Proc. Nati. Acad. Sci. USAVol. 80, pp. 7229-7233, December 1983Cell Biology

Reciprocal modulation of growth and differentiated functions ofmature rat hepatocytes in primary culture by cell-cell contactand cell membranes

(cell density dependency/plasma membranes/cell growth/liver functions)

TOSHIKAZU NAKAMURA, KATSUHIKO YOSHIMOTO, YASUO NAKAYAMA*, YUMIKo TOMITA, ANDAKIRA ICHIHARAtInstitute for Enzyme Research, School of Medicine, University of Tokushima, Tokushima 770, Japan

Communicated by, Van R. Potter, August 18, 1983

ABSTRACT In primary monolayer cultures of rat mature he-patocytes, many metabolic functions as well as cell growth are reg-ulated by cell density. There are two types of regulatory responseof these functions to change of cell density. Growth-related func-tions, such as DNA synthesis, induction of glucose-6-phosphatedehydrogenase, 2-aminoisebutyric acid transport, synthesis ofcellular protein, and cholesterogenesis, are stimulated by low celldensity. In contrast, functions related to hepatocyte-specific char-acters, such as the inductions of tyrosine aminotransferase, serinedehydratase, and malic enzyme and synthesis of triglycerides, arestimulated by high cell density. The reciprocal responses of thesecellular activities to cell density were mimicked by addition ofplasma membranes purified from adult rat liver to hepatocytescultured at low cell density. The modulator activity was heat labileand trypsin sensitive. The activity was also found in plasma mem-branes from kidney, brain, and erythrocytes, although the spe-cific activities of these preparations seemed to be different. Theseresults suggest that the reciprocal regulations of cell growth andhepatocyte-specific functions are mediated by some surface com-ponents via cell-cell contact.

Adult rat hepatocytes in primary culture retain in vivo levels ofvarious liver functions and respond to various hormones (1-4).However, adult rat hepatocytes do not proliferate, even whenthey are cultured in 10% fetal calf serum. Several investigatorshave found that insulin and epidermal growth factor (EGF)stimulated DNA synthesis in mature rat hepatocytes in primaryculture, although the cells did not divide (5-8). Recently, wefound that mature hepatocytes could proliferate when culturedat lower cell density in the presence of insulin and EGF, andwe suggested that a density-dependent mechanism regulatesentry of cells into the cell cycle (9). Michalopoulos et al. (8) alsoobserved the importance of cell density for DNA synthesis andcell proliferation in primary cultured hepatocytes. Subse-quently, we noticed density-dependent control of many othercellular activities of mature hepatocytes besides proliferationand we classified the activities of the cells into two types, growth-related functions and hepatocyte-specific characters.

This paper reports that the two types of function are con-trolled reciprocally by cell density and that their regulation athigh cell density can be mimicked by addition of hepatic plasmamembranes to cultures at low density. We suggest that cell growthand various liver functions are regulated by cell surface com-ponents via cell-cell contact in vivo.

MATERIALS AND METHODSMaterials. The materials used for cell isolation and culture

were as reported (10). Insulin, glucagon, trypsin (type III), trypsininhibitor (type I-S), acetyl-coenzyme A, and malonyl-coenzymeA were obtained from Sigma; dexamethasone (Dex) was fromSchering (Berlin); 3,3',5'-triiodo-DL-thyronine (T3) was fromNakarai Chemical (Kyoto, Japan); EGF was from CollaborativeResearch (Waltham, MA); NADP and glucose 6-phosphate werefrom Oriental Yeast Co. (Osaka, Japan); Percoll was from Phar-macia; [1-14C]acetate (57.6 mCi/mmol; 1 Ci = 3.7 X 10k' Bq),L-[4,53H]leucine (55 Ci/mmol), [1-'4C]-2-aminoisobutyric acid(AIB) (61 mCi/mmol), and [methy1-3H]thymidine (54.2 Ci/mmol)were from the Radiochemical Centre.

Cell Isolation and Monolayer Culture. Parenchymal hepa-tocytes were isolated from male Wistar rats weighing 150-200 gby in -situ perfusion on the liver with collagenase, essentially asdescribed by Seglen (11). The isolated cells were suspended at5 X 10' cells per ml in Williams medium E containing 2% calfserum and 2 nM insulin. The cell density at inoculation wasvaried from 5 X 105 to 2.5 x 106 cells per 5-cm (diameter)Corning plastic culture dish. Over 80% of the cells became at-tached in 1 hr at 37C under 5% C02/30% 02 in air. After cul-ture for 2 hr. the medium -was changed to hormone-free Wil-liams medium E containing 5% calf serum. In experiments onthe effect of addition of plasma membranes, rat hepatocyteswere inoculated at a low density (7 x 105 cells per 5-cm dish).After 2 hr. the medium was changed to hormone-free mediumcontaining 5% calf serum and then various amounts of plasmamembranes were added to the cultures. Hormones were addedto the medium after 22 hr of culture. Incubation was continuedfor 24 or 45 hr further and then various activities were assayed.

Assay of DNA Synthesis. DNA synthesis was assayed bymeasuring incorporation of [3H]thymidine into DNA with orwithout hydroxyurea in 2 hr. Radioactivity in the hot trichlo-roacetic acid-soluble fraction was measured as described (6).

Assays of Enzyme Activities. The cells were harvested witha rubber policeman in 0.2-1.0 ml of 20 mM potassium phos-phate buffer (pH 7.5) containing 0.1 mM dithiothreitol and5 mM EDTA for the glucose-6-phosphate dehydrogenase (Glc-6-P-DHase) assay, in 0.1-0.5 ml of 5 mM Tris-HCl buffer (pH

Abbreviations: EGF, epidermal growth factor; T3, 3,3',5'-triiodo-DL-thyronine; Glc-6P-DHase, glucose-phosphate dehydrogenase; T-NH2Tase, tyrosine aminotransferase; Ser-dehydratase, serine dehydra-tase; AIB, 2-aminoisobutyric acid; Dex, dexamethasone; mU, miMi-unit(s).* On leave from Otsuka Pharmaceutical Co., Ltd., Tokushima ResearchInstitute, Tokushima 77-01, Japan.

t To whom reprint requests should be addressed.

7229

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertise-ment" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

July

22,

202

1

7230 Cell Biology: Nakamura et al.

7.4) containing 1 mM dithiothreitol, 0.1 mM EDTA, and 0.25M sucrose for assay of malic enzyme, and in 0.1-0.5 ml of 0.1 Mpotassium phosphate buffer (pH 8.0) containing 1 mM EDTAand 50 AM pyridoxal phosphate for assays of serine dehydratase(Ser-dehydratase) and tyrosine aminotransferase (T-NH2Tase).The volume of homogenizing buffer was adjusted to give a con-stant protein concentration. The cells were disrupted by twocycles of freeze-thawing, the homogenate was centrifuged at30,000 x g for 30 min, and the resulting supernatant was usedas the enzyme preparation for assays of Glc-6-P-DHase (12),malic enzyme (13), Ser-dehydratase (14), and T-NH2Tase (15).Units of activity are defined as the amounts forming 1 Amol ofNADPH per min at 240C for Glc-6-P-DHase and malic enzyme,1 Amol of pyruvate per min at 370C for-Ser-dehydratase, and1 ,umol of p-hydroxyphenylpyruvate per min at 37°C for T-NH2-Tase. Protein was measured by the method of Bradford (16)with bovine serum albumin as standard and specific activity wasexpressed as units/mg of protein.

Assay of Lipogenesis. Monolayers of hepatocytes in 5-cmdishes were incubated with 2 ml of culture medium containing5 mM Na-[1-14C]acetate (1 ,Ci/ml) for 3 hr. The cells werewashed twice with phosphate-buffered saline (PI/NaCl), har-vested in 1 ml of Pi/NaCl, and homogenized in a Polytron ho-mogenizer. Lipids in the homogenates were extracted by themethod of Bligh and Dyer (17). The extract was evaporated un-der N2, and the concentrated extract in CHCl3 was spotted onsilica gel 60 plates (Merck, Darmstadt, FRG) and developedwith n-hexane/ethyl ether/acetic anhydride, 70:20:1 (vol/vol).Spots were located with iodine vapor, scraped off into vials, andassayed for radioactivity in toluene scintillator.

Assays of AIB Transport and Protein Synthesis. Transportof AIB was determined by the method of Kletzien et al. (18).Cells were preincubated with insulin (0.1 ,uM) for 2 hr and thenwashed and incubated in Hanks/Hepes solution containing 0.2mM [1-14C]AIB (0.1 pCi/ml) for 30 min. Then the medium wasdiscarded and the cells were washed three times with ice-coldPi/NaCI and solubilized in 0.2 M NaOH and their radioactivitywas measured. Syntheses of intra- and extracellular proteinswere measured by incubating the cells in leucine-free Williamsmedium E containing L-[4,5-3H]leucine (0.1 ACi/nmol) andthen measuring radioactivity incorporated into intra- and ex-tracellular proteins as reported (10).

Preparations of Plasma Membranes from Various Tissues.Hepatic plasma membranes were purified from adult rat liveron a Percoll gradient as described (19); the membranes werepurified about 20-fold from the homogenate, judging by their5' nucleotidase activity. Renal basolateral membranes were pu-rified by the method of Sacktor et al. (20); about 6.5-fold pu-rification from a homogenate of renal cortex was achieved,judging from the activity of Na+,K+-ATPase. Synaptosomalmembranes were purified by the method of Jones and Matus(21) from a cerebral homogenate; these membranes were pu-rified 9-fold, judging by their Na+,K+-ATPase activity. Eryth-rocyte ghosts were prepared by the method of Dodge et al. (22).The purified membranes were sterilized by irradiation from anUV lamp at a distance of about 5 cm for 10 min.

Treatments of Purified Plasma Membranes. A suspensionof purified plasma membranes (2 mg of protein per ml) in Pi/NaCl was heated for 10 min in boiling water, cooled, and usedfor assay of modulation of activity. A suspension of the mem-branes (2.5 mg of protein per ml) in 10 mM Hepes buffer (pH7.5) containing 0.15 M NaCl and 10 mM CaCl2 was mixed with1/9 vol of 4 mg of trypsin preparation per ml, which contained50% of the activity, and incubated for 3 hr at 37C. The reactionwas stopped by the addition of the same volume of 6 mg ofsoybean trypsin inhibitor per ml. A control suspension of plasmamembranes was incubated without trypsin for 3 hr at 37TC and

then was mixed with trypsin previously inactivated with soy-bean trypsin inhibitor.

RESULTSEffects of Cell Density on Various Metabolic Activities of

Hepatocytes. Previously we showed that mature rat hepato-cytes do not proliferate when cultured at high cell density, evenin the presence of insulin and EGF, but that they can prolif-erate when cultured at low cell density with these hormones(9). These findings prompted us to examine whether other met-abolic activities of hepatocytes depend on cell density. Isolatedhepatocytes were plated at various inocula of 5 X 105 to 2.5 X107cells per 5-cm dish. Microscopic examination showed thatat the lowest cell density (2.5 x 10' cells per cm2) the cells werepresent in cultures singly or as small aggregates, whereas at thehighest cell density (12.5 x 104 cells per cm2) they were pres-ent as a sheet with tight cell-cell contact almost completelycovering the surface of the plate. Fig. 1 shows that at cell densi-ties of >5 x 1(Y cells per cm2, the inductions of T-NH2Taseand Ser-dehydratase by Dex and Dex with glucagon, respec-tively, were markedly increased, whereas the inductions of DNAsynthesis and Glc-6-P-DHase by insulin with EGF were mark-edly decreased. At confluency (12.5 x 104 cells per cm2), theinductions of T-NH2Tase by Dex and Ser-dehydratase by Dexwith glucagon reached maximal levels of 10-13 times the con-trol values, whereas in sparse cultures (2.5 x 104 cells per cm2)their inductions were less than twice the control levels. Thestimulation of cAMP formation by glucagon did not show celldensity-dependent regulation (data not shown). This fact sug-gests that the density-dependent change in induction of Ser-dehydratase by glucagon with Dex is not due to modulation ofthe glucagon receptor-adenylate cyclase system in the plasmamembranes.,An interesting finding was that Glc-6-P-DHase, which is

-

0o-4

M; 5

to3

x

0C6 3

-- 50

-I40.4.40to

,30s

*20

-lox 10C.)E5 0

0

100 I

80

.440 -

be

405C.)

20R

_~

2 4 6 8 10 12 14

Cell density, x 104 cells per cm2

40 -

.4.4 ev

930 u~

20 '

A.10 ID

0 4

FIG. 1. Cell density-dependent controls of DNA synthesis and in-ductions of Glc-6-P-DHase, T-NH2Tase, and Ser-dehydratase in pri-mary cultured hepatocytes. The numbers of cells indicated were in-oculated and cultured (see text). Activities were measured 24 hr afterhormone addition, except Glc-6.P-DHase activity, which was assayed45 hr after hormone addition. o, DNA synthesis induced by insulin (0.1M) with EGF (20 ng/ml); A, Glc-6-P-DHase activity induced by in-sulin with EGF; *, T-NH2Tase activity inducedby Dex (10 pM) A, Ser-dehydratase activity inducedby Dexwith glucagon (0.5 pM). The basal

activities (no hormone) at different cell densities were within ranges of1,468-1,697 dpm/hr per mg of protein for DNA synthesis, 6-8 mil-liunits (mU)/mg protein for Glc-6-P-DHase, 7-9mU/mg ofprotein forT-NH2Tase, and 4-6 mU/mg of protein for Ser-dehydratase, respec-tively. Values are means for duplicate experiments.

-

IF

I -

lk

Proc. Natl. Acad. Sci. USA 80 (1983)

EdI

Dow

nloa

ded

by g

uest

on

July

22,

202

1

Proc. Natl. Acad. Sci. USA 80 (1983) 7231

thought to be a key enzyme for the supply of NADPH in lipo-genesis, was induced much more at low cell density, whereasmalic enzyme, which is another key enzyme for the supply ofNADPH in lipogenesis, was induced by insulin and T3 muchmore at high cell density (Table 1). Therefore, Glc-6-P-DHasemay be more important in supplying ribose 5-phosphate for cellgrowth than for lipogenesis, as discussed earlier (23), whereasmalic enzyme may be more important for lipogenesis. Con-sistent with this idea, the density-dependent regulation of malicenzyme was found to correlate with that of synthesis of tri-glycerides.

Another interesting reciprocal relation in lipogenesis wasfound: cholesterol synthesis was greater at low cell density,whereas triglyceride synthesis was much higher at high celldensity (Table 1). Possibly, cholesterol is required for synthesisof cell membranes during growth, whereas triglycerides are de-posited in resting cells. Similarly, synthesis of cellular proteinwas stimulated at low cell density, whereas synthesis of serumproteins was not affected by cell density. The rate of AIB trans-port, which is generally thought to depend on cell growth, wasalso stimulated at low cell density. Thus, there seem to be twotypes of responses of various metabolic activities to change ofcell density: one is that of activities related to cell growth andanabolism, and the other is that of activities related to the dif-ferentiated characters of hepatocytes. As summarized in Table1, the properties related to cell growth tended to respond moreto low cell density, whereas those related to specific charactersof hepatocytes were stimulated more at high cell density thanat low cell density. The basal activities of these functions with-out added hormones also changed in the same way as with hor-mones, although these changes were <10% of those with hor-

Table 1. Reciprocal effects of cell density on growth-related andhepatocyte-specific functions in primary cultured hepatocytes

% of maximal activity

25,000 75,000 125,000per per per

Function Stimulators cm2 cm2 cm2DNA synthesis Insulin 100 36 5

+ EGFAB transport Insulin 100 67 42Protein synthesis

Intracellular 100 84 52Secreted 100 100 100

LipogenesisCholesterol 100 75 37Triglycerides Insulin 33 88 100

Enzyme inductionGlc-6-P-DHase Insulin 100 38 23

+ EGFMalic enzyme Insulin 34 50 100

+ T3T-NH2Tase Dex 11 47 100Ser-dehydratase Dex 13 44 100

+ glucagon

Experimental conditions were as for Fig. 1, except that Glc-6-P-DHase,malic enzyme, and lipogenesis were assayed 45 hr after hormone ad-dition. Values are expressed as % of maximal activity at the indicatedcell densities. The maximal values were 5.6 x 104 dpm/hr per mg ofprotein forDNA synthesis, 5.12 nmol/10 min per mgofprotein for AMBtransport, 4.23 x 105 dpm/hr per mg of protein and 1.3 x 105 dpm/hrper mg of protein for syntheses of intracellular and secreted proteins,respectively, 2.91 x 103 dpm/hr per mg of protein for cholesterol syn-thesis, 4.25 x 103 dpm/hr per mg of protein for triglyceride synthesis,48 mU/mg of protein for Glc-6P-DHase, 31 mU/mg of protein for malicenzyme, 89 mU/mg of protein for T-NH2Tase, and 38 mU/mg of pro-tein for Ser-dehydratase. Values are means of three to five experi-ments.

mones (data not shown). Moreover, some functions, such as thesyntheses of protein and cholesterol, changed markedly by celldensity without added hormones. These results suggest thatcell density affects both the basal expression of cellular func-tions and their responses to hormones.

Effects of Plasma Membranes on Growth and Specific Func-tions of Hepatocytes. Next we examined the mechanism re-sponsible for the cell-density dependent controls for these manycellular activities of adult rat hepatocytes. Two mechanisms arepossible: one is that the cell density-dependent regulations ofcell growth and various cellular activities may result from ac-cumulation of some soluble factors released from hepatocytesinto the medium. However, this possibility is excluded by thefact that addition of conditioned medium from cultures at highcell density to cultures at low density did not mimic the effectof high cell density (data not shown). The other possibility isthat direct cell-cell contact, perhaps involving interaction ofcomplementary cell surface components on adjacent cells, pro-vides a signal for regulation of cell growth and various differ-entiated characters of hepatocytes. If this possibility is correct,contact between intact cells would not necessarily be requiredfor density-dependent regulation. This consideration promptedus to examine whether cell surfice components of isolated plasmamembranes could mimic the effect of increase in cell densityon the growth and various cellular activities of cells at low den-sity.

Isolated plasma membranes were prepared from adult ratliver. The inductions of T-NH2Tase by Dex (Fig. 2) and malicenzyme by insulin with T3 (Fig. 3) were markedly enhanced inhepatocytes at low cell density by addition of these plasmamembranes. The effect of the plasma membranes was dose de-pendent and reached a maximum at 300 Ag of membrane pro-tein per 5-cm dish. This amount of membranes corresponds tothat of about 2.5 x 106 cells, calculated on the basis of the pu-rity and yield of membranes during purification. In contrast,

0._

20

;z

so

.6;P.

*

Eag.zErm

...

'S0

0

biD

x

._

4)

V4)

z

0 100 200 300 500Hepatic membrane, Ag per 5-cm dish

FIG. 2. Stimulation of T-NH2Tase induction and inhibition ofDNAsynthesis in sparse cultures of hepatocytes by addition of hepatic plasmamembranes. Hepatocytes were seeded at an initial cell density of 7 X105 cells per 5-cm dish (3.5 x 10' cells per cm2) and cultured. The mem-branes were added to the culture medium 2 hr after plating. Dex (10MM) for T-NH2Tase induction and insulin (0.1 pM) with EGF (20 ng/ml) for DNA synthesis were added to the culture medium 22 hr afteraddition ofthe membranes. T-NH2Tase activity andDNA synthesis wereassayed 24 hr after hormone addition. o, TAT activity without Dex; e,T-NH2Tase activity with Dex; *,DNA synthesis with insulin and EGF.The basal activity (mean ± SD) ofDNA synthesis was 1,605 ± 450 dpm/hr per mg of protein (three experiments). Values are means for dupli-cate experiments.

Cell Biology: Nakamura et al.

Dow

nloa

ded

by g

uest

on

July

22,

202

1

7232 Cell Biology: Nakamura et al.

50

0._.

20

bjD

s

*5...

cis

40 _

30

20

10

0

0 100 200 300 400Hepatic membrane, ,ug per 5-cm dish

FIG. 3. Stimulation of induction of malic enzyme in sparse cul-tures of hepatocytes by addition of hepatic plasma membranes. He-patocytes were plated and treated as for Fig. 2. Insulin (0.1 ^M) and T3(0.5 MM) were added to the culture medium 22 hr after addition of themembranes. Activity was measured 45 hr after hormone addition. o,No hormone; 9, insulin with T3. Values are means for duplicate ex-

periments.

the membrane preparation strongly inhibited DNA synthesisstimulated by insulin with EGF in cells at low cell density (Fig.2). This inhibition was also dose dependent and was maximalwith about 150 ,ug of membrane protein per 5-cm dish, whichis less than half the amount for maximal stimulations of T-NH2Tase and malic enzyme. Addition of plasma membranesalso affected the basal activities of the functions shown in Figs.2 and 3 and Table 2 of cells at low density without added hor-mones, although the changes were very small. These resultssuggest that the reciprocal controls of DNA synthesis and dif-ferentiated characters of hepatocytes are mediated by some cellsurface components via direct cell-cell contact.The ability of hepatic plasma membranes to mimic the re-

ciprocal effect of increased cell density on cellular activities wasalmost lost by digestion of the membranes with trypsin (Table2). Moreover, the modulator activity of hepatic plasma mem-branes was significantly decreased by heat treatment (in boilingwater for 10 min) of the membranes. Thus, the active com-

ponent(s) present in the cell surface may be a protein(s).Next we examined whether the cell surface modulator(s) is

specific to hepatocyte membranes. Table 3 shows that plasmamembranes purified from rat kidney, brain, and erythrocytes

Table 3. Tissue distribution of cell surface modulator inplasma membranes

T-NH2Taseactivity,

Dose, kqg of mU/mg of proteinPlasma protein per Without With Induction,

membranes 5-cm dish Dex Dex foldNone 6.0 9.2 1.53Liver 62.5 7.6 30.5 4.01

250 8.0 81.2 10.2Kidney 62.5 5.5 37.5 6.82

125 7.0 58.9 8.41250 7.5 87.0 11.6

Brain 62.5 8.5 20.6 2.42125 8.0 31.5 3.94250 7.9 37.2 4.71

Erytbrocytes 62.5 7.2 40.0 5.56125 6.5 49.3 7.58250 7.9 63.2 8.00

Experimental conditions were as described for Fig. 2. Values are meansfor duplicate dishes. Dex was added at 10 MM.

have similar modulator activities, as shown by their enhance-ment of T-NH2Tase induction, although the activities of prep-arations from different tissues varied. Renal membranes seemedto contain at least the same amount of cell surface modulator(s)as hepatic membranes, although the specific activities of themembranes could not be compared exactly because the puritiesof these membranes were different.

It should be mentioned that the effects of membranes onDNA synthesis and T-NH2Tase induction were similar whetherthe membranes were isolated from intact adult liver or fromhepatocytes cultured at high density.

DISCUSSIONCell-cell contact is known to be an important mechanism intissue formation, cytodifferentiation, and cell growth. Studieson this phenomenon have mainly been carried out on the slimemold (24-26) and chicken neural retina (27-29). Recently, itwas found that in slime molds, some surface components dif-fering from cell adhesive factors regulate induction of the post-aggregative enzyme UDP-glucose pyrophosphorylase (26).The phenomenon of "contact inhibition" of cell proliferation

has also been studied extensively in 3T3 cells and is usually ex-plained as due to either depletion of growth factors (30, 31) orsecretion of growth inhibitors (32, 33). As an alternative, Glaser

Table 2. Effects of various treatments of plasma membranes on their cell surfacemodulation activity

Dose, pg of T-NH2Tase activity, DNA synthesisHepatic protein per mU/mg of protein dpm x 10-4/hr per

membranes 5-cm dish Without Dex With Dex mg of protein %None 10.1 ± 0.3 18.8 ± 2.1 3.89 + 0.29 100Untreated 25 9.3 ± 0.2 41.7 ± 1.9

50 9.2 ± 0.7 86.5 ± 5.7 3.83 + 0.31 99100 17.8 2.2 141.9 ± 6.9 2.77 0.07 71200 33.9 4.7 147.1 ± 7.9 1.70 0.18 44

Trypsinized 100 8.4 ± 0.2 28.9 ± 1.6 3.82 ± 0.60 104200 7.4 ± 0.4 45.8 ± 2.7 4.28 ± 0.29 117

Heated 200 9.9 ± 3.2 55.8 ± 3.5 3.12 ± 0.18 80

Membranes were treated as described in the text. Hepatocytes were plated and treated as for Fig. 2.The membranes used in these experiments were purified 38-fold from liver homogenatesjudgingby their5' nucleotidase activity. Values are means ± SD of three experiments. Basal activity for DNA synthesiswithout hormone with different amounts of membranes was as shown in Fig. 2.

Proc. Nad. Acad. Sci. USA 80 (1983)

Dow

nloa

ded

by g

uest

on

July

22,

202

1

Proc. Natl. Acad. Sci. USA 80 (1983) 7233

and his colleagues (34, 35) proposed that it might be due to in-hibitors on the cell membrane, because they found that 3T3 cellgrowth was inhibited by addition of plasma membranes of 3T3cells. However, details of the regulatory mechanism of cell-cellcontact are still unknown.

In this work we showed that growth-related functions andhepatic characters were regulated reciprocally by cell densityand that these regulations were mediated by a cell surface com-ponent of the plasma membrane via direct cell-cell contact.The surface component seems to be a protein(s), because itsactivity was heat labile and sensitive to trypsin (Table 2). Re-cently, we succeeded in solubilizing the cell surface modula-tor(s) from hepatic membranes with a nonionic detergent, oc-tylglucoside, with 4 M guanidineHCl and partially purified themodulator(s) by Sephacryl S-400 filtration. This membrane-freemodulator(s) showed activities for both modulation of growth-related functions and hepatocyte-specific characters. The de-tails of properties of the partially purified modulator will bereported elsewhere.

These results in vitro suggest that various functions of he-patocytes in vivo at the mature state and during developmentand regeneration are all regulated by cell-cell contact and thatstable functions are not maintained unless the cells form tissueand receive complex regulation from their cytosocial environ-ment. In liver regeneration there is swelling of liver lobulesand disappearance of gap junctions morphologically and de-creased transfer of substances of low molecular weight and ofelectrical coupling functionally (36, 37). Liver tissue consists ofnot only parenchymal hepatocytes but also other cells, such ascells of the bile ducts, endothelium, and connective tissues.The participation of these other types of cells in regulation ofliver functions cannot be excluded, but the population of pa-renchymal cells is the largest and these cells are in contact witheach other. Therefore, the effect of cell-cell contact of hepa-tocytes on their activities must be the most important.The molecular mechanism of intracellular transduction of

membrane signals to regulate growth and liver functions is un-known. It should involve control of gene transcription of liverfunctions, because the changes in activity of Glc-6-P-DHaseand Ser-dehydratase in cells at different densities are parallelwith their mRNA levels and rates of enzyme syntheses (un-published data). Moreover, the cell density-dependent regu-lation does not depend on the type of hormone or its action butdepends on the type of cellular activity. These considerationssupport the idea that the mechanism of density-dependent reg-ulation is a nuclear event involving control of gene transcrip-tion.

Cell surface modulators involved in the regulations of cellfunctions by cell density were not present only on hepatocytemembranes but were also found in the plasma membranes ofother tissues (Table 3). Their wide distribution seems reason-able, because cell surface modulators do not induce new ex-pression of dormant genes but regulate phenotypic genes thathave already been switched on during terminal differentiation.Therefore, cell surface modulators of various tissues could havea common mechanism in regulating the phenotypes of varioustissues, although these phenotypes are tissue specific.

It would be interesting to know whether the same factor incell membranes is responsible for inhibition of growth-relatedfunctions and stimulation of hepatocyte-specific characters. Thisproblem must be examined by further purification and char-acterization of the factor(s) responsible for these activities. Ourrecent findings show that solubilized and partially purified fac-tor regulates reciprocally several activities reported in this pa-per. This suggests that density-dependent changes in activitiesmay all be mediated by one or a few kinds of factors. This in-formation would be helpful in understanding the process of as-

sembly of liver cells to organize liver tissue in vivo and themechanisms of change of gene expression in liver during de-velopment, regeneration, and carcinogenesis.

This work was supported by grants for Cancer Research (57010052)and Scientific Research (00544029) from the Ministry of Education,Science and Culture of Japan.

1. Ichihara, A., Nakamura, T. & Tanaka, K. (1982) Mol. Cell. Biochem.43, 145-160.

2. Nakamura, T., Shinno, H. & Ichihara, A. (1980)J. Biol. Chem. 255,7533-7535.

3. Nakamura, T., Aoyama, K., Tomomura, A. & Ichihara, A. (1981)Biochim. Biophys. Acta.678, 91-97.

4. Noda, C., Nakamura, T. & Ichihara, A. (1983)J. Biol. Chem. 258,1520-1525.

5. Richman, R. A., Claus, T. H., Pilds, S. J. & Friedman, D. L. (1976)Proc. Natl. Acad. Sci. USA 73, 3589-3593.

6. Tomita, Y., Nakamura, T. & Ichihara, A. (1981) Exp. Cell Res. 135,363-371.

7. McGowan, J. A., Strain, A. J. & Bucher, N. L. R. (1981)J. Cell.Physiol. 108, 353-363.

8. Michalopoulos, G., Cianciulli, H. D., Novotny, A. R., Kliger-man, A. D., Strom, S. C. & Jirtle, R. L. (1982) Cancer Res. 42,4673-4682.

9. Nakamura, T., Tomita, Y. & Ichihara, A. (1983) J. Biochem. (To-kyo) 94, 1029-1035.

10. Tanaka, K., Sato, M., Tomita, Y. & Ichihara, A. (1978)J. Biochem.(Tokyo) 84, 937-946.

11. Seglen, P. 0. (1976) Methods Cell Biol. 13, 29-83.12. Nakamura, T., Yoshimoto, K., Aoyama, K. & Ichihara, A. (1982)

J. Biochem. (Tokyo) 91, 681-693.13. Wise, E. M. & Ball, E. G. (1964) Proc. Nati. Acad. Sci. USA 52,

1255-1263.14. Suda, M. & Nakagawa, H. (1970) Methods Enzymol. 17B, 346-

351.15. Granner, D. K. & Tomkins, G. M. (1970) Methods Enzymol. 17A,

633-637.16. Bradford, M. (1976) Anal. Biochem. 72, 248-254.17. Bligh, E. G. & Dyer, W. J. (1959) Can. J. Biochem. Physiol. 37,

911-917.18. Kletzien, R. F., Pariza, M. W., Becker, J. E., Potter, V. R. &

Butcher, F. R. (1976) J. Biol. Chem. 251, 3014-3020.19. Nakamura, T., Tomomura, A., Noda, C., Shimoji, M. & Ichihara,

A. (1983) J. Biol. Chem. 258, 9283-9289.20. Sacktor, B., Rosenbloom, I. L., Liang, T. & Cheng, L. (1981)

J. Membr. Biol. 60, 63-71.21. Jones, D. H. & Matus, A. I. (1974) Biochim. Biophys. Acta 356,

276-287.22. Dodge, J. T., Mitchell, C. & Hanahan, D. J. (1983) Arch. Biochem.

Biophys. 100, 119-130.23. Yoshimoto, K., Nakamura, T. & Ichihara, A. (1983) J. Biol. Chem.

258, in press.24. Okamoto, K. & Takeuchi, I. (1976) Biochem. Biophys. Res. Com-

mun. 72, 739-746.25. Cheung, S., Landfear, S. M., Blumberg, D. D., Cohen, N. S. &

Lodish, H. F. (1981) Cell 24, 785-797.26. Kaleko, M. & Rothman, F. G. (1982) Cell 28, 801-811.27. Morris, J. E. & Moscona, A. A. (1971) Dev. Biol. 25, 420-444.28. Hausman, R. E. & Moscona, A. A. (1976) Proc. Natl. Acad. Sci.

USA 73, 3594-3598.29. Thiery, J.-P., Brackenbury, R., Rutishauser, U. & Edelman, G.

M. (1977)1. Biol. Chem. 252, 6841-6845.30. Holley, R. W. & Kiernan, J. A. (1968) Proc. Natl. Acad. Sci. USA

60, 300-304.31. Stoker, M. G. P. (1973) Nature (London) 246, 200-203.32. Steck, P. A., Blinis, J., Voss, P. G. & Wang, J. L. (1982)J. Cell

Biol. 92, 523-530.33. Holley, R. W., Bohlen, P., Fava, R., Baldwin, J. H., Kleeman, G.

& Armour, R. (1980) Proc. Natl. Acad. Sci. USA 77, 5989-5992.34. Whittenberg, B. & Glaser, L. (1977) Proc. Natl. Acad. Sci. USA

74, 2251-2255.35. Whittenberg, B., Raben, D., Lieberman, M. A. & Glaser, L. (1978)

Proc. Natl. Acad. Sri. UYSA 75, 5457-5461.36. Yee, A. & Revel, J.-P. (1978) J.Cell Biol. 78, 554-564.37. Meyer, D. J., Yancey, S. B. & Revel, J.-P. (1981)J. Cell Biol. 91,

505-523.

Cell Biology: Nakamura et al.

Dow

nloa

ded

by g

uest

on

July

22,

202

1