap junctions appear as bright fluorescentspots between adjacent cells. Few connexin32 gap junctionswere seen between cells cultured without dexamethasone. Theaddition of the steroid, however, markedly increased thenumber of connexin32 gap junctions. Note that the cytoplasmof these cells also shows some fluorescence. This cytoplasmicstaining was apparently due to background staining sincesimilar levels were also noted in cells stained with thesecondary antibody alone (data not shown).

DiscussionThis study demonstrates that dye-coupling was enhanced incultures of rat hepatocytes and MH,C| cells by the glucocort-icoids dexamethasone and hydrocortisone, in a dose- and time-dependent manner and that these effects were correlatedwith increases in connexin32 expression, but not connexin43expression. Connexin26 expression was also increased slightlyby dexamethasone in the hepatocyte cultures. These actionswere limited to glucocorticoid steroid hormones since othertypes of steroids did not affect dye-coupling or connexinexpression. The magnitude of these effects in hepatocytes wasalso dependent upon the type of culture medium utilized.Glucocorticoid enhancement of hepatocyte dye-coupling andconnexin expression was greater in L-15 medium than inRPMI 1640 medium. This may be due to the higher quantitiesof amino acids in the former medium or other differencesbetween the two media. For instance, L-15 contains sodiumpyruvate and DL-alanine, which are lacking from RPMI 1640.Both components enhanced mouse hepatocyte dye-couplingwhen added to RPMI 1640 medium (42).

In contrast to hepatocytes and M H ^ cells, WB-F344cells expressed high levels of connexin43 and no detectableconnexin32 or connexin26 as we have previously reported(39). In these cells, dexamethasone had no effect on connexinexpression. This effect did not appear to be due to the culturemedium since these cells do not express connexin32 whentreated with dexamethasone and cultured in Dulbecco's MEM(43) or RPMI 1640 (P.Ren and R.J.Ruch, unpublished).WB-F344 cells have been reported to be capable of differentiat-ing into hepatocytes when injected into the rat liver (44),but apparently are not differentiated enough to respond todexamethasone in vitro in a manner similar to hepatocytes.

The observed enhancement of connexin32 expression bydexamethasone in MHiCi cells was due, at least in part, toincreased transcription of the connexin32 gene. The nuclearrun-on assay (Figure 7) revealed that connexin32 gene trans-cription was more active in nuclei isolated from dexametha-sone-treated MHjC, cells than in control cells.

These results are consistent with the well-described effectsof glucocorticoids on gene expression (45,46). In most cases,steroid actions on gene expression result from interactions ofactivated steroid receptors with m-regulatory DNA elementslocated within steroid-responsive genes and subsequent modu-lation of transcriptional activity. A consensus sequence forthe glucocorticoid response element (GRE; 5'-GGTA-CAnnnTGTTCT-3') has been reported (45,46). No GREs havebeen identified in the rat connexin32 and connexin26 genomicsequences published thus far (47,48). However, a similarsequence (5'-GCCACAAAGTGGTCT-3') is located at nucleo-tide positions -409 to -395 upstream of the rat connexin32promoter (47). One or more GREs may also be present innon-sequenced regions of the connexin32 and connexin26genes; their identification will require more extensive sequenceinformation.

Since no GREs have yet been identified in the connexin32and connexin26 genes, it remains unclear whether glucocorti-coids induced hepatic cell connexin32 and connexin26 expres-sion directly or indirectly. The increase of connexin32 mRNAdetected in hepatocytes within 4 h after the addition ofdexamethasone (Figure 5) occurs rapidly enough to suggest adirect response of the gene, although an indirect mechanismmight also occur within this time period. Glucocorticoids havebeen reported to alter the activity or expression of other factorsinvolved in connexin32 and connexin26 gene regulation. Suchan effect would indicate an indirect mechanism. For instance,cyclic AMP and P-adrenergic agonists have been reported toinduce connexin32 expression (20,21) and putative cyclic AMPresponsive elements have been described in the rat and mouseconnexin32 promoter regions (47,48). Dexamethasone hasbeen reported to enhance cyclic AMP metabolism and theactivity of p-adrenergic receptors (49). Glucocorticoids havealso been reported to alter the morphology of primary culturedhepatocytes, enhance hepatocyte cell-cell contact, and alterthe composition of their extracellular matrices (50). Theseeffects in turn have been shown to affect hepatocellularconnexin gene expression (25,29,51). Increased connexinmRNA stabilization and translational rates might also havecontributed to the enhancement of dye-coupling and con-nexin32 and connexin26 mRNA levels. Recent studies fromtwo independent laboratories suggest that hepatic connexin32and connexin26 mRNAs are regulated post-transcriptionallyin vivo (52,53). All of these potential mechanisms need to beaddressed experimentally.

The present data also raise the possibility that glucocorticoidsmay affect connexin32 and connexin26 expression in vivo.Several other types of steroids have been reported to alter dye-coupling and connexin expression in non-hepatic cells andtissues (30-33). We are currently examining the effects ofadrenalectomy and dexamethasone administration on connexinexpression in the rat.

Recent studies strongly implicate the involvement of GJICin the regulation of growth, phenotype and neoplastic trans-formation (1-4). Glucocorticoids have been widely recognizedfor their ability to induce hepatic differentiation and inhibitthe growth of normal and neoplastic tissues (54-61). Whether

1811

at University of C

alifornia, Santa Cruz on D

ecember 2, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

P.Ren et al

the enhancement of GJIC and connexin expression by glucocor-ticoids is a cause or consequence of hepatic differentiationand growth inhibition remains to be determined.

AcknowledgementsThe authors are grateful to Dr Ronald N.Hines (Wayne State University,Detroit, MI) for providing technical expertise and critically reading themanuscript. This study was supported by NCI-CA57612 awarded to RJ.R.

References1. Loewenstein.W.R. (1990) Cell-cell communication and the control of

growth. Am. Rev. Respir. Dis., 142, S48-S53.2.TroskoJ.E., Madhukar.B.V. and Chang, C.C. (1993) Endogenous and

exogenous modulation of gap junctional intercellular communication:toxicological and pharmacological implications. Life Sci., 53, 1-19.

3. KlaunigJ.E. and Ruch.R.J. (1990) Role of intercellular communication innongenotoxic carcinogenesis. Lab. Invest., 62, 135-146.

4. Yamasaki.H. (1990) Gap junctional intercellular communication andcarcinogenesis. Carcinogenesis, 11, 1051-1058.

5.Lee,S.W., Tomasetto.C, Paul.D.L., Keyomarsi.K. and Sager.R. (1992)Transcriptional down-regulation of gap junctional proteins blocksjunctional communication in human mammary tumor cell lines. /. CellBioL, 118, 1213-1221.

6.Katoh,R, Klein J.-L., Bignami.M. and Yamasaki.H. (1993) Association ofviral oncogene-induced changes in gap junctional intercellularcommunication and morphological transformation in BALB/c3T3 cells.Carcinogenesis, 14, 435-440.

7.Musil,L.S., Cunningham.B.A., Edelman.G.M. and Goodenough.D.A.(1990) Differential phosphorylation of the gap junction protein connexin43in junctional communication-competent and -deficient cell lines. J. CellBioL, 111, 2077-2088.

8. Brissette J.L., Kumar.N.M., GilulaJsl.B. and Dotto.G.P. (1991) The tumorpromoter 12-O-tetradecanoylphorbol-13-acetate and the ras oncogenemodulate expression and phosphorylation of gap junction proteins. Mol.Cell BioL, 11,5364-5371.

9. Paul.D.L. (1986) Molecular cloning of cDNA for rat liver gap junctionprotein. J. Cell BioL, 103, 123-134.

lO.Beer.D.G., Neveu.MJ., Paul.D.L., Rapp.U.R. and Pitot,H.C. Expressionof the c-raf protooncogene, Y-glutamyltranspeptidase, and gap junctionprotein in rat liver neoplasms. Cancer Res., 48, 1610-1617, 1988.

ll.Neveu^MJ., HullyJ.R., Paul.D.L. and Pitot.H.C. (1990) Reversiblealteration in the expression of the gap junctional protein connexin 32during tumor promotion in rat liver and its role during cell proliferation.Cancer Commun., 2, 21-31.

12. Htzgerald.DJ., Mesnil.M., Oyamada,M., Tsuda,H., Ito,N. and Yamasaki.H.(1989) Changes in gap junction protein (connexin 32) gene expressionduring rat liver carcinogenesis. J. Cell. Biochem., 41, 97-102.

13.Krutovskikh,V.A., Oyamadajd. and Yamasaki.H. (1991) Sequentialchanges of gap-junctional intercellular communications during multistagerat liver carcinogenesis: direct measurement of communication in vivo.Carcinogenesis, 12, 1701-1706.

14. KlaunigJ.E., Ruch.RJ., HamptonJ.A., Weghorst,C.M. and HartnettJ.A.(1990) Gap junctional intercellular communication and murine hepaticcarcinogenesis. Prog. Clin. BioL Res., 331, 277-291.

15,Oyamada,M., Krutovskikh.V.A., Mesniljvl., Partensky.C, Berger.F. andYamasaki,H. (1990) Aberrant expression of gap junction gene in primaryhuman hepatocellular carcinomas: increased expression of cardiac-typegap junction gene connexin 43. MoL Carcinogenesis, 3, 273-278.

16. Janssen-Timmen,U., Traub.O., Dermietzel.R., Rabes.H.M. and Willecke.K.(1986) Reduced number of gap junctions in rat hepatocarcinomas detectedby monoclonal antibody. Carcinogenesis, 7, 1475-1482.

l7.Wilgenbus,K.K., Kirkpatrick.CJ., Knuechel.R., Willecke.K. and Traub.O.(1992) Expression of Cx26, Cx32 and Cx43 gap junction proteins innormal and neoplastic human tissues. Int. J. Cancer, 51, 522-529.

18.Willeckc,K., Hennemann.H., Dahl.E., Jungbluth.S. and Heynkes.R. (1991)The diversity of connexin genes encoding gap junctional proteins. Eur. J.Cell BioL, 56, 1-7.

19.MehtaJ>.R, Yamamoto.M. and Rose.B. (1992) Transcription of the genefor the gap junctional protein connexin43 and expression of functionalcell-to-ccll channels are regulated by cAMP. MoL BioL Cell, 3, 839-850.

20. SaezJ.C., Gregory,W.A., Watanabe.T, Dermietzel.R., Hertzberg,E.L.,ReicLL.M., BennettJU.V.L. and Spray.D.C. (1989) cAMP delaysdisappearance of gap junctions between pairs of rat hepatocytes in primaryculture. Am. J. Physiol., 257, Cl-Cll .

21. Traub.O., LookJ., Paul.D. and Willeckc.K. (1987) Cyclic adenosine

monophosphate stimulates biosynthesis and phosphorylation of the 26 kDajunction protein in cultured mouse hepatocytes. Eur. J. Cell BioL, 43,48-54.

22. Rogers.M., BerestckyJ.M., Hossain,M.Z., Guo.H., Kadle.R., Nicholson.B J.and Bertram J.S. (1990) Retinoid-enhanced gap junctional communicationis achieved by increased levels of connexin 43 mRNA and protein. Mol.Carcinogenesis, 3, 335—343.

23.Guo,H., Acevedo,P., Parsa.F.D. and BertramJ.S. (1992) Gap-junctionalprotein connexin 43 is expressed in dermis and epidermis of human skin:differential modulation by retinoids. J. Invest. DermatoL, 99, 460-467.

24.Zhang,L.-X., Cooney.R.V. and BertramJ.S. (1991) Carotenoids enhancegap junctional intercellular communication and inhibit lipid peroxidationin C3H/10T1/2 cells: relationship to their cancer chemopreventive action.Carcinogenesis, 12, 2109-2114.

25. Spray.D.C., Fujita>1., SaezJ.C., Choi.H., Watanabe.T., Hertzberg.E.,Rosenberg.L.C. and Reid,L.M. (1987) Proteoglycans and glycosaminoglycansinduce gap junction synthesis and function in primary liver cultures. J.Cell BioL, 105,541-551.

26. Dermietzel.R., Yancey.S.B., Traub.O., Willecke.K. and RevelJ.-P. (1987)Major loss of the 28-kDa protein of gap junction in proliferatinghepatocytes. J. Cell BioL, 105, 1925-1934.

27.Goulet,F., Normand.C. and Morin.O. (1988) Cellular interactions promotetissue-specific function, biomatrix deposition and gap junctionalcommunication of primary cultured hepatocytes. Hepatology, 8,1010-1018.

28.Mesnil,M., Piccoli.C. and Yamasaki.H. (1993) An improved long-termculture of rat hepatocytes to detect liver tumour-promoting agents: resultswith phenobarbital. Eur. J. Pharmacol. Environ. Toxicol. Pharmacol., 248,59-66.

29.Fujita,M., Spray.D.C., Choi.H., SaezJ.C, Watanabe.T., Rosenberg.L.C.,Hertzberg.E.L. and Reid.L.M. (1987) Glycosaminoglycans and proteo-glycans induce gap junction expression and restore transcription of tissue-specific mRNAs in primary liver cultures. Hepatology, 7, 1S-9S.

30. MacKenzie.L.W. and Garfield.R.E. (1985) Hormonal control of gapjunctions in the myometrium. Am. J. Physiol., 248, C296-C308.

31.Tabb,T, Thilander.G., Grover,A., Hertzberg.E. and Garfield.R. (1992) Animmunochemical and immunocytologic study of the increase in myometrialgap junctions (and connexin 43) in rats and humans during pregnancy.Am. J. Obstet. GynecoL, 167, 559-567.

32. Hendrix.E.M., Mao.SJ.T, Everson.W. and Larsen.WJ. (1992) Myometrialconnexin 43 trafficking and gap junction assembly at term and in pretermlabor. Mol. Reprod. Dev., 33, 27-38.

33.Matsumoto,A., Arai.Y., Urano.A. and Hyodo.S. (1991) Androgen regulatesgap junction mRNA expression in androgen-sensitive motoneurons in therat spinal cord. Neurosci. Lett., 131, 159-162.

34. Ruch.R.J. and KlaunigJ.E. (1992) Enhancement of rodent hepatocyte gapjunctional intercellular communication by dexamethasone. In Vitro Toxicol.,5, 103-111.

35. Beyer.E.C, Paul.D.L. and Goodenough.D.A. (1987) Connexin43: a proteinfrom rat heart homologous to a gap junction protein from liver. J. CellBioL, 105, 2621-2629.

36.ZhangJ.-T. and Nicholson,BJ. (1989) Sequence and tissue distribution ofa second protein of hepatic gap junctions, Cx26, as deduced from itscDNA. J. Cell BioL, 109, 3391-3401.

37. Richardson.U.I., Tashjian.A.H.Jr and Levine.L. (1969) Establishment ofa clonal strain of hepatoma cells which secrete albumin. J. Cell BioL, 40,236-247.

38.Tsao,M.S., SmithJ.D., Nelson.lCG. and GrishamJ.W. (1984) A diploidepithelial cell line from normal adult rat liver with phenotypic propertiesof oval cells. Exp. Cell Res., 154, 38-52.

39.Ruch,RJ., Bonney.WJ., Sigler.K., Guan,X., Matesic.D., Schafer.L.D.,Dupont.E. and TroskoJ.E. (1994) Loss of gap junctions from DDT-treatedrat liver epithelial cells. Carcinogenesis, 15, 301-306..

40. FeinbergAP and Vogelstein.B.A (1983) A technique for radiolabelingDNA restriction endonuclease fragments to high specific activity. AnaLBiochem., 132,6-12.

41.Marzluff,W.F. and Huang.R.C.C. (1984) Transcription in isolated nuclei.In Hames.B.D. and Higgins.SJ. (eds), In Vitro Transcription andTranslation: A Practical Approach. IRL Press, Oxford, pp. 89-129.

42. Klaunig J.E., Ruch.RJ., Hampton J.A., Weghorst,C.M. and HartnettJA.(1990) Gap junctional intercellular communication and murinehepatocarcinogenesis. Prog. Clin. BioL Res., 331, 227-291.

43. Stutenkemper.R., Geisse.S., Schwarz,HJ., LookJ., Traub.O.,Nicholson.BJ. and Willecke.K. (1992) The hepatocyte-specific phenotypeof murine liver cells correlates with high expression of connexin32 andconnexin26 but very low expression of connexin43. Exp. Cell Res., 201,43-54.

44.Coleman,W.B., Wennerberg^A.E., Smith.GJ. and GrishamJ.W. (1993)

1812

at University of C

alifornia, Santa Cruz on D

ecember 2, 2014

http://carcin.oxfordjournals.org/D

ownloaded from

Glucocorticoid effects on connexin expression

Regulation of diploid and some aneuploid rat liver epithelial (stemlike)cells by the hepatic microenvironment. Am. J. Pathot., 142, 1373-1382.

45.Beato, M. (1989) Gene regulation by steroid hormones. Cell, 56, 335-344.46. Evans, R.M. (1988) The steroid and thyroid hormone receptor superfamily.

Science, 240, 889-895.47. Miller.T., Dahl,G., and Wemer.R. (1988) Structure of a gap junction gene:

rat connexin-32. Bwsci. Rep., 8, 455-464.48.Hennemann,H., Kozjek.G., Dahl.E., Nicholson.B. and Willecke.K. (1992)

Molecular cloning of mouse connexin26 and -32: Similar genomicorganization but distinct promoter sequences of two gap junction genes.Eur.J. Cell Bio/.. 58, 81-89.

49.Christofferson,T, Refsnes.M., Bronsta,G.O., Ostby,E., HuseJ., Haffner.E,Sand.T., Hunt,N.H. and Sonne.O. (1984) Changes in hormoneresponsiveness and cyclic AMP metabolism in rat hepatocytes duringprimary culture and effects of supplementing the medium with insulin anddexamethasone. Eur. J. Biochem., 138, 217-226.

50. Marceau,N., GoyetteJ*., PelleUer.G. and Antakly.T. (1983) Hormonally-induccd changes in the cytoskeleton organization of adult and newbornrat hepatocytes cultured on fibronectin precoated substratum: effect ofdexamethasone and insulin. Cell. Mol. BioL, 29, 421-435.

51.JeffersonX>M., Clayton.D.F., DarnelU.R., and Reid,L.M. (1984) Post-transcnptional modulation of gene expression in cultured rat hepatocytes.Mol. Cell. BioL, 4, 1929-1939.

52. Rosenberg.E., Spray,D.C. and Reid.L.M. (1992) Transcriptional andposttranslational control of connexin mRNAs in periportal and pericentralrat hepatocytes. Eur. J. Cell BioL, 59, 21-26.

53.Kren, B.T., Kumar, N.M., Wang, S., Gilula, N.B. and Steer, CJ. (1993)Differential regulation of multiple gap junction transcripts and proteinsduring rat liver regeneration. J. Cell BioL, 123, 707-718.

54. Baribault,H and Marceau,N. (1986) Dexamethasone and dimethylsulfoxideas distinct regulators of growth and differentiation of cultured sucklingrat hepatocytes. J. Cell. Physiol., 129, 77-84.

55. Lescoat,G., Pasdeloup,N., Kneip.B., and Guguen-Guillouzo.C. (1987)Modulation of alpha-fetoprotein, albumin, and transferrin gene expressionby cellular interactions and dexamethasone in cocultures of fetal rathepatocytes. Eur. J. Cell BioL, 44, 128-134.

56. Castellano.TJ., Schiffman,R.L., JacobJvl.C. and LoebJ.N. (1978)Suppression of liver cell proliferation by glucocorticoid hormone: acomparison of normally growing and regenerating tissue in the immaturerat. Endocrinology, 102, 1107-1112.

57.Cook,P.W., Swanson.KX, Edwards.C.P. and Firestone.G.L. (1988)Glucocorticoid receptor-dependent inhibition of cellular proliferation indexamethasone-resistant and hypersensitive rat hepatoma cell variants.MoL Cell. BioL, 8, 1449-1459.

58. Vintermyr.O.K. and Doskeland.S.O. (1989) Characterization of theinhibitory effect of glucocorticoids on the DNA replication of adult rathepatocytes growing at various cell densities. J. CelL Physiol., 138, 29-37.

59.Morishige,W.K. and MontesAM. (1989) The effect of corticosterone onthe growth of perinatal rat lung fibroblasts in culture: modulation by theextracellular matrix. Dev. Growth Diff., 31, 95-102.

60. Braunschweiger.P.G., Ting.H.L. and Schiffer.L.M. (1983) Correlationbetween glucocorticoid receptor content and the antiproliferative effect ofdexamethasone in experimental solid tumors. Cancer Res., 43, 4757-4761.

61.Webster,M.K., GuthrieJ. and Firestone.G.L. (1990) Suppression of ratmammary tumor cell growth in vitro by glucocorticoids requires serumproteins. J. BioL Chenu, 265, 4831-4838.

Received on April 21, 1994; revised on June 1, 1994; accepted on June 10, 1994

1813

at University of C

alifornia, Santa Cruz on D

ecember 2, 2014

http://carcin.oxfordjournals.org/D

ownloaded from