Enzyme Induction and Modulation
-
Upload
others
-
View
4
-
Download
0
Embed Size (px)
Citation preview
Developments in molecular and cellular biochemistry
Najjar, Victor A., ed.: Biological Effects of Glutamic Acid and Its
Derivatives, 1981. ISBN 90-6193-841-4
Najjar, Victor A., ed.: Immunologically Active Peptides, 1981. ISBN
90-6193-842-2
Enzyme Induction and Modulation
edited by
v.A. NAJJAR Division of Protein Chemistry Tufts University School
of Medicine Boston, Massachusetts, U.S.A.
Reprinted from Molecular and Cellular Biochemistry Volumes 53/54,
1983
1983 MARTINUS NIJHOFF PUBLISHERS a member of the KLUWER ACADEMIC
PUBLISHERS GROUP BOSTON / THE HAGUE / DORDRECHT / LANCASTER
Distributors
for the United Stares and Canada: Kluwer Boston, Inc., 190 Old
Derby Street, Hingham, MA 02043. USA for aff other countries:
Kluwer Academic Publishers Group. Distribution Center, P.O. Box
322, 3300 AH Dordrecht, The Netherlands
Library or CongrtSS Cltlloging in Publication Dltl
LI.,,,, .f CHI"" (. ,.1,,1., I. P.bll"H • • D." Mooi= .=t ry ""d.r
title;
(Ile ... l<>p .. nto Ie 1I01uular ""d ~dlul.r blcehu.1.try)
Orl,laailT pulol1lb*<l. ..... S3 .. 5. or Nool.eular ..,d
.dlular blochnhtq. 1. taQ"U l_eU"" .... M~lIeo. eo .. ¥I.
leetu:-...
2. Llvu e.ll ... _A4~ ........ ..,. •• lee"'...... 3. eUI .l:.l.lur
..... Add. ........ . . """" l.el........ 1. lo.JJI.l" . nel or".
II. !!o1«lIIar Uld e.lllll~ hloc.bea1.tq. Itt. Se r in. W'6ol.!'517
1983 m '.Ol925 B:HI13t1
ISBN-13: 978- 1-4613-3881-9 001 : 10. 1007/978- 1-461
3-3879-6
Copyright
e-ISBN- 13: 978- 1-46 13-3879-6
© 1983 by Martinus Nijhorr Publishers, Boston. Soflcover reprint of
the hardcover lSI edition 1983 All rights reserved. No part of this
publication may be reproduced. stored in a retrieval system, or
transmitted in any rorm or by any means. mechanical, photocopying,
recording. or otherwise. without the prior written permission of
the publishers, Martinus Nijhorr Publishers, 190 Old Derby Street,
Hingham. MA 02043. USA.
Contents
Part I 9 T. D. Gelehrter, P. A. Barouski-Miller, P. L. Coleman and
B. J. Cwikel: Hormonal regulation of
plasminogen activator in rat hepatoma cells II J. W. Grisham: Cell
types in rat liver cultures: their identification and isolation 23
C. Guguen-Guillouzo and A. Guillouzo: Modulation offunctional
activities in cultured rat hepatocytes 35 D. F. Haggerty, E. B.
Spector, M. Lynch, R. Kern, L. B. Frank and S. D. Cederbaum:
Regulation of
expression of genes for enzymes of the mammalian urea cycle in
permanent cell-culture lines of hepatic and non-hepatic origin
57
R. Barouki, M.-N. Chobert, J. Finidori, M.-C. Billon and J.
Hanoune: The hormonal induction of gamma glutamyltransferase in rat
liver and in a hepatoma cell line 77
L. J. Crane and D. L. Miller: Plasma protein induction by isolated
hepatocytes 89
Part II III D. K. Granner and J. L. Hargrove: Regulation of the
synthesis of tyrosine aminotransferase: the
relationship to mRNATAT 113 J. M. Masserano and N. Weiner: Tyrosine
hydroxylase regulation in the central nervous system 129
Part III 153 A. J. Fulco, B. H. Kim, R. S. Matson, L. O. Narhi and
R. T. Ruettinger: Nonsubstrate induction of a
soluble bacterial cytochrome P-450 monooxygenase by phenobarbital
and its analogs 155 G. Kikuchi and T. Yoshida: Function and
induction of the microsomal heme oxygenase 163
Part IV 185 W. L. Miller, D. C. Alexander, J. C. Wu, E. S. Huang,
G. K. Whitfield and S. H. Hall: Regulation of
,B-chain mRN A of ovine follicle-stimulating hormone by
17,B-estradiol 187 W. E. G. Muller, A. Bernd and H. C. Schroder:
Modulation of poly (A)( +)mRN A-metabolizing and
transporting systems under special consideration of microtubule
protein and actin 197 P. H. Pekala and J. Moss: 3T3-Ll preadipocyte
differentiation and poly(ADP-ribose) synthetase 221 P. K. Sarkar
and S. Chaudhury: Messenger RN A for glutamine synthetase 233
Part V 245 R. Sasaki and H. Chiba: Role and induction of
2,3-bisphosphoglycerate synthase 247 J. G. Cory and A. Sato:
Regulation of ribonucleotide reductase activity in mammalian cells
257 M. R. Waterman and R. W. Estabrook: The induction of microsomal
electron transport enzymes 267 V. Rubio, H. G. Britton and S.
Grisolia: Activation of carbamoyl phosphate synthetase by
cryoprotectants 279
6
G. C. Ness: Regulation of 3-hydroxy-3-methylglutaryl coenzyme A
reductase 299 J. W. Porter and T. L. Swenson: Induction of fatty
acid synthetase and acetyl-CoA carboxylase by
isolated rat liver cells 307
Preface
In addition to performing its prime function as a vehicle for
scientific communications of varied colora tions, Molecular and
Cellular Biochemistry is again focusing on two subjects which it
treats in depth. One of these is a book issue dealing with the
transglutaminase reaction. The other is this issue that deals with
induction and modulation of enzymes. This is a very broad subject
that calls for broader coverage than could be included in one book
issue. However, I have elected to include only certain
contributions that serve as general examples of the principles
involved.
There are six articles on enzyme regulation in hepatocyte culture.
These include arginase and argino-succi nate synthetase,
y-glutamyl transferase and plasminogen activitor. Other regulatory
enzymes that are discussed are protein kinases,
2,3-bisphosphoglycerate synthetases, carbamoyl phosphate
synthetase, heme oxygenase, cytochrome P-450, tyrosine hydroxylase,
fatty acid synthetase, acetyl eoA carboxylase, among others. Also
included is the regulation of several enzyme messengers RNAs.
As in the past, subscribers to the journal will receive this double
volume as a continuation of regular issues of the journal Molecular
and Cellular Biochemistry. This double volume will also be
available separately in the form of a book issue to interested
readers through Martinus Nijhoff Publishers.
v.A. Najjar
Part I
Molecular and Cellular Biochemistry 53;54,11-21 (1983). © 1983,
Martinus 1\ ij hoff Publishers. Boston. Printed in The
Netherlands.
Hormonal regulation of plasminogen activator in rat hepatoma
cells
Thomas D. Gelehrter, Patricia A. Barouski-Miller, Patrick L.
Coleman and Bernard J. Cwikel Departments of Internal Medicine and
Human Genetics, University of Michigan Medical School, Ann Arbor,
MI48109, U.S.A,
Summary
Plasminogen activators are membrane-associated, arginine-specific
serine proteases which convert the inactive plasma zymogen
plasminogen to plasmin, an active, broad-spectrum serine protease.
Plasmin, the major fibrinolytic enzyme in blood, also participates
in a number of physiologic functions involving protein processing
and tissue remodelling, and may play an important role in tumor
invasion and metastasis. In HTC rat hepatoma cells in tissue
culture, glucocorticoids rapidly decrease plasminogen activator (P
A) activity. We have shown that this decrease is mediated by
induction of a soluble inhibitor of P A activity rather than
modulation of the amount of PA. The hormonally-induced inhibitor is
a cellular product which specifically inhibits PA but not plasmin.
We have isolated variant lines of HTC cells which are selectively
resistant to the glucocorticoid inhibition ofP A but retain other
glucocorticoid responses. These variants lack the hormonal
ly-induced inhibitor; P A from these variants is fully sensitive to
inhibition by inhibitor from steroid-treated wild-type cells.
Cyclic nucleotides dramatically stimulate P A activity in HTC cells
in a time- and concentra tion-dependent manner. Paradoxically,
glucocorticoids further enhance this stimulation. Thus glucocorti
coids exert two separate and opposite effects on P A activity. The
availability of glucocorticoid-resistant variant cell lines,
together with the unique regulatory interactions of steroids and
cyclic nucleotides, make HTC cells a useful experimental system in
which to study the multi hormonal regulation of plasminogen
activator.
Introduction
Plasminogen activators (PAs) are membrane associated
arginine-specific serine proteases found in a variety of tissues
(I). P A selectively hydro lyses a single Arg-Val bond of the
plasma zymo gen, plasminogen, to yield the active serine pro
tease, plasmin, the major fibrinolytic activity in blood (2, Fig.
I). Plasmin is a broad-spectrum endopeptidase which can act on a
variety of pro teins. Because plasminogen is present in plasma in
relatively high concentrations (1.5 to 2 ~M, or 0.5% of all plasma
proteins), the plasminogen ac tivator-plasmin cascade provides
considerable po tential proteolytic activity (2, 3). Thus
generation
of plasmin both amplifies P A activity and broad ens the substrate
specificity. In addition to plas min's well-known role in
fibrinolysis, it is also in volved in many normal physiologic
functions which involve protein processing, cell migration and
tissue remodelling (I, 3, 4, Table I). By acting directly on fibrin
and directly or indirectly (via activation of procollagenase) on
connective tissue matrix (5,6), the plasminogen activator-plasmin
cascade may al so play an important role in tumor invasion and
metastasis (1,3,4,6).
Not surprisingly for an enzyme of such biological importance,
plasminogen activator is subject to regulation by a variety of
effectors (see 7 for re view). Steroid (8-16) and polypeptide
hormones
12
Table I. Physiological/ pathological processes mediated by
plasminogen activator/ plasmin.
Fibrinolysis Proteolytic processing of cellular and serum
proteins
Complement activation Kinin formation Proinsulin conversion
Migration of macrophages during inflammation Tissue remodelling and
destruction
Rupture of the ovarian follicle during ovulation Implantation of
the mouse embryo Involution of the mammary glandfollowing
laclation
Neoplasia Invasiveness lvlelasla/is
(17,18), growth factors (4), cyclic nucleotides (13, 14,17-22),
retinoids(22-25), lectins(26) and tumor promoters (4, 22, 25, 27)
are all modulators of P A activity in some tissues. Inhibitors of P
A and! or plasmin are also important regulators of protease
activities (2, 28-33), and in some experimental sys tems these
inhibitors may also be subject to physio logic regulation (10,
11,28,33). The mechanisms by which P A activity is regulated are
largely unknown, but could involve regulation of the rates of
synthe sis or degradation of the P A protein, activation of P A
itself or of a P A precursor, or regulation of specific inhibitors
of PA.
HTC cells are an established line of rat hepatoma
cells in long-term tissue culture, which provide a favorable
experimental system for studying the regulation of plasminogen
activators. This line is extremely well characterized with respect
to hor monal regulation of multiple functions, particular ly for
the actions of glucocorticoids, insulin, and cyclic nucleotides
(34-38). Furthermore, it is possi ble to isolate variant HTC cell
lines altered in hor monal regulation of various properties, and
several such lines have been described (9, 39, 40). Over the past
several years we have exploited these features of HTC cells to
study the hormonal regulation of plasminogen activator and the role
ofPA in various cellular functions modulated by hormones. We have
described two unique mechanisms of regula tion of P A: first, the
glucocorticoid induction of a specific inhibitor of plasminogen
activator (10, II, 28); and second, a paradoxical effect of
glucocorti coids on P A regulation in which glucocorticoids alone
inhibit P A activity but together with cyclic nucleotides enhance
the dramatic stimulation of PA activity by the latter (41).
Materials and methods
Cell culture HTC cells were routinely grown in spinner or
monolayer culture in Minimal Essential Medium (Eagle's) without
antibiotics, modified to contain 50 mM tricine, a nonvolatile
buffer, 0.5 g/ I sodium bicarbonate, and supplemented with2 mM
glutam ine and 5% calf and 5% fetal bovine serum. Experi ments
were performed in a chemically-defined me dium identical to the
growth medium ~xcept that it lacked serum and was supplemented with
neomycin and, where applicable, with 0.1 % bovine serum al
bumin.
Assays of plasminogen activator P A was routinely assayed in either
conditioned
medium (CM) or 0.2% Triton X-100 extracts of cells using either an
125I-fibrinolytic assay (42) or the esterolytic assay (43, 44)
developed in this la boratory. HTC cells have no demonstrable
plas minogen-independent fibrinolytic, caseinolytic, or estero
lytic activity. Direct addition of dexametha sone or cyclic
nucleotides to the assay mixture has no effect on PA activity.
Inhibitory activity was measured by incubating CM or cell extracts
to be
13
tested with either CM or cell extracts of untreated HTC cells (as a
source of P A) or urokinase (a human urinary plasminogen activator)
for 20 min at 37 0 C or 30 min at 25 0 C prior to assaying P A
activity. Inhibitory activity was quantitated by ti trating serial
dilutions of CM or cell extracts from dexamethasone-treated cells
against a fixed amount of UK or HTC cell PA.
Characterization of plasminogen activator in HTC cells
Multiple molecular weight forms of P A were separated by SDS
polyacrylamide gel electropho resis under nonreducing conditions.
Following electrophoresis, SDS was removed from the gels and
proteins by exchange with the nonionic deter gent Triton X-100,
allowing recovery ofPA activity (45). P A activity was then
localized and assayed either by elution ofPA from homogenized gel
slices and assay of plasminogen-dependent fibrinolysis, by the
fibrin-agar underlay method of Granelli-Pi perno & Reich (45),
or by plasminogen-dependent caseinolytic activity in the gel (46,
47).
Enucleation of HTC cells Anucleate HTC cells (cytoplasts) were
prepared
by centrifugation of cells from suspension cultures through a
discontinuous Ficoll gradient in the pres ence of cytochalasin B.
The efficiency of enuclea tion was routinely greater than 92%.
Cytoplast preparations maintained their membrane integrity for at
least 24 hours in culture in the absence of serum (48).
Isolation of variant HTC cells An agar-fibrin overlay technique (9,
49) was used
to identify colonies with plasminogen-dependent fibrinolytic
activity. Colonies which expressed P A activity in the presence of
dexamethasone (and were thus presumably resistant to the dexametha
sone inhibition of P A) were isolated through the agar-fibrin
overlay and propagated in the absence of dexamethasone. Variant
cell lines highly resis tant to the dexamethasone inhibition of P
A activity were obtained after several cycles of such treatment
(9,39).
Materials Tissue culture media and sera were obtained
from Gibco. Dexamethasone was the kind gift of
14
Results
Glucocorticoid regulation of plasminogen activa tor
Plasminogen activator actiVIty is found in the membrane fraction of
HTC cells from which it can be released by detergents such as
Triton X-I 00 (10). Activity is also found in serum-free medium
condi tioned by HTC cells (9, 41, 43). Analysis of PA activity
from both cell extracts and conditioned
94K __ 68K
:3 C D
Fig. 2. Schematic diagram of multiple molecular weight forms of HTC
cell plasminogen activator. Monolayer cultures of HTC cells were
incubated for 24 hours in serum-free medium without hormones (Lane
A). withO.II'M dexamethasone(Lane B), with 3 mM 8-bromo-cAMP; I mM
MIBX (Lane C), or with 3 mM 8-bromo-cAMPfI mM MIBX;O.II'M
dexamethasone (Lane D). Conditioned media were subjected to SDS
polyacrylamide gel electrophoresis and PA activity was localized on
a fibrin-agar underlay as described in Materials and methods.
Molecular weight markers are phosphorylase B (94K). bovine serum
al bumin (68K), ovalbumin (43K) and soybean trypsin inhibitor
(2IK).
o 4
Time (hrs)
6
Fig. 3. Time course of dexamethasone inhibition of HTC cell
plasminogen activator. HTC cells were incubated for the times
indicated in serum-free medium containing I I'M dexametha sone.
Fibrinolytic activity of cell extracts was measured as de scribed
in Materials and methods and was normalized for the amount or
protein in each sample. Each point represents the average of
duplicate assays on a single culture. Reproduced from reference
10.
medium on SDS polyacrylamide gels under nonre ducing conditions
reveals two major molecular weight forms of P A activity of 110000
and 66000 daltons. A minor band of activity is sometimes observed
at 33 000 daltons (Fig. 2). There does not appear to be any
difference between the molecular weights of the PAs associated with
the cell and those released into the medium.
Glucocorticoids rapidly inhibit the activity ofPA in HTC cells.
Inhibition is half-maximal after ap proximately 90 min and maximal
after 4 to 6 hours of incubation (Fig. 3). The magnitude of
inhibition is usually 75 to 100%. Inhibition of P A activity is
also observed in conditioned medium at later times than in cell
extracts. Half-maximal inhibition is achieved at 5 nM
dexamethasone, the same concen tration that half-maximally induces
tyrosine ami notransferase and half-maximally inhibits amino acid
transport in HTC cells; maximal inhibition is achieved at 10 to 100
nM dexamethasone (39).
Dexamethasone could inhibit PA activity by de creasing the amount
ofP A protein or by decreasing its activity (possibly by inducing
an inhibitor of this protease), or by some combination of these
mech anisms. When increasing amounts of an extract of
dexamethasone-treated cells are mixed with a fixed amount of an
extract from control cells (as a source of P A activity), there is
a concentration-dependent inhibition of P A activity, demonstrating
that the dexamethasone-treated cells contain an inhibitor of
40 ?: 0
~:~ ~ g 30 U .2 "'"0 (,) f! 20 ~n; go . .::'" 10 ~O
*- o
III conditioned medium
Fig. 4. Inhibition of HTC cell plasminogen activator by condi
tioned medium from dexamethasone-treated HTC cells. HTC cells were
incubated for 18 hours with 0.1 I'M dexamethasone (e) or without
hormones(O). Increasing amounts of conditioned medium from these
cultures were incubated for 30 min at 25°C with 10 I'g of cell
extract from untreated HTC cells (as a source of P A) and
fibrinolytic activity assayed as described. Each point represents
the average of duplicate assays.
P A (10, 11). Cell fractionation experiments have demonstrated that
the inhibitor is found primarily in the soluble 100 000 X g
supernatant fraction (10), as well as in medium conditioned by
dexametha sone-treated cells (Fig. 4).
An intact nucleus is required for this hormonal regulation of PA
activity. Cytoplasts prepared by
e 100 Control "E
20 e '" 1;;
0 w 5
III conditioned medium
Fig. 5. Inhibition of human urokinase by conditioned medium from
dexamethasone-treated HTC cells. HTC cells were incu bated for 18
hours with 0.1 I'M dexamethasone (e) or without hormones (0).
Increasing amounts of conditioned medium were incubated for 20 min
at 37 0 C with 5 milliPloug units of human urokinase and
esterolytic activity assayed as described (44). Each point
represents the average of duplicate assays.
15
centrifugation through Ficoll gradients in the pres ence of
cytochalasin B maintain P A activity at lev els comparable to or
higher than those of intact cells. However, anucleate HTC cells are
not respon sive to glucocorticoid regulation of P A activity.
Dexamethasone does not decrease either intracellu lar or
extracellular P A activity of the anucleate cells, or induce
production of the soluble inhibitor of PA activity (52).
The dexamethasone-induced inhibitor of P A ac tivity (PAl) also
inhibits the plasminogen-depen dent fibrinolytic or estero lytic
activity of HeLa cells (Coleman, unpublished work) and of human
uro kinase (Fig. 5), as well as P A from HTC cells. Thus, PAl
inhibits both major immunochemical types of P A, urokinase-like and
tissue activator (I). In con trast, plasmin is not inhibited by
conditioned medi um from HTC cells incubated with dexamethasone.
The specificity of the inhibitor for plasminogen activation was
demonstrated directly by the inhibi tion of urokinase-catalyzed
activation of 125I-plas minogen to 1251-plasmin (28). These
results show that the inhibition is not directed against plasmin,
but is specific for plasminogen activator.
Because certain cell types can take up serum pro tease inhibitors
from the serum in medium (31, 33, 53) and release them to
serum-free medium condi tioned by these cells (31), we
investigated the origin of this hormonally-induced inhibitor in HTC
cells. SF HTC-HI, a line of cells selected for their ability
to-grow in serum-free medium (54), were grown for 76 days (at least
30 generations) in the presence or absence of serum; dexamethasone
induced equiva lent amounts of inhibitory activity in cells grown
under either condition. Furthermore, the inhibitory activity from
HTC cells is stable to pH 3 for 2 hours at 37 DC, a treatment which
inactivates fibrinolytic inhibitors in serum. These results
indicate that the dexamethasone-induced inhibitor is a cellular
pro duct which differs from serum-derived fibrinolytic inhibitors
(28).
The inhibitor is inactivated by boiling and by treatment with
pepsin under acidic conditions, sug gesting that it is a protein.
The P A inhibitory activi ty in CM from dexamethasone-treated
cells mi grates as a single band of approximately 45 000 daltons
upon SDS polyacrylamide gel electropho resis under nonreducing
conditions (Cwikel & Ge lehrter, unpublished work). PAl is
clearly different from the 38 000 dalton protease inhibitor,
protease
16
nexin, described by Low et al. (32) which inacti vates both
thrombin and urokinase. Conditioned medium from
dexamethasone-treated cells, which readily inhibited urokinase, had
no effect on thrombin activity, even in the presence of heparin
which accelerates the thrombin-protease nexin in teraction
(Coleman & Gelehrter, unpublished work).
We have begun to investigate the interaction of P Al with various
PAs. Preliminary evidence sug gests that P AI forms an
irreversible covalent com plex with urokinase. I n contrast, its
interaction with P A from HTC cells can be reversed under
conditions of SDS polyacrylamide gel electropho resis. Taking
advantage of the latter fact, we have asked whether dexamethasone
decreases P A activi ty directly, in addition to the effects
mediated by P AI. Cell extracts and CM from dexamethasone treated
cells, which have less than 10% of control PA activity when assayed
by the fibrin plate assay,
AGAR-FIBRIN OVERLAY TECHNIQUE
PREPARATION OF AGAR-FIBRIN
POUR OYER CELLS p .... TEO OUT ON 601ft.,. CMSH(S
Figure 6. Schematic diagram of the agar-fibrin overlay tech nique
used to isolate variant HTC cell lines resistant to the
glucocorticoid inhibition of PA. Experimental details in Mate
rials and methods and in references 9 and 39. Reproduced from
reference 9.
exhibit PA activity comparable to the controls when measured by the
fibrin-agar underlay (Fig. 2) or by direct assay of gel eluates
following S OS polyacrylamide gel electrophoresis. These results
suggest that PA and PAl dissociate during electro phoresis, and
that there is no decrease in the amount of PAin cell extracts or
medium from dexamethasone-treated cells (Cwikel, Coleman,
Barouski-Miller & Gelehrter, unpublished work).
Isolation and characterization oj variant hepatoma cells resistant
to hurmunal regulation oJplasmino gen activator
Utilizing the agar-fibrin overlay technique (Fig. 6), we have
isolated a number of variant HTC cell lines which are highly
resistant to the dexametha sone inhibition of PA activity (9, 39).
These var iants are resistant to a concentration of dexametha
sone I 000 times greater than that necessary to completely inhibit
P A activity in wild-type cells (39). The nature of the defect in
these cells was defined by mixing experiments analogous to those
described above. Dexamethasone-treated variant cells show no PAl
activity, whereas the PA from these cells is fully sensitive to
inhibitor from gluco corticoid-treated wild-type cells. Thus the
basis of hormone resistance in the variants appears to be the
failure of dexamethasone to induce PAl (10, II).
The growth rate and cloning efficiency of variant and wild-type
lines, both on plastic and in soft agar, are indistinguishable.
Morphologically there are no consistent differences between the
variant and wild type cells. Fluctuation analyses support the hy
pothesis that resistance to dexamethasone arises randomly and is
not induced by the hormone. It was not possible to determine the
rate at which stable variant cells arise, but only to quantitate
the frequency of the first step in this process. The fre quency
with which colonies from a wild-type popu lation form fibrinolytic
plaques in the presence of dexamethasone is high (approximately
10-3) and this rate is not altered by treatment with two differ
ent mutagens: ethylmethane sulfonate and UV light. This observation
suggests that mutations are not the primary cause of resistance in
this cell line. The karyotypic variability of HTC cells raises the
possibility that variants might arise from chromo somal
segregation events (39).
Biochemical analysis of the variant cell lines
demonstrates that they have a lesion specific for the regulation of
plasminogen activator. The hormonal resistance is apparently not
due to deficient or de fective steroid receptor function since the
variants show wild-type induction of tyrosine aminotrans ferase.
The lesion iIi these variants must therefore be at some step distal
to the entry of the hormone receptor complex into the nucleus. We
have also shown that 6 of 7 variant cell lines tested show
wild-type inhibition of amino acid transport by glucocorticoids
(one variant is partially resistant to the inhibitory effect of
dexamethasone on trans port). These findings indicate that there
is not a generalized resistance to all membrane-associated
dexamethasone responses, but that the cells are selectively
resistant to the inhibition of P A (39). Other variant f-!:TC cell
lines which are selectively resistant to the dexamethasone
induction of tyro sine aminotransferase (40) show wild-type
inhibi tion of P A activity by glucocorticoids (55). These results
indicate that the various glucocorticoid-me diated responses in
HTC cells are independently rather than coordinately regulated (11,
55). The selective resistance of these HTC variants is unique and
in contrast to the great majority of glucocorti coid-resistant
variant cell lines previously described, essentially all of which
have been shown to have deficient or defective glucocorticoid
receptors (56, 57).
We utilized these variant lines to study the role of plasminogen
activator in the hormonal regulation of other membrane properties.
HTC cells exhibit increased levels of adhesion to a substrate as
well as decreased P A activity when incubated with dexa methasone
(58). In a variety of cultured cells, adhe sion to a substrate
requires specific cell surface glycoproteins and intact
cytoskeletal elements, and there is evidence that plasmin may
affect both of these components of adhesion (59-61). Using the
variant HTC cell lines, we tested the hypothesis that dexamethasone
induces cell adhesion by decreasing the activity of P A, which in
turn allows the accumu lation of specific cell surface
glycoproteins neces sary for adhesion. If this hypothesis were
correct there should be little or no dexamethasone inhibi tion of
adhesion in the variant cell lines. A sensitive quantitative assay
was developed which measures the strength of attachment of
radioactively-labeled cells to glass scintillation vials following
exposure to shearing force (62). We found that dexametha-
17
sone induces the adhesiveness of variant HTC cells to the same
extent as that of wild-type cells. This was true when adhesion was
measured in serum free medium, in serum-containing medium, or in
serum-containing medium depleted of, or reconsti tuted with,
plasminogen. These results indicated that neither P A itself nor
plasmin plays a major role in cell adhesion in HTC cells and
suggested that the dexamethasone induction of adhesion might oper
ate through the synthesis of cell surface glycopro teins (II ,
62).
Cyclic nucleotide regulation of plasminogen activa tor
activity
Incubation of HTC cells with cAMP deriva tives stimulates
cell-associated P A activity 8- to 20-fold and extracellular P A
activity 30- to I 300- fold. This time- and concentration-dependent
in crease is enhanced by phosphodiesterase inhibi tors such as
l-methyl-3-isobutylxanthine (MIBX). Maximal stimulation of P A
activity is observed at 3 mM 8-bromo-cAMP and half-maximal stimula
tion at 0.2 mM. A similar concentration depend ence is noted with
dibutyryl-cAMP. N6-monobu tyryl-cAMP also stimulated P A activity
but cAMP itself did not; dibutyryl-cGMP inhibited PA activi ty.
Increases in PA activity in cells incubated with 8-bromo-cAMP are
first detectable at 4 hours in cell extracts and 6 hours in
extracellular medium, and are maximal by 8 and 12 hours,
respectively (Fig. 7). MIBX increases the level of maximal stim
ulation, but does not significantly alter the time course
(41).
As noted above, dexamethasone decreases P A activity by induction
of an inhibitor. Paradoxically, dexamethasone added simultaneously
with cAMP derivatives causes a further 4-fold enhancement of the
cyclic nucleotide stimulation of P A activity (41). Analysis of the
molecular weight forms ofPA under these conditions indicates that
the same 110 K and 66 K forms are present in cells treated with
cyclic nucleotides or cyclic nucleotides plus dexamethasone as in
control and dexametha sone-treated cells. In addition a minor 33 K
dalton form of PAis sometimes found in cells incubated with cAMP
derivatives plus or minus dexametha sone (Fig. 2). Dexamethasone
also profoundly al ters the time course of the cyclic nucleotide
en hancement of P A activity: increased activity is
18
8.0
.s
12 16 20 24
Time (hours]
Fig. 7. Time course of 8-bromo-cAMP action in the presence and
absence of dexamethasone. HTC cells were incubated for 0-24 hours
in serum-free medium containingO.1 I'M dexameth asone (A), 3 mM
8-bromo-cAMP (0), 3 mM 8-bromo-cAMP/ I mM MIBX (0), or 3 mM
8-bromo-cAMP/ I mM MIBX/ 0.1 I'M dexamethasone ("). (e), Control
(no additions). Upper panel: cell-associated PA activity. Lower
panel: extracellular PA activity. Inset: expanded ordinate scale to
demonstrate the effect of dexamethasone. In these experiments, the
fibrinolytic activities of several different amounts of each sample
were mea sured and the slope of the line (percentage of fibrin
solubilized per microgram cellular protein or microliter of
extracellular medium) was determined by linear regression analysis.
P A activ ity is expressed as micrograms fibrin solubilized per
microgram of cellular protein or microliter of extracellular
medium. Repro
duced from reference 41.
detected at4 hours in cells incubated with 8-bromo cAMP and MIBX
but not until 12 hours in cells incubated with dexamethasone as
well (41). Induc tion of inhibitor by dexamethasone might explain
this delay in appearance of the cyclic nucleotide stimulated
increase in PA activity. Glucocorticoids thus exert two separate
and opposite effects on P A activity: induction of an inhibitor and
amplification of cyclic nucleotide action. Although
permissive
and synergistic effects of dexamethasone on cyclic nucleotide
action have been reported previously (38,63), glucocorticoid
regulation of P A activity is unique in that the amplification of
cyclic nucleotide effects by dexamethasone opposes its regulatory
action toward a specific enzyme.
In variant HTC cells selectively resistant to the glucocorticoid
inhibition of P A activity, dexameth asone still enhances the
stimulation of P A activity by cyclic nucleotides (Fig. 8).
Furthermore, in con trast to its effect in wild-type cells,
dexamethasone does not alter the time course of 8-bromo-cAMP
stimulation of P A activity in variant cells; en hancement is
first observed at 6 hours and is max imal by 12 hours incubation.
These results appear to dissociate the glucocorticoid induction of
inhibi tor from its enhancement of cyclic nucleotide stim ulation
of P A activity.
The steroid specificity of the glucocorticoid en hancement effect
appears to be similar to that for glucocorticoid inhibition of P A
activity and amino
!: :~ 20 ti co 0 'ij 10 ~
'iii B '0 0 4 8 12 16
*' Variant
~ u::
o Time, hr
Fig. 8. Time course of8-bromo-cAMP stimulation of PA activi ty in
the presence or absence of dexamethasone in wild-type and variant
cells. Wild-type and variant HTC cells were incubated for 0-24
hours in serum-free medium containing: no additions (e), 0.1 I'M
dexamethasone (A), 3 mM 8-bromo-cAMP/ I mM MIBX (0), orO.II'M
dexamethasonej3 mM 8-bromo-cAMP/ I mM MIBX ("). Cell-associated P A
activity (in 2.5 micrograms cellular protein of wild-type cells or
I microgram cellular protein of variant cells) was measured on
"'I-fibrin plates as described in Materials and methods, and
expressed as a percentage of total radioactivity released.
acid transport (64), as well as induction of tyrosine
aminotransferase (65). The optimal inducers, or full agonists,
dexamethasone and cortisol, show a sim ilar concentration
dependence curve for all of these phenomena. In each case,
dexamethasone was ten times more potent than cortisol. The partial
ago nists, II f3-hydroxyprogesterone and deoxycortico sterone,
cause submaximal enhancement of the cyc lic nucleotide stimulation
of P A activity and do so only at higher steroid concentrations.
Tetrahydro cortisol, which does not interact with the glucocor
ticoid receptor, fails to enhance cyclic nucleotide stimulation of
P A activity. The glucocorticoid an tagonist,
17a-methyltestosterone, which has no ef fect on the enhancement of
cyclic nucleotide stimu lation by itself, blocks the enhancement
by dexa methasone in a concentration-dependent manner. These
observations suggest that the steroidal en hancement of cyclic
nucleotide stimulation of P A activity is mediated by the same
glucocorticoid re ceptor mechanism which mediates the induction of
transaminase, and the inhibition of amino acid transport and PA
activity (7, Barouski-Miller & Gelehrter, unpublished
work).
Incubation of HTC cells with cAMP derivatives also alters cell
morphology, causing cell elongation and extension of processes
followed by flattening of the cells. Plasminogen activator has been
reported to alter cell morphology in several lines either di
rectly (66) or by production of plasmin (67). We have shown,
however, that the cyclic nucleotide effects on cell morphology are
not caused by the stimulation of P A activity and can be
dissociated from them. The morphologic changes appear with in 30
to 60 min incubation with cyclic AMP derivatives, long before any
detectable changes in either intracellular or extracellular P A are
appar ent. Upon removal of cyclic nucleotides from the medium,
cell morphology returns to normal within two to four hours, a time
at which P A activity is still significantly elevated. Furthermore,
when protein synthesis is blocked by cycloheximide, the cyclic
nucleotide stimulation of P A activity is completely blocked;
however, induction of mor phologic changes still occurs. Analogous
to the si tuation described above for glucocorticoid regula tion
of P A and cell adhesion, these results suggest independent
regulation by cyclic nucleotides of P A activity and cell
morphology (7, Barouski-Miller & Gelehrter, unpublished
work).
19
Discussion
HTC cells provide a useful experimental model for studying the
hormonal regulation of plasmino gen activator and the role of PAin
several cellular functions modulated by hormones. In addition to
various well-characterized hormonal responses in these cells,
variant cell lines which are resistant to specific hormone-mediated
functions have been is olated (9,39,40,55,68). Our studies have
revealed two unique regulatory mechanisms: glucocorticoids inhibit
P A activity by inducing a soluble inhibitor rather than by
regulating the amount of enzyme (10, II, 28); and glucocorticoids
together with cy clic nucleotides paradoxically enhance the
dramat ic stimulation of P A activity by cyclic nucleotides
(41).
Further investigation of this system should yield interesting
information about mechanisms of hor monal regulation of this
important protease. The ability to study the multiple molecular
weight forms ofPA should allow studies on the hormonal regula tion
of the expression of these forms. We can inves tigate whether
various hormones cause differential regulation of these forms of
plasminogen activator and whether they affect interconversion of
these forms. The isolation and characterization of the
dexamethasone-induced inhibitor (PAl) and the preparation of
specific antibodies to it should allow studies on the hormonal
regulation of PAl at a molecular level. Finally, the paradoxical
effects of glucocorticoids on PAin this system provide a unique
opportunity to study the nature of glucocor ticoid-cyclic
nucleotide interactions.
Acknowledgements
This work was supported by Grant CA 22729 from the National Cancer
Institute. P.A.B-M. was supported by Predoctoral Training Grant G M
97544 from the National Institutes of Health. We thank Ms Judy
Worley for secretarial assistance, and Denis Lee for helping create
Fig. I.
References
I. Christman, J. K., Silverstein, S. C. and Acs, G .. 1977. In:
Proteinases in Mammalian Cells and Tissues. (Barrett, A. J., ed.),
pp. 91-149, New York: North Holland Publishing Co.
20
2. Lijnen, H. R. and Collen, D., 1982. Seminars in Thrombosis and
Hemostasis 8: 2·10.
3. Reich, E., 1978. In: Biological Markers of Neoplasia: Basic and
Applied Aspects. (Ruddon, R. W., Jr., ed.), pp. 491-500, New York:
Elsevier-North Holland.
4. Weinstein, I. B., Wigler, M., Yamasaki, H. et aI., 1978. In:
Biologlcal Markers of Neoplasia: Basic and Applied As pects.
(Ruddon, R. W., Jr., ed.), pp. 451-471, New York: Elsevier-N orth
Holland.
5. Werb, Z., Mainardi, C. L."Vater, C. A. and Harris, E. D., Jr.,
1977. N. Engl. J. Med. 296: 1017-1023.
6. Quigley, J. P., 1979. In: Surfaces of Normal and Malignant
Cells. (Hynes, R.O., ed.), pp. 247-285, Chichester: John
Wiley.
7. Miller, P. A. Barouski-, 1982. Ph.D. thesis, University of
Michigan.
8. Wigler, M., Ford, J. P. and Weinstein, I. B., 1975. In: Pro
teases and Biological Control. (Reich, E., Rifkin, D. B. and Shaw,
E., eds.), pp. 849-856, New York: Cold Spring Harbor
Laboratory.
9. Carlson, S. A. and Gelehrter, T. D., 1977. J. Supramolecular
Structure 6: 325-331.
10. Seifert, S. C. and Gelehrter, T. D., 1978. Proc. Natl. Acad.
Sci. U.S.A. 75: 6130-6133.
II. Gelehrter, T. D., Seifert, S. C. and Fredin, B. L., 1979. Cold
Spring Harbor Conf. Cell Prolif. 6: 259-267.
12. Laishes, B. A., Roberts, E. and Burrowes, c., 1976. Bio chern.
Biophys. Res. Commun. 72: 462-471.
13. Vassali, 1.-D., Hamilton, J. and Reich, E., 1976. Cell 8:
271-281.
14. Granelli-Piperno, A., Vassalli, J.-D. and Reich, E., 1977. J.
Exp. Med. 146: 1693-1706.
15. Roblin, R. and Young, P. L., 1980, Cancer Research 40:
2706-2713.
16. Werb, Z., 1978. J, Exp. Med. 147: 1695-1712. 17. Beers, W. H.,
Strickland, S. and Reich, E., 1975. Cell 6:
387-394. 18. LaCroix, M. and Fritz, I. B., 1982. Molec. and Cell.
Endo
crinol. 26: 247-258. 19. Laug, W. E., Jones, P. A., Nye, C. A. and
Benedict, W. F.,
1976. Biochem. Biophys. Res. Commun. 68: 114-119. 20. Rosen, N.,
Piscitello, J., Schneck, J. et aI., 1979. J. Cell.
Physiol. 98: 125-136. 21. Rosen, N., Schneck, J., Bloom, B. R. and
Rosen, O. M.,
1978. J. Cyclic Nucleotide Research 5: 345-358. 22. Wilson, E. L.
and Reich, E., 1979. Cancer Research 39:
1579-1586. 23. Strickland, S. and Mahdavi, V., 1978. Cell 15:
393-403. 24. Schroder, E. W., Chou, I.-N. and Black, P. H., 1980.
Cancer
Research 40: 3089-3094. 25. Miskin, R., Easton, T. G. and Reich,
E., 1978. Cell 15:
1301-1312. 26. Mochan, E., 1979. Biochim. Biophys. Acta 558:
273-278. 27. Wigler, M. and Weinstein, I. B., 1976. Nature 259:
232-233. 28. Coleman, P. L., Barouski, P. A. and Gelehrter, T. D.,
1982.
J. BioI. Chern. 257: 4260-4267. 29. Loskutoff, D. J. and Edgington,
T. S., 1977. Proc. Natl.
Acad. Sci. U.S.A. 74: 3903-3907. 30. Roblin, R. 0., Young, P. L.
and Bell, T. E., 1978. Biochem.
Biophys. Res. Commun. 82: 165-172.
31. Rohrlich, S. T. and Rifkin, D. B., 1981. J. Cell. Physiol. 109:
1-15.
32. Low, D. A., Baker, J. B., Koonce, W. C. and Cunningham, D. D.,
1981. Proc. Natl. Acad. Sci. U.S.A. 78: 2340-2344.
33. Finlay, T. H., Katz,J., Rasums, A. etal., 1981. Endocrinol ogy
108: 2129-2136.
34. Thompson, E. B., 1979. In: Glucocorticoid Hormone Ac tion.
(Baxter, J. D. and Rousseau, G. G., eds.), pp. 203-217, Heidelberg:
S pringer-Verlag.
35. Higgins, S. J., Baxter, J. D. and Rousseau, G. G., 1979. In:
Glucocorticoid Hormone Action. (Baxter, J. D. and Rous seau, G.
G., eds.), pp. 135-160, Heidelberg: Springer-Verlag.
36. Gelehrter, T. D., 1979. In: Glucocorticoid Hormone Action.
(Baxter, J. D. and Rousseau, G. G., eds.), pp. 561-574, Hei
delberg: S pringer-Verlag.
37. Gelehrter, T. D., 1979. In: Glucocorticoid Hormone Action.
(Baxter, J. D. and Rousseau, G. G., eds.), pp. 583-591, Hei
delberg: S pringer-Verlag.
38. Granner, D. K., 1979. In: Glucocorticoid Hormone Action.
(Baxter, J. D. and Rousseau, G. G., eds.), pp. 593-611, Hei
delberg: S pringer-Verlag.
39. Seifert, S. C. and Gelehrter, T. D., 1979. J. Cell. Physiol.
99: 333-342.
40. Thompson, E. B., Aviv, D. and Lippman, M. E., 1977. En
docrinology 100: 406-419.
41. Barouski-Miller, P. A. and Gelehrter, T. D., 1982. Proc. Natl.
Acad. Sci. U.S.A. 79: 2319-2322.
42. Strickland, S. and Beers, W. H., 1976. J. BioI. Chern. 251:
5694-5702.
43. Coleman, P. L. and Green, G. D. J., 1981. Annals, N.Y. Acad.
Sci. 370: 617-626.
44. Coleman, P. L. and Green, G. D. J., 1981. Meth. Enzymol. 80:
408-414.
45. Granelli-Piperno, A. and Reich, E., 1978. J. Exp. Med. 148:
223-234.
46. Huessen, C. and Dowdle, E. B., 1980. Anal. Biochem. 102:
196202.
47. Miskin, R. and Soreq, H., 1982. Anal. Biochem. 118:
252-258.
48. McDonald, R. A. and Gelehrter, T. D., 1981. J. Cell BioI. 88:
536-542.
49. Jones, P., Benedict, W., Strickland, S. et aI., 1975. Cell 5:
323-329.
50. Deutsch, D. G. and Mertz, E. T., 1970. Science 170:
1095-1096.
51. Gilbert, L. R. and Wachsman, J. T., 1976. Anal. Biochem. 72:
480-484.
52. Barouski, P. A. and Gelehrter, T. D., 1980. Biochem. Bio phys.
Res. Commun. 96: 1540-1546.
53. Van Leuven, F., Cassiman, J-J. and Van den Berghe, H., 1979. J.
BioI. Chern. 254: 5155-5160.
54. Thompson, E. B., Anderson, C. U. and Lippman, M. E., 1975. J.
Cell. Physiol. 86: 403-412.
55. Thompson, E. B., Granner, D. K., Gelehrter, T. D. et aI., 1979.
Molec. and Cell. Endocrinol. 15: 135--150.
56. Pfahl, M. R., Kelleher, R. J. and Bourgeois, S., 1978. Molec.
and Cell. Endocrinol. 10: 193-207.
57. Yamamoto, K. R., Gehring, U., Stampfer, M. R. and Sibley, C.
H., 1976. Rec. Prog. Hormone Res. 32: 3-23.
58. Ballard, P. and Tomkins, G. M., 1970. J. Cell BioI. 47:
222-234.
59. Culp, L., 1978. Curf. Top. Membf. Transp. 11: 327-395. 60.
Pollack, R. and Rifkin, D., 1975. Cell 6: 495-506. 61. Weber, 1.
M., 1975. Cell 5: 253-261. 62. Fredin, B. L., Seifert, S. C. and
Gelehrter, T. D., 1979. Na
ture 277: 312-313. 63. Rousseau, G. G., 1977. Eur. J. Biochem. 76:
309- 316. 64. Gelehrter, T. D. and McDonald, R. A., 1981.
Endocrinology
109: 476-482.
21
65. Samuels, H. H. and Tomkins, G. M., 1970.1. Mol. BioI. 52:
57-74.
66. Quigley, J. P., 1979. Cell 17: 131-141. 67. Urquhart, c., Whur,
P., Gordon, M. et aI., 1978. Exptl. Cell
Res. 113: 31-38. 68. Grove, 1. R. and Ringold, G. M., 1981. Proc.
Natl. Acad.
Sci. U.S.A. 78: 4349-4353.
Received 24 September 1982.
Molecular and Cellular Biochemistry 53/54, 23-33 (1983), © 1983,
Martinus Nijhoff Publishers, Boston. Printed in The
Netherlands.
Cell types in rat liver cultures: their identification and
isolation
J. W. Grisham Department of Pathology. University of North
Carolina. School of Medicine, Chapel Hill, NC 27514. U.S.A.
Abstract
This paper reviews the various types of cells in the liver in vivo
and in hepatic cellular suspensions produced by perfusion ofthe
liver with collagenase solutions. Methods to identify and isolate
different types of hepatic cells are discussed. In vitro culture of
various types ofliver cells is reviewed and the identification of
cultured cells is considered.
I. Introduction
The major aim of this paper is to consider in broad outline methods
and strategies for isolating and culturing the several types of
cells that compose the liver. A secondary aim is to discuss the
problems and possibilities of identifying specific types of cells
in cultures of liver tissue in order to relate the cultured cells
to cells in the liver in vivo.
Culture of liver tissue and of cells isolated from liver affords
the investigator the opportunity to examine hepatic function at the
cellular level, un impeded by those constraints that follow from
the fact that the liver in vivo is a complex tissue com posed of
several types of cells and unaffected by the influences of cells
and cellular products from other parts of the body. Although much
progress has been made in the culture of liver and in the disper
sion of liver tissue into popUlations of single cells for culture,
much less effort has been directed to ward the separation of the
mixed population that comprises a suspension of total liver cells
into sub popUlations containing only one type of cell. Many
investigators seem to assume that an enzymatic suspension ofliver
parenchyma contains a virtually pure population of hepatocytes,
although such an assumption can readily be shown to be incorrect
by
using relatively simple cytological or tissue culture techniques.
An unquestioning attitude about the cellular composition of an
enzymatic suspension of liver cells has already led to possibly
incorrect con clusions concerning the primacy of the hepatocyte in
diverse areas of hepatic function.
A related pro blem pertains to the identity of cells in long-term
cell cultures derived from dispersed liver cells. When a suspension
of liver cells is main tained in culture for several weeks,
hepatocytes are replaced by small cells with clear cytoplasm, some
of them epithelial in character, which can often be subcultured.
Such long-term epithelial cultures de rived from liver have been
used for a variety of experimental investigations. Although they
bear no morphological resemblance to authentic hepato cytes, these
continually culturable hepatic epithelial cells are frequently
termed hepatocytes. Since there is strong evidence that cultured
hepatic epithelial cells are derived from cells in the original
suspen sion other than mature hepatocytes, the designa tion of
these cells as hepatocytes is misleading. If data from the analysis
of hepatic cells in long-term culture are to be applied to an
interpretation of hepatic function in vivo, then identification of
the cell of origin in the liver is of considerable impor
tance.
24
A variety of techniques for culturing liver are now available,
including organ and explant cul tures of intact liver, primary
culture of total hepatic cellular suspensions or of partially
purified subfrac tions of the whole suspension, and secondary
(prop agable) cell cultures. Many valuable studies have already
been carried out using these techniques. Recent reviews show that
primary cultures of dispersed hepatic cells are suitable
preparations to analyze many facets of hepatic function (1,2), as
well as to evaluate the metabolism of chemicals and drugs and to
ascertain the cytotoxicity of their me tabolites (3-5). The
phenomenon of carcinogenesis in epithelial cells (as contrasted to
mesenchymal cells) was first examined in propagable hepatic
epithelial cells (reviewed in 3). Perhaps the greatest handicap to
the accurate interpretation of the re sults of studies employing
cultured liver cells stems from the combination of not knowing the
exact composition ofthe mixed cellular population one is studying
and/ or not being able to identify the cell one is studying in
vitro with a cell type present in the liver in vivo. This paper
focuses on these problems and, although it presents few answers,
aims to stim ulate studies that will yield solutions.
II. Types of cells in the liver
The liver is a cytologically and structurally com plex tissue (6,
7). In the normal liver, hepatocytes predominate in both number and
volume, but the liver contains several additional types of cells
and these non parenchymal cells are responsible for many important
hepatic functions. As a point of departure for discussing the
identification and iso lation of different types of hepatic cells,
it is useful to consider liver tissue in terms of its major
structu ral compartments-portal tracts and lobular paren chyma.
Portal tracts comprise a dense collagenous matrix containing
afferent blood vessels (both arte rial, portal venous, and
capillary), secondary and larger order bile ducts, nerves, and
lymphatic ves sels (6, 7). The spaces in the collagenous matrix
also contain a variable population of cells, including fibroblasts,
hematopoietic stem cells, and variable numbers of leukocytes,
including neutrophiles, lymphocytes, plasma cells, mast cells,
macrophages, and eosinophiles. Included in the tissue of the por
tal tracts are epithelial cells of bile ducts, endothe-
lial cells of blood and lymphatic vessels, smooth muscle cells of
arteries and veins, cells that form nerves, and a variety of
mesenchymal cells, includ ing fibroblasts and inflammatory cells.
The lobular parenchyma is histologically less complex than por tal
tracts, consisting predominantly of interposed plates of
hepatocytes and sinusoids, and of the ef ferent blood vessels into
which the sinusoids drain (6,7). The lobular parenchyma is
supported by only a light investment of collagen fibers, which form
a network around efferent vessels and along the inter face between
sinusoids and plates of hepatocytes. As in the whole liver,
hepatocytes are the most numerous cellular component of the lobular
paren chymal compartment, and they comprise the major volume
fraction of this compartment (8, 9). How ever, sinusoidal
endothelial cells and Kupffer cells, although small in size, are
also numerous constitu ents of the lobular parenchymal compartment
(9, 10); fat storing (Ito) cells (II) and so-called pit cells (12)
are also represented in the lobular parenchymal compartment of the
normal liver, residing in the tissue space between hepatocytic
plates and sinus oids. Epithelial cells of the terminal bile ducts
(usu ally called bile ductules) and fibroblasts and smooth muscle
cells in the walls of efferent hepatic veins are frequently
overlooked cellular constituents of the lobular parenchyma. Bile
ductules connect the bile canaliculi of hepatocytic plates to bile
ducts in por tal tracts, providing a pathway for the outflow of
bile (13, 14). Terminal ends of bile ductules are connected by
attachment complexes to hepatocytes at the portal ends of
hepatocytic plates; bile duc tules protrude for short distances
into the lobular parenchymal compartment without a heavy in
vestment of connective tissue (13, 14).
Not only is there heterogeneity of cell types in the lobular
parenchyma, but hepatocytes located in dif ferent parts of the
hepatocellular plates differ in form and function (15, 16). In
general, structure and function of hepatocytes follow gradients
along the distance of hepatocellular plates from the vicini ty of
portal tracts (Zone I) to the vicinity of termi nal hepatic veins
(Zone 3). In adult rats, hepato cytes of Zone I are diploid,
whereas cells of Zone 3 are polyploid. Zone I hepatocytes are
smaller and contain fewer but larger mitochondria than do hep
atocytes of Zone 3. A larger fraction of the typical hepatocyte in
Zone I is occupied by the Golgi appa ratus than is true of the
Zone 3 hepatocyte, whereas
the opposite situation obtains in the instance of the smooth
endoplasmic reticulum. Functionally, hep atocytes of Zone I appear
to be predominantly involved in gluconeogenesis, while cells in
Zone 3 are mainly involved in glycolysis. Analogous varia tions in
protein and lipid metabolism, drug metabo lism, and bile formation
may also typify hepato cytes located in different parts of the
lobular parenchymal compartment.
The cellular composition of the liver described here typifies the
normal liver of the adult rat. N or mal livers of some other
species and pathologically altered livers of all species may show a
heavier investment of collagenous connective tissue in the lobular
parenchyma and different distributions and arrangements of various
types of cells. In addition, cells in the normal livers of the
developing animal may vary considerably from those described, espe
cially by containing a large population of hemato poietic cells.
Techniques that allow the separation of cells from normal livers
may not apply to devel oping or pathologically altered
livers.
III. Features of different hepatic cells that allow selective
recovery
The remainder of this discussion focuses on the cells of the
lobular parenchymal compartment, since it is these cells that are
predominantly released by limited perfusion of the liver with
collagenase solutions, the method used most frequently to dis
perse liver cells. Each of the types of cells that make up the
tissue of the lobular parenchymal compart ment are morphologically
and functionally distinc tive and can be accurately identified in
vivo by their characteristic morphologies, synthetic capacities,
enzymatic features, metabolic responses to specific challenge,
capacity to store metabolites, locations of receptors and
receptor-mediated internalization ofligands, and ability to
phagocytically ingest large particulates. This section emphasizes
some features of various hepatic cells that facilitate their
selective recovery from an enzymatic suspension of cells of the
whole liver and their identification in long-term cultures. The
differential features of the various types of hepatic cells that
are presently useful in selectively recovering them from a mixed
popula tion are size, density, surface membrane receptors and
enzymes, and differences in the content of sub-
25
stances that can be tagged with fluorochromes. For purposes of
identification of different types of cells, these features plus the
histochemical demonstra tion of selected enzymes are useful.
Hepatocytes in the adult rat are large cells that vary from about
15 /-Lm to nearly 25 /-Lm in diameter and have densities that vary
from about 1.100 to 1.140 gm/ ml (17). In livers of immature rats
and in livers that are the sites of pathological oval cell
reactions, smaller hepatocytes are also found (18). Isolated
hepatocytes are spherical and their sur faces are completely
covered by densely numerous microvilli (17, 19). Ultrastructurally,
hepatocytes contain a rich endowment of organelles, including large
aggregates of rough and smooth endoplasmic reticulum, prominent
Golgi apparatuses, many large mitochondria, numerous Iysosomes, and
per oxisomes (17, 19). Hepatocytes from well fed ani mals contain
large stores of glycogen. Isolated hep atocytes possess the
functional ability to synthesize several serum proteins, including
albumin, trans ferrin, and coagulation factors (1,2). They
metabol ize bilirubin and possess a broad capability to me
tabolize exogenous chemicals by mixed-function mono-oxidation
(3-5). Numerous membrane re ceptors have been identified on
hepatocytes (20). Hepatocyte-specific monoclonal antibodies have
been developed (21).
Isolated Kupffer cells are 8-12 /-Lm in diameter (22) and stellate
in shape with a kidney-shaped nucleus (23). Their peak density has
been measured at about 1.076 gm/ ml (22). The surface membranes of
isolated Kuppfer cells form complex folds and microvilli (23) and
internally these cells are distin guished by a large number of
phagosomes (22). They are actively phagocytic in vitro, avidly ac
cumulating relatively large (0.8 /-Lm) particles (22-24). The
surface membranes of Kupffer cells contain Fe and C3 receptors
(23,24), receptors for mannose and N-acetyl glucosamine (25-26),
and they bind avidly to glass (23, 24), as do other mac rophages.
They contain high levels of peroxidase (22, 23), acid phosphatase
(and other lysosomal enzymes) (22, 23), and glucose-6-phosphate
dehy drogenase (27).
Isolated endothelial cells are smaller than Kupffer cells,
measuring 4-6 /-Lm in diameter (28). Their surfaces are more or
less smooth, and when isolated they are round and have a high
nuclear cytoplasmic ratio (22, 26). Ultrastructurally, the
cytoplasm of
26
freshly isolated cells has a lace-like appearance, which is
interpreted to represent the remnants of in-situ sieve plates (22).
Their density is about 1.060 to 1.080 gm/ ml (22). Endothelial
cells are said to synthesize type IV collagen (29). Whether or not
sinusoidal endothelial cells can synthesize and store coagulation
factor VIII, as do endothelial cells of other vessels, is now a
matter of dispute. Factor VIII has been localized immunologically
by fluo rescence microscopy to sinusoidal endothelial cells in
sections of liver tissue (29), but recent studies on isolated
endothelial cells are said to demonstrate that these cells lack
Factor VIII (28). Sinusoidal endothelial cells contain receptors
for glycopro teins that have a terminal mannose or N-acetyl
glucosamine residue (26).
Bile ductular epithelial cells are 6-12 Mm in di ameter( 18) and
their surface membrane is occupied by microvilli (13, 14, 31).
These cells are round and they contain a round nucleus (18). Their
cytoplas mic organelles are not distinctive (18). The density of
bile ductular cells ranges from 1.075 to 1.1 00 gm/ ml with a peak
of about 1.095 gm/ ml (18,32). Antibodies have been developed
against isolated biliary epithelial cells and shown to bind to bile
ductules in vivo (32). Bile ductular epithelial cells contain large
amounts of ),-glutamyl transpepti dase (18,32) and leucine
aminopeptidase (32).
F at-storing cells can be detected in isolated popu lations of
normal hepatic cells by the presence of lipid droplets (11) or
vitamin A fluorescence (33). Presence oflipid should make them much
less dense than cells that do not contain lipid, but the density of
these cells apparently has not been measured. Although these cells
show bright vitamin A fluores cence in vivo, attempts to
distinguish a subpopula tion of isolated nonparenchymal cells that
contains a high content of vitamin A (34) or lipid droplets (35)
have not succeeded.
A potentially useful feature for identifying hepat ic cells and,
perhaps, for rapidly separating them as multicellular clumps is
their tendency to form ag gregates of similar cells when
maintained in suspen sion cultures. Many investigators have
observed the reaggregation of dispersed hepatocytes in culture with
the formation of intercellular attachment complexes (17, 19). Other
studics have demonstrat ed that other hepatic cells, including
biliary ductu lar cells, endothelial cells, and Kupffer cells,
reag gregate in suspension culture (36).
IV. Dispersion of liver cells
The development of methods to enzymatically separate liver tissue
into suspensions of viable sin gle cells has been a signal
technological accomp lishment (37,38). U sing now standard
techniques of liver perfusion with solutions of collagenase, it is
possible routinely to separate the lobular paren chyma into
suspensions of single cells that are im permeable to trypan blue
and that retain for at least short periods a high level
offunctional integrity (39, 40). It is possible to harvest
selectively the cells of the lobular parenchyma by controlled
collagenase perfusion. The structures of the larger portal tracts
and of the larger hepatic veins are only poorly released from their
dense stromal matrix by the usual techniques of collagenase
perfusion, and they are left behind in the undigested residue. In
con trast, cells of the lobular parenchyma are more readily
dispersed since they lack a dense investment of collagenous
connective tissue. Recent studies from several laboratories have
shown that it is pos sible to harvest the cells ofthe lobular
parenchymal compartment virtually quantitatively (28). By ap
plying selective proteolysis with pronase or trypsin, hepatocytes
can be removed without rendering the nonparenchymal cells permeable
to trypan blue (22, 23, 31). However, strong proteases must be used
cautiously because they may damage surface membrane receptors,
impeding further separation of cells or analysis of function.
The nonparenchymal cell popUlation remaining after removal of all
hepatocytes from of a whole liver cell suspension contains K upffer
cells, en dothelial cells, bile ductular epithelial cells, Ito
cells, and various cells from blood (31, 35). By detecting cells
that are histochemically positive for peroxidase or that can ingest
0.8 Mm latex spheres, about 15 to 20% of the nonparenchymal cell
popu lation can be shown to represent Kupffer cells (22, 23). The
remaining cells are frequently termed en dothelial cells (22), but
the subpopulation is heterog enous. Histochemical detection of
)'-glutamyl transpeptidase-positive cells shows that 5-15% of the
nonparenchymal popUlation, excluding K upffer cells, represents
biliary epithelial cells (Grisham, unpublished work).
V. Isolation of specific types of celIs from suspen sions
Several techniques are now available by which to separate the
suspension of single cells of the hepatic parenchymal compartment
into more or less hom ogenous populations. Separation techniques
have generally been directed toward the separation of hepatocytes
from non parenchymal cells, the sub fractionation of
nonparenchymal cells into the var ious types of cells, and the
subfractionation of hep atocytes into different classes. It is
obvious that any separation technique must be nontoxic if cells are
to be used for short-term functional studies or if they are to be
propagated in culture. Tissue culture also requires that sterility
be maintained during the sep aration procedures.
The separation methods most widely used today employ sedimentation
of cells in gravitational fields. Separation of cells by
sedimentation techniques is determined by differences in the size
and/ or density of cells in the mixed cellular suspension.
Isopycnic sedimentation utilizes density gradients and rela tively
high centrifugal forces to band cells at a level of equal density
in the gradient. Attempts have been made to separate hepatocytes
and nonparenchymal cells by isopycnic sedimentation using gradients
of Ficoll (41-43), Percoll (44), or Metrizamide (45, 46). In a few
investigations, the objective has been to subfractionate
hepatocytes (47) or nonparen chymal cells (22, 32). Velocity
sedimentation de pends mainly on differences in cell size for
separa tion and utilizes low centrifugal forces, in most instances
for short periods of time. Commonly em ployed methods of velocity
sedimentation are unit gravity sedimentation, isokinetic
sedimentation, and elutriation. U nit gravity sedimentation has
been widely used as a means to partially separate hepatocytes from
nonparenchymal cells (48-50). Isokinetic sedimentation in albumin
(5 I) or Ficoll (47,52) has been used to separate populations of
hepatocytes with varying purity. Similar methods have been used to
partially subfractionate nonpar enchymal cells (32). Iso kinetic
sedimentation on Ficoll in a reorienting zonal rotor has been
applied to the separation of altered hepatocytes and non
parenchymal cells from livers of rats treated with the
hepatocarcinogen diethylnitrosamine (53). A combination of velocity
and isopycnic sedimenta tion methods, employing sedimentation at
unit
27
gravity to remove most of the hepatocytes and sed imentation on a
density gradient ofMetrizamide to separate non parenchymal cells,
was used to isolate a subfraction of nonparenchymal cells, termed
oval cells, from livers of rats that had been treated with the
hepatocarcinogen ethionine(l8). Reverse phase elutriation has been
used to effectively separate hepatocytes and nonparenchymal cells
(54, 55), to subfractionate nonparenchymal cells (56), and to
subfractionate hepatocytes (57). Separation of he patic cell
suspensions has also been attempted by sedimentation on
discontinuous gradients of Ficoll (32, 58) or by partitioning on a
phasic system of dextran and polyethylene glycol (59), but these
methods generally have not yielded satisfactory separation of mixed
hepatic cells. Interesting at tempts to make different categories
of hepatic cells more separable by stimulating the accumulation of
selected cytoplasmic organelles or storage products have been
reported. Selective hypertrophy of hep atocyte smooth endoplasmic
reticulum and stimu lation of glycogen storage has been used to
modify the size and density range of hepatocytes and there by
influence their separability ( 17,60). Kupffer cells have been
induced to phagocytize various materials with a similar objective
in mind (22). Attempts to selectively remove K upffer cells that
have ingested iron by passing hepatic suspensions through a
magnetic field have shown this imaginative method to be inefficient
(61).
In general, velocity sedimentation methods ap pear to more
effectively separate hepatic cells than do isopycnic sedimentation
methods. Separation of the much larger hepatocytes from small
nonparen chymal cells is fairly efficient with velocity sedimen
tation methods. However, clean separation of non parenchymal cells
of nearly the same size and density is more difficult. Even though
high degrees of purity of separation have been claimed for non
parenchymal cells in some instances (22), the meth ods employed to
distinguish different types of cells were inadequate to detect many
types of cells in the nonparenchymal population. Most studies have
clearly distinguished only the Kupffer cell subpopu lation and
then assumed that the remainder of the nonparenchymal population
was composed of en dothelial cells. 0 bviously other types of
cells are known to be present and some of them can be identified by
applying appropriate methods. Based on ploidy distribution,
elutriation seems clearly to
28
be superior to isopycnic sedimentation on Ficoll gradients to
subfractionate hepatocytes (17). Unit gravity sedimentation enjoys
the advantage of sim plicity, since it does not require a
centrifuge, and allows partial separation of the large hepatocytes
from the small nonparenchymal cells. Since it af fords a cheap and
easy way to enrich the fraction of hepatocytes in the hepatic cell
suspension that is cultured, a unit gravity sedimentation step is
incor porated in many protocols for establishing hepato cyte
primary cultures, but complete separation is not achieved. Another
disadvantage of unit gravity sedimentation is its slowness. Both
velocity and isopycnic sedimentation methods require expensive
centrifuges and rotors, but these techniques are rapid and can
quickly separate large numbers of cells. Separation of the large,
dense hepatocytes from small, usually less dense nonparenchymal
cells by these methods appears to be quite good if not absolute.
However, it seems unlikely that sin gle-step sedimentation
techniques will ever be able to yield absolutely precise
separations of different types of cells, especially cells in the
nonparenchy mal population, which differ little in size and densi
ty. Complete separation of such cells that are close ly similar in
size and density will require methods that take advantage of
distinct, nonoverlapping dif ferences in some property such as a
surface enzyme or other antigen. The latter methods may usefully
start with populations of cells that are partially separated by
sedimentation methods.
More specific methods to purify various subpop ulations of
hepatocytes and nonparenchymal cells are under investigation in
many laboratories. These methods attempt to take advantage of
differences in the surface properties of cells, in the content of
proteins (including enzymes) and other antigens, and in the content
of other macromolecules to yield more precise separation of cells.
A particularly promising general method is the use of antibodies to
selected cellular antigens to facilitate cellular separation by
secondary means. Differential at tachment and release of cells
from surfaces, either through nonspecific or poorly understood
mechan isms or through specific binding of cellular recep tors to
immobilized ligand, offer the possibility of precise separation of
some types of cells in hepatic cell suspensions. Various cells in
the nonparen chymal cell population may be enriched selectively by
nonspecific attachment to surfaces of glass or
plastic. Attachment of Kupffer cells to glass or plastic surfaces
allows these cells to be highly puri fied from other
nonparenchymal cells (23). Specific attachment of cells by
receptors to ligands immobil ized on surfaces is a potentially
powerful method that has not been widely applied to the separation
of different types of hepatic cells. Receptor-ligand combinations
that may be used are limited only by the requirement that they be
differentially localized to some ofthe various types of hepatic
cells. Recep tors for sugar residues of glycoproteins provide a
good example of the general attributes of a poten tially useful
system. Hepatocytes have receptors for oligosaccharide chains that
terminate in galactase residues, whereas sinusoidal endothelial
cells and Kupffer cells contain specific receptors for complex
sugars that terminate in N-acetyl glucosamine or mannose residues
(26). Endothelial cells are sixfold more active in internalizing
the latter residues than are Kupffer cells (26). Other types of
hepatic cells have not been discerned to bind these sugars, al
though this possibility needs more rigorous study. Hepatocytes bind
avidly to polyacrylamide sur faces containing galactose via a
Ca++-requiring reaction, and the bound cells may be released (62).
If other hepatic cells lack receptors for galactose and man nose,
it should be possible to separate hep atocytes and sinusoidal
cells (endothelial and Kupffer cells) from other hepatic cells with
consid erable precision by passing hepatic cell suspensions
sequentially over surfaces to which galactose and man nose are
bound. The number of receptors on hepatic cells is vast, and it is
possible that other receptors are located differentially on various
types of hepatic cells. For example, hepatocytes that con tain
concanavalin-A binding sites have been isolat ed by binding to the
ligand (63). Other cells of the nonparenchymal population may have
cell specific receptors. Epithelial cells of terminal biliary ducts
selectively bind certain steroids (64). For cells to be effectively
separated by receptor-ligand binding they must be dispersed by
techniques that preserve the integrity of receptor molecules.
Proteases, in cluding pronase and those proteases that contami
nate the relatively crude collagenase preparations used to disperse
hepatic tissue, may damage or destroy receptors and prevent
specific binding to ligands.
Fluorescence activated cell sorting may be used to separate cells
that can be specifically tagged with
a fluorescent material. Hepatic cells have been sort ed into
different population groups based on ploidy by using a fluorescent
compound that binds stoi chiometrically to DNA (65). This
technique separ ates diploid cells (all nonparenchymal cells and a
fraction of the hepatocytes) from polyploid cells of other classes
(several fractions of hepatocytes). Fluorescently tagged antibodies
to cellular antigens that are located on the surface membrane of
cells may also be used to separate these cells by fluores cence
activated cell sorting. Epithelial cells from biliary ductules have
been sorted after reacting them with a specific antibody to the
membrane enzyme 'Y-glutamyl transpeptidase (66). An anti body to
the hepatocyte galactose receptor protein, which has been isolated
(67), might be used to sep arate hepatocytes. Antibodies,
monoclonal and po Iyclonal, that bind specifically to unknown
antigens of hepatocytes or bile duct epithelium have been reported
(21, 32). Antibodies directed toward these and other hepatic cells
could provide the basis for precise fractionation and separation of
hepatic cells by fluorescence activated cell sorting. Fluorescence
activated cell sorting is limited by the relatively small number of
cells that can be economically separated by this method. Although
sufficient cells can be readily sorted to establish tissue
cultures, it is not economical today to use this technique to iso
late cells for biochemical studies.
Separation of cells by electrophoresis has been advocated for many
years, but apparently has not been recently used to attempt to
separate hepatic cells.
A frequently overlooked and simple method to isolate specific cells
is to plate suspensions as single cells and to selectively
subculture colonies that have the desired morphological and
functional proper ties (68). Selective culture media and
conditions (biomatrix) show promise of allowing the propaga tion
of specific types of cells, although these meth ods are still
poorly defined.
VI. Long-term culture of isolated cells
Hepatocytes can be maintained in primary cul ture for several
weeks, although their major func tions deteriorate markedly during
this period of time (69). Changes in culture media allow the long
term expression of some functional properties by
29
hepatocytes in primary cultures (70) but the hepat ocytes cannot
be subcultured. The major impedi ment to subculturing hepatocytes
has been their inability to divide in culture to form daughter
cells. Recent reports indicate that under appropriate conditions
hepatocytes may proliferate in culture (71, 72). Coupling of
culture conditions that allow hepatocytes to cycle with those that
facilitate the maintenance of critical functions may allow the
eventual development of propagable (subcultur able) populations of
hepatocytes and the estab lishment of clonally derived cultures of
hepatocytes from single cells.
Several investigators have established long-term, propagable
diploid cultures of liver cells that have so-called epithelial,
fibroblastic, or macrophagic patterns of growth in vitro (reviewed
in 3). When mass cultures are established from crude isolates of
liver cells, it is impossible to precisely identify the origin of
the cultured cells from a specific type of cell present in the
intact liver tissue. However, the use of a pure population of
isolated cells or the establishment of clonal culture populations
from single isolated cells as starting material has allowed the
clear demonstration that Kupffer cells and pre sumed bile ductular
epithelial cells can be main tained in culture in vitro for
relatively long periods with the maintenance of major functional
capabili ties. Preliminary evidence suggests that endothelial
cells may also be maintained in culture in pure populations.
However, propagable cultures of Kupffer cells or endothelial cells
that have been rigorously identified before cultures were estab
lished apparently have not been reported.
Kupfter cells have been grown in liquid medium after isolation from
a crude nonparenchymal cell population by adherence to glass or
plastic (23). After plating, Kupffer cells flatten but retain a
stel late shape (23). When Kupffer cells proliferate, they form
clusters typical of other macrophages, but confluent sheets of
cells are not formed (73). Cul tured cells are distinguished by
their phagocytic capability, by the presence of Fe receptors, and
by the high specific activity of lysosomal enzymes and peroxidase
(23, 73). Kupffer cells have been main tained in culture for up to
several weeks (73). Re cently a cell line isolated from liver
several decades ago was shown to have functional characteristics of
macrophages, presumably Kupffer cells (74).
Presumptive endothelial cells flatten and form
30
continuous sheets after attachment (Grisham, un published work).
Ultrastructurally, they are simple and they lack distinctive
morphologic features. Fen estrations have not been reported in
endothelial cells that have been maintained in culture. Endothe
lial cells fo{m type IV collagen in culture, but this does not
distinguish them precisely from bile ductu lar cells (29). A
recent study suggests that cultured endothelial cells do not
contain Factor VIII (29). Lacking a specific marker for endothelial
cells, such as Factor VIII, it is difficult at present to identify
these cells once they have been cultured.
After plating, small epithelial cells that are not hepatocytes and
that are presumed to be bile ductu lar epithelial cells flatten
and attach to form a con tinuous sheet of cells joined by
attachment com plexes (68). Because the origin of these cells from
bile ductular epithelium is unproved, these cultured cells are best
termed, noncommittally, hepatic epithelial cells. Cultured hepatic
epithelial cells contain relatively high levels of I'-glutamyl
trans peptidase (Grisham, unpublished work). They syn thesize
collagen of types I, III, and IV (75) and they ha ve levels of
receptors for epidermal growth factor (Earp and Grisham,
unpublished work) and insulin (76) that are similar to the receptor
levels found on isolated hepatocytes. Many propagable hepatic
epithelial lines and strains possess a few hepatocyte like
functions, but no single line possesses a large combination of
functions reminiscent of mature hepatocytes. Cells from some
individual lines and strains may synthesize and secrete one or more
serum proteins, including albumin, transferrin, (X
fetoprotein, and fibrinogen (reviewed in 3). Greatly increased
fractions of biliary ductular cells, having properties similar to
cells from normal livers, can be isolated from livers that are
sites of oval cell reac tions (IS).
VII. Identification of cells in long-term cultures
The principles of identification of cells in long term cultures
are similar to the principles of identi fication of cells in the
liver in vivo or in freshly isolated suspensions; one can identify
cells in long term cultures if it can be shown that they have
properties that place them in one of the major types of cells in
the liver in vivo and that exclude them from other types. Although
it is not possible at
present to uneq uivocally identify all types of cells in long-term
cultures, as we gain further detailed in formation on the
differences in characteristics of the various types of liver cells
such identification should become possible. Antibodies to specific
types of hepatic cells would provide potentially the most precise
and accurate method to identify cells in long-term culture.
At present, perhaps the most difficult hindrance to the precise and
facile identification of cells in long-term cultures is the
possible phenotypic alter ation (dedifferentiation) of cells under
conditions of culture so that they no longer possess the identi
fying characteristics ofthe cell of origin in vivo (77). Although
recent observations on the macrophagic characteristics of a long
established line of liver cells suggests that culture-associated
changes in mor phology and function may not prevent identifica
tion (74), opinions differ. This is a point of conten tion as
regards the origin of lines of continuously propagable hepatic
epithelial cells. Recognizable hepatocytes deteriorate structurally
and functional ly during the first 7 -10 days in culture, and
colonies of small, morphologically simple epithelial cells appear.
It has been suggested that these colonies of simple epithelial
cells arise from the altered hepato cytes (77). The analysis of
this problem has been approached directly in two ways: by
determining the effect of proteolytic destruction of hepatocytes by
trypsin or pronase on formation of epithelial colonies in culture
and by assessing the ability of different types of single liver
cells to proliferate to form epithelial clones in vitro. The number
of epithelial colonies formed when standard inocu lums of hepatic
cell suspension are plated does not decrease even though all
discernable hepatocytes have been destroyed (6S). Establishment of
colonies by primary cloning from single cells isolated from
collagenase-dispersed liver cell suspensions demon strated that
hepatocytes never formed clones, while clones developed from small
nonparenchymal cells (6S). Nonparenchymal clonogenic cells
were6-S!lm in longest diameter and had scant, nongranular cytoplasm
(6S). These results indicate that hepatic epithelial colonies can
originate from a cell or cells in the non parenchymal cell
popUlation and exclude the required involvement of mature or
differentiated hepatocytes. Although these studies demonstrate that
hepatocytes do not produce epithelial clones under the conditions
of these experiments, they do
not exclude the possibility that mature hepatocytes may
phenotypically modulate or retrodifferentiate under the artificial
conditions of culture and yield epithelial clones that grow in
long-term culture. Further studies are needed to corroborate or ex
clude this possibility.
VIII. Suggested nomenclature for liver cells in long term
culture
To avoid ambiguity, cells in long-term culture should not be given
the same name as a type of cell in the liver unless they can be
shown unequivocally to be derived from that £ell. Single cell
cloning appears at present to be the only certain way to prove that
a cultured cell population arises from a specific type of cell.
Even this circumstance requires that the single cell be precisely
identified if the cultured population is to be accurately traced to
a specific type of cell in vivo. Unless the cell in ques tion
possesses unique features that may be dis cerned without rendering
it unculturable, it may not be possible to provide sufficiently
strong evi dence to enable one to accurately trace a cultured cell
population to a particular type of cell in the liver. Culturing of
populations of cells, even popu lations that are greatly enriched
in a single type of cell, cannot be used as unequivocal proof of
the origin of the cultured population. No method of fractionation
of single cell suspensions of liver yet has the precision required
to exclude the presence of small numbers of different types of
cells which may grow preferentially in vitro and yield a subcul
turable population. Cell sorting on the basis of a unique cellular
characteristic (i.e., a feature not shared by other hepatic cells)
may eventually allow the establishment of pure cultures from
specific types of hepatic cells. More experience is needed.
Cultured cells whose origin is not rigorously es tablished
cIonogenically, but whose functional at tributes suggest their
origin from a particular type of cell in the liver, should be
designated as being like that particular cell, e.g.,
hepatocyte-like, Kupffer cell-like, etc. Cultured cells whose
origin clonogeni cally is not certain and whose functional
properties are unknown should be given a less committal de
signation. For example, cultured cells that possess epithelial
characteristics but lack a history or func tional qualities that
would limit their origin from
31
either hepatocytes or bile ductular cells should be designated as
hepatic epithelial cells. Epithelial cells cultured from liver are
the source of perhaps the greatest terminological ambiguity.
Although such cells are often loosely referred to as hepatocytes in
the literature, there is no firm evidence at present that they may
be derived from hepatocytes while it has been demonstrated that
they can be derived from a small, non parenchymal cell that may be
derived from the epithelium of bile ductules or some other source.
The situation is made even more ambiguous because cultured hepatic
epithelial cells that are rigorously shown not to be derived from
hepatocytes may demonstrate limited hepatocyte like functional
characteristics in vitro. Although there is suggestive evidence
that bile ductular epithelial cells and hepatocytes may undergo
phe notypic interchanges under some normal and patho logical
conditions in vivo, and in tissue culture, these potential
interconversions are as yet obscure. Until such potential cellular
interchanges are un equivocally documented and their mechanisms
bet ter discerned, it is best not to refer to any hepatic
epithelial cell in long-term culture as a hepatocyte. To do so
leads at this time to ambiguity and confu sion.
Some hepatic cells may grow in culture as coher ent sheets, but
cannot be determined unequivocally to be either epithelial or
endothelial. Such cultured cells should be given some no~specific
designation that indicates their origin from liver and, perhaps,
describes a major attribute, but is noncommittal about their
derivation from a specific type of cell in the liver. The
designation hepatic clear cells can be used for this purpose.
IX. Conclusions
The investigation of cultured hepatic cells has already led to many
new insights on the cellular functions of the liver and improved
the accuracy and precision of previous observations from in vivo
studies. In order to maximize the utility of the observations made,
investigators who study hepatic cells in vitro need to be as
precise as possible about the types of hepatic cells that are being
examined in vitro and about the need not to confuse or obscure the
functions of different types of hepatic cells by oversimplifying
their correlation in vivo and in vi-
32
tro. The liver is a cytologically complex tissue that should not be
analyzed as if the hepatocyte were the sole type of cell present,
and, indeed, hepatocytes themselves cannot be treated as a
homogenous population. Separation and culture of hepatic cells has
already facilitated the development of new knowledge on the
contributions to integrated he patic functions by different types
of hepatic cells, and additional insights on cellular functions,
can be confidently predicted.
Studies on cultured epithelial cells from liver - both he