| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Center for Engineering in Medicine and Surgical Services, Massachusetts General Hospital, Harvard Medical School, and Shriners Hospital for Children, Boston, Massachusetts 02114, USA
1Correspondence: Center for Engineering in Medicine, Massachusetts General Hospital, Bigelow 1401, 55 Fruit St., Boston, MA 02114, USA. E-mail: mtoner{at}sbi.org
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONCELL-CELL INTERACTIONS IN THE...COCULTIVATION OF HEPATOCYTES AND...EFFECT OF COCULTURE ON...MECHANISM OF INDUCTION OF...METHODS TO EXAMINE INFLUENCE...MICROFABRICATED COCULTURES:...Role of the Extent...Role of Heterotypic Interactions...Role of...CLINICAL APPLICATIONS OF...SUMMARYREFERENCES |
|---|
Key Words: liver • coculture • bioartificial liver • hepatocyte morphology
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONCELL-CELL INTERACTIONS IN THE...COCULTIVATION OF HEPATOCYTES AND...EFFECT OF COCULTURE ON...MECHANISM OF INDUCTION OF...METHODS TO EXAMINE INFLUENCE...MICROFABRICATED COCULTURES:...Role of the Extent...Role of Heterotypic Interactions...Role of...CLINICAL APPLICATIONS OF...SUMMARYREFERENCES |
|---|
, 2)
, pathophysiology (3
, 4)
,
cancer (5
, 6)
, developmental biology (7
, 8)
,
wound healing (9
, 10)
, and attempts to replace tissue
function through `tissue engineering' (11
, 12)
. Further
understanding of how cell–cell interactions modulate tissue function
will allow us to gain fundamental biological insight as well as suggest
approaches that will allow the manipulation of tissue function in
vitro for therapeutic applications.
In particular, heterotypic interactions play a fundamental role in
liver function. The formation of this vital organ from the endodermal
foregut and mesenchymal vascular structures is thought to be mediated
by heterotypic interactions (13
, 14)
. Heterotypic
interactions have also been implicated in adult liver physiology (i.e.,
localization of enzymes in zones of the liver) and pathophysiology
(i.e., cirrhosis, and response to injury) (15
16
17)
. As we
describe in this review, even in vitro, heterotypic
interactions have proved useful in stabilizing liver-specific functions
in isolated hepatocytes.
Despite extensive work in this area, the details by which cell–cell interactions modulate the hepatocyte phenotype in vitro remain unelucidated. Here, we summarize the existing works on cocultivation of hepatocytes with nonparenchymal cells: the experimental approaches, the outcome, and proposed mechanisms of interaction. In addition, recent advances in cell culture techniques (micropatterning) are discussed as they facilitate examination of these model systems. Finally, we present and discuss various approaches to the incorporation of hepatocyte cocultures into clinical liver support systems.
|
CELL–CELL INTERACTIONS IN THE LIVER IN VIVO |
|---|
|
TOPABSTRACTINTRODUCTIONCELL-CELL INTERACTIONS IN THE...COCULTIVATION OF HEPATOCYTES AND...EFFECT OF COCULTURE ON...MECHANISM OF INDUCTION OF...METHODS TO EXAMINE INFLUENCE...MICROFABRICATED COCULTURES:...Role of the Extent...Role of Heterotypic Interactions...Role of...CLINICAL APPLICATIONS OF...SUMMARYREFERENCES |
|---|
. The fibrous and hemopoietic tissue
and Kupffer cells of the liver are also derived from the mesodermal
septum transversum. It is thought that the mesenchyme induces the
endoderm to proliferate, branch, and differentiate (19)
.
It has been shown experimentally in chimeric avian and mouse livers
that differentiated hepatocytes arise from the endodermal compartment
and mesenchyme gives rise to the endothelial lining of the adult
sinusoids (13)
. In addition, when endoderm was cultivated
alone, it failed to differentiate; however, tissue interactions between
hepatic endoderm and mesenchyme induced hepatocyte differentiation
in vitro. More recently, specific cytokines and
transcription factors have been identified as important mediators of
this process (20
, 21)
.
In contrast, the adult form of the liver, a complex multicellular
structure, is seen in Fig. 1
(reprinted from Kaplowitz, ref 22
). It consists of
differentiated hepatocytes (H) separated from a fenestrated endothelium
(E) by the Space of Disse. Lipocytes (stellate or Ito cells) are
elaborate, extensive processes that encircle the sinusoid,
well-positioned for both communication with hepatocytes and the
potential to modify the extracellular space by secretion of
extracellular matrix. Biliary ductal cells contact hepatocytes toward
the end of the hepatic sinusoid (not depicted) and Kupffer cells (the
resident macrophage), and Pit cells (a type of natural killer cell) are
free to roam through the blood and tissue compartment. Thus, the adult
liver provides a scaffold for many complex cell–cell interactions that
allow for effective, coordinated organ function.
|
The information about cell–cell interactions in liver development and terminal differentiation implies an essential role for cell signaling between parenchymal and nonparenchymal tissue compartments. Cocultivation of hepatocytes with other cell types in vitro offers a unique model for in-depth study of these critical pathways.
|
COCULTIVATION OF HEPATOCYTES AND NONPARENCHYMAL CELLS |
|---|
|
TOPABSTRACTINTRODUCTIONCELL-CELL INTERACTIONS IN THE...COCULTIVATION OF HEPATOCYTES AND...EFFECT OF COCULTURE ON...MECHANISM OF INDUCTION OF...METHODS TO EXAMINE INFLUENCE...MICROFABRICATED COCULTURES:...Role of the Extent...Role of Heterotypic Interactions...Role of...CLINICAL APPLICATIONS OF...SUMMARYREFERENCES |
|---|
, 24)
, mutagenesis (25
26
27
28)
, xenobiotic
toxicity (29
, 30)
, response to oxidative stress
(31)
, lipid metabolism (32
, 33)
, and
induction of the acute-phase response (34
35
36
37)
. In
addition, the ability to preserve key features of the hepatocyte
phenotype in vitro may have important applications in
hepatocyte-based therapies for liver disease.
First noted by Langenbach et al. in 1979 (38)
through work
with hepatocytes atop irradiated feeder layers of human fibroblasts and
later elucidated by Guguen-Guillouzo et al. (39)
by a
mixed coculture of hepatocytes with live isolated rat liver epithelial
cells, the effect of cell–cell interactions on the hepatocyte
phenotype has become an active area of investigation. Figure 2
shows the earliest images, to our knowledge, of retained hepatocyte
morphology and function in vitro due to cocultivation with
another live cell type (39)
. Note intracellular albumin
staining throughout the hepatocyte island regardless of proximity to
the heterotypic interface. In this review, the term heterotypic
interface will be used to describe the spatial dimension over which
fibroblasts and hepatocytes undergo coplanar cell–cell contact (i.e.,
in an island of hepatocytes surrounded by fibroblasts, this would
correspond to the island perimeter). Figure 3
depicts the functional outcome of this culture method and the clear
demonstration of retention of a liver-specific function, albumin
secretion, for many weeks (>5 wk) (37)
. In fact,
relatively stable albumin production has been observed as long as 65
days (40)
.
|
|
This stabilization of liver-specific functions has been reported
for cocultures with both liver-derived cell types as well as
non-liver-derived endothelia and fibroblasts. Table 1
summarizes studies of both liver and non-liver-derived cell types in
hepatocyte cocultures. Liver-derived cell types include rat liver
epithelial cells of presumed biliary origin (31
, 37
, 39
40
41
42
43
44
45
46)
, stellate (Ito, fat-storing) cells
(47
48
49)
, sinusoidal endothelial cells (50
, 51)
, Kupffer cells (24
, 52
53
54
55
56)
, and the entire
`nonparenchymal' fraction of isolated liver cells
(57
58
59)
. Although this effect on morphological and
functional differentiation was originally thought to be species
specific, many other cell types from other organ systems and species
have since been shown to influence isolated rat hepatocytes. This
effect has been demonstrated with rat hepatocytes, to varying degrees,
using embryonic murine 3T3 and C3H 10T1/2 cells (29
, 32
, 45
, 51
, 60
61
62
63
64)
, rat dermal fibroblasts (51)
,
Chinese hamster cells (25
, 28)
, canine kidney epithelia
(65)
, bovine aortic endothelia (51
, 66)
, and
human fibroblasts and lung epithelia (26
, 27
, 65
, 67
, 68)
.
In addition, similar findings have been observed for adult human and
fetal rat, chick, and porcine hepatocytes (28
, 30
, 68
69
70
71)
. Finally, the effects of cell–cell communication are
also reciprocal; stabilization of function and responsiveness of
nonparenchymal cells when cocultured with hepatocytes has also been
reported (48
, 59)
.
|
Typically, the time course of events in hepatocyte cocultures is
similar, independent of the choice of secondary cell type, culture
configuration, or cell concentration. Most cultures have preserved
hepatospecific synthetic functions for prolonged periods (1 to 10 wk).
The effects on hepatocyte function are inducible for 3–7 days, after
which hepatocyte `rescue' is unattainable (41
, 69)
. This
is noted graphically in Fig. 3
, which demonstrates the comparable
efficacy of initiating coculture on both day 1 and day 7 of hepatocyte
culture. In addition, the time course over which albumin synthetic
capability increases before stabilization appears to remain fairly
constant, 6–10 days.
Culture configurations for many of these systems employed variations in
the ratio of cell types and media composition. Typically, investigators
have explored ratios of cell numbers of ~1:1 (nonparenchymal:
hepatocyte); however, this has varied among studies from 10:1 to 1:10
(56
, 58)
. In addition, many media formulations have
included additions of insulin and glucocorticoids such as
hydrocortisone or dexamethasone to inhibit fibroblastic overgrowth.
Last, both serum-free and fetal bovine serum formulations have been
used successfully. In addition to viable cells in the above culture
configurations, experiments have been performed with feeder layers,
including irradiated (38)
, desiccated and heated
(58)
, glutaraldehyde-fixed (72)
, or mitomycin
C-treated (61)
nonparenchymal cells. One study compared
the relative effect of viable cells vs. feeder layers and reported
comparable effects on the examined markers, DNA synthesis
(58)
. Similarly, glutaraldehyde-fixed endothelial cells
elicited a comparable response to viable cells (72)
when
cocultured with hepatocytes.
In general, a variety of coculture models have met with significant success in maintenance of many hepatospecific functions. A summary of the existing data on the morphological, mitotic, and biochemical effects of coculture on hepatocytes is presented below.
|
EFFECT OF COCULTURE ON HEPATOCYTE MORPHOLOGY AND FUNCTION |
|---|
|
TOPABSTRACTINTRODUCTIONCELL-CELL INTERACTIONS IN THE...COCULTIVATION OF HEPATOCYTES AND...EFFECT OF COCULTURE ON...MECHANISM OF INDUCTION OF...METHODS TO EXAMINE INFLUENCE...MICROFABRICATED COCULTURES:...Role of the Extent...Role of Heterotypic Interactions...Role of...CLINICAL APPLICATIONS OF...SUMMARYREFERENCES |
|---|
); however,
when isolated and cultured on tissue culture plastic, they adhere and
spread, losing many of their characteristic features (73)
.
Over time, the nuclei undergo karyolysis (loss of nucleus associated
with cell death), the cell borders become indistinct, the cytoskeleton
undergoes rearrangement with actin stress fiber formation and a
`fibroblastic' appearance, and cells ultimately detach and die. In
contrast, hepatocyte cocultures exhibit stereotypical polygonal
morphology with distinct nuclei and nucleoli, well-demarcated
cell–cell borders, and a visible bile canalicular network for many
weeks.
Cocultures have also been shown to express many liver-specific proteins
such as albumin (Table 2
). Murine 3T3's have been shown to induce the highest levels of albumin
secretion by hepatocytes (4.2 to 15 µg/106
cells/h) (74
, 75)
, followed by rat liver endothelial cells
(3.1 µg/106 cells/h) (51)
, rat
dermal fibroblasts (3.1 µg/106 cells/h)
(51)
, rat liver epithelial cells (2.9
µg/106 cells/h) (39)
, and bovine
aortic endothelial cells (1 µg/106 cells/h)
(51)
.
|
The level of regulation involved in induction of liver-specific protein
production has also been investigated. The cause of the increases
observed in protein synthesis and mRNA was studied using in
vitro transcriptional assays from isolated nuclei as well as
`rescue' experiments wherein mRNA was allowed to decline and then
observed to reappear. Albumin, pyruvate kinase, transferrin, and
various subunits of glutathione S-transferase were found to be
regulated primarily at the transcriptional level, with at least some
component of post-transcriptional mRNA stabilization (41
, 44)
.
Markers of detoxification capability, such as cytochrome P-450 enzyme
activity, have also been observed to increase in amount and stability.
For the most part, P-450 isoenzymes 1A1, 2B1 and 3A1 seem to be the
best stabilized after 1 wk (45
, 46
, 65
, 76
77
78
79
80
81)
. In
comparison to conventional hepatocyte cultures such as Matrigel, total
P-450 content was found to be elevated twofold in uninduced cocultures
(62)
. In addition, hepatocytes retained inducibility of
P-450 enzymes by prototypic inducers (61
, 62
, 77
, 79
80
81)
.
One study showed a 12- to 15-fold increase in mRNA levels for CYP2B1
after 7 days of induction of cocultures by phenobarbital as compared to
hepatocytes cultured alone (81)
. Although each isoenzyme
shows a different induction pattern, some have been reported to be
inducible for up to 2 months (82)
. On the other hand,
cocultivation does not preserve all isoenzyme activities; some, such as
2C11, 2C6, and 2E1, were reported to decline continuously (45
, 46
, 79
, 80)
.
The influence of hepatocyte coculture on other markers of
detoxification pathways such as the conjugating (phase II) systems have
also been studied. The glutathione-S-transferase (GST) family, a family
of dimeric enzymes that catalyze the conjugation of reduced
glutathiones to electrophiles, is the most commonly studied of the
phase II systems. Like the cytochrome P-450 isoenzymes, GSTs are
generally more stable in coculture than pure hepatocyte culture
(31
, 43
44
45
46
, 83
, 84)
. Stable expression of some subunits
has been noted for 12 days in coculture, with significant quantitative
differences between various nonparenchymal cell sources (44
, 78)
. In addition, as seen in the P-450 family, quantitative
differences for each subunit are noted, with subunits 3 and 4 being the
most stable. The mechanism of stabilization of the GSTs in coculture is
thought to be both at the transcriptional and mRNA stabilization level
(44)
. GSTs may also be induced in cocultures by
stereotypical inducers (i.e., phenobarbital), with each subunit
responding differentially (43
, 46)
. Finally, the GST
family differs from the cytochrome P-450 family in one important arena:
expression of a fetal GST (7)
is induced in both pure
hepatocyte and hepatocyte cocultures, though not typically seen in the
adult liver (43
, 44
, 83
, 84)
.
Functional contacts were also observed in hepatocyte cocultures. Tight
junctions were detected by the presence of ZO-1 in cocultures
(47)
. Gap junctions (connexin 32) were localized by
indirect immunofluorescence and/or by microinjection of Lucifer yellow.
In general, gap junctions were formed only in homotypic hepatocyte
interactions (40)
. One notable exception is the formation
of heterotypic functional gap junctions between hepatocytes and
fat-storing cells (connexin 43) (49)
. In addition, the
degree of induction of albumin synthesis correlated with increased
levels of connexin 43 in various fat-storing cell clones. This
observation is particularly significant because fat-storing cells
maintain direct contact with hepatocytes in vivo. This model
suggests that hepatocyte function may be influenced by the degree of
cell–cell interaction (and thus the formation of heterotypic gap
junctions). Therefore, though homotypic hepatocyte gap junctions are
commonly noted in coculture, the presence and influence of heterotypic
gap junctions are cell type dependent.
Spatial and temporal distribution of gap junctions between hepatocytes
was also examined under various coculture conditions. Mesnil et al.
(40)
noted that the number of dye-coupled hepatocytes per
injection gradually increased with coculture time from a single cell
early in coculture to 9 cells by 25 days. Once hepatospecific functions
stabilized, all hepatocytes in a given colony were found to express
both functional gap junctions and albumin regardless of the proximity
to the heterotypic nonparenchymal cell type (i.e., even hepatocytes
that do not undergo direct contact with nonparenchymal cells retain
function). This data provided indirect evidence that the signal for
induction of liver-specific function is not confined to the heterotypic
interface, i.e., signals may propagate through confluent hepatocyte
populations. Thus, studies of gap junction expression in hepatocyte
cocultures suggested that both temporal and spatial variations exist.
Even though these model systems offer the opportunity to study complex modes of cell–cell communication, there are significant confounding factors in such randomly distributed cocultures. Hepatocyte colony size varies with cell seeding density as well as hepatocyte adhesion, aggregation, and migration. The approximate size of hepatocyte colonies (estimated from published micrographs) in these studies was 100–200 µm in diameter containing 5–15 cells. These phenomena may ultimately be better examined by using a culture system that produces spatially uniform cell–cell interactions.
An additional notable feature of certain hepatocyte cocultures as
compared to pure hepatocyte cultures is the ability to synthesize DNA
in vitro. This effect has been noted in hepatocyte
cocultures of both liver-derived and non-liver-derived cell types. An
important distinction must be made between DNA synthesis and cell
growth per se, especially in light of the known ability of
hepatocytes to multinucleate both in vivo and in
vitro. Given this caveat, two investigators have reported
significant levels of DNA synthesis/division in cocultures. When rat
liver cells were cocultured with the entire nonparenchymal liver
fraction on felt templates, parenchymal cells of 15–30 µM diameter
increased in number by 10-fold over 48 days as measured by enzymatic
separation of cultures and counts of cell populations by size. In
addition, thymidine incorporation was measured over 48 days and found
to reach a maximum at 24 days of culture (82)
. In
contrast, Shimaoka et al. (58)
found an increase of
labeling index from 13% of hepatocytes in pure cultures to 35% of
hepatocytes in cocultures with nonparenchymal cells. This stimulatory
effect of nonparenchymal cells on DNA synthesis by adult hepatocytes
varied in a dose-dependent manner, where cultures with low hepatocyte
densities demonstrated a twofold increase in labeling index over high
hepatocyte densities. Furthermore, DNA synthesis reached a maximum at 3
days of culture.
DNA synthesis was also examined by coculture with non-liver-derived 3T3
clones, producing varied results. Some investigators have reported
20–30% labeling indices (61)
whereas others have
reported minimal thymidine uptake (58
, 62)
. In other
non-liver-derived cell types such as human embryonic lung, canine
kidney, and monkey kidney epithelia, minimal thymidine uptake was
reported (65)
. Thus, in general, it appears that very
little hepatocyte growth occurs in coculture configurations with
non-liver-derived nonparenchymal cells. This suggests that
growth-arrest of the `alternative cell type' in this type of
hepatocyte cocultures may afford adequate control over preservation of
approximately constant cell numbers for precise study of both
subpopulations.
In summary, cocultivation of hepatocytes with nonparenchymal cells has
been shown to preserve stereotypical hepatocyte morphology and a
variety of synthetic, metabolic, and detoxification functions of the
liver. Although cell communication clearly plays a role in the
regulation of these hepatospecific functions, the complex rules that
govern the influence of homotypic cell interactions, heterotypic cell
interactions, cell density, and ratio of cell populations remain
undetermined. These issues may be elucidated by use of a model system
that allows precise control over these interactions. One such model
system, based on cellular micropatterning techniques, was recently
developed and is discussed in detail elsewhere in this review
(85)
.
|
MECHANISM OF INDUCTION OF LIVER-SPECIFIC FUNCTION IN HEPATOCYTES |
|---|
|
TOPABSTRACTINTRODUCTIONCELL-CELL INTERACTIONS IN THE...COCULTIVATION OF HEPATOCYTES AND...EFFECT OF COCULTURE ON...MECHANISM OF INDUCTION OF...METHODS TO EXAMINE INFLUENCE...MICROFABRICATED COCULTURES:...Role of the Extent...Role of Heterotypic Interactions...Role of...CLINICAL APPLICATIONS OF...SUMMARYREFERENCES |
|---|
Many studies attempting to discern the contribution of soluble factors
in coculture systems have produced contradictory results. Morin et al.
(72)
reported that a transmembrane culture system using
hepatocytes seeded on a 0.45 µM pore size filter and endothelial
cells in an underlying well induced similar levels of albumin secretion
as control cocultures with sinusoidal cells in contact with hepatocytes
on similar filters. In contrast, Donato et al. (79)
reported no significant improvement in P450 activity when hepatocytes
were cultured on the bottom of a similar trans-well system with a 0.4
µM pore size and MS epithelial cells on top of the insert over pure
hepatocyte cultures. The differences in these findings suggest that
perhaps culture configuration (i.e., hepatocyte adhesion to a transwell
filter as compared to tissue culture dishes) is important. In addition,
use of media conditioned by the second cell type on pure hepatocyte
cultures has been shown to be almost universally ineffective (58
, 61
, 62)
. At least one dissenting study showed a partial effect
of rat liver epithelial cell conditioned media on hepatocyte cultures
(half-maximal increases in levels of glutamine synthetase activity
relative to control cocultures). Conditioned media obtained from
cocultures showed no effect on glutamine synthetase activity (the only
function tested). This implies that any potential soluble factor is not
present in excess in cocultures due to its uptake/degradation by
hepatocytes (42)
.
Comparatively, studies of extracellular matrix-mediated effects
on liver-specific gene expression have been even less conclusive.
Although many groups have reported matrix deposition patterns specific
to cocultures, no causative effects of this matrix have been shown. In
particular, reticulin fibers were observed in cocultures but absent in
both types of pure culture (39
, 51
, 86)
. Other
extracellular matrix components have been observed in cocultures with
indirect immunofluorescent techniques including collagens I, IV,
fibronectin, laminin, and entactin (37
, 47
, 51)
.
Mesenchymal cells typically are characterized by their ability to
produce collagen I and fibronectin matrix molecules, whereas
hepatocytes have been shown to primarily produce collagen IV and
laminin. As a result, the cellular source of extracellular (ECM)
deposition in cocultures is unclear. In addition, endothelial cells
were found to produce perlecan in vivo (heparan
sulfate-proteoglycan), a known mediator of some liver-specific
functions (87)
, which may implicate proteoglycans in some
component of the coculture effect. However, this ECM effect on liver
cells is unlikely to be descriptive of the mechanism by which stellate
cells induce hepatospecific function since they were consistently
negative for perlecan. Finally, two groups have attempted to modulate
the effect of potentially ECM-mediated events by (1)
crudely assessing the distance over which the signal can travel from
the heterotypic interface (42)
and by (2)
treating feeder layers with enzymes specific for ECM destruction
(58)
. Shrode et al. (42)
found up-regulation
of glutamine synthetase production up to a few millimeters from the
heterotypic interface; they suggest that large, insoluble ECM molecules
are likely mediators since they would have limited diffusivity at
critical concentrations. In contrast, the effects of direct cell
contact communicated via gap junctions are discounted by the authors as
they hypothesize that such a signal would travel over a limited
distance. Finally, Shimaoka et al. (58)
reported that the
DNA synthesis they monitored in cocultures was acid-, trypsin-, and
collagenase-sensitive, implicating some protein containing collagen. In
addition, precultured feeder layers induced DNA synthesis earlier than
fresh feeder layers, indicating that the presence of some material was
rate-limiting. The authors suggest that the insoluble molecules (ECM or
membrane receptors) in the feeder layers were responsible for the
observed effects, although soluble factors entrapped in the feeder
layers may also have played a role.
Until recently, the role of direct contact of cells, the other
potential mechanism involved with induction of liver-specific function,
has remained unclear. Mesnil et al. (40)
showed that only
hepatocytes in close proximity to epithelial cells in sparse cultures
remained viable and differentiated as compared to those that appeared
to lack heterotypic contact. The authors suggest the importance of cell
contact based on this indirect evidence; however, it seems clear that
local deposition of ECM or local concentrations of critical soluble
factors cannot be ruled out as causes for the preservation of viability
and differentiation. More rigorous evidence supporting the role of
membrane contact as a potential mechanism was reported in 1991 by Corlu
et al. (88)
. These authors identified a cell surface
protein (liver-regulating, or LRP) that seemed to be involved in the
establishment and maintenance of hepatocyte differentiation in
coculture with liver epithelial cells. They demonstrated the ability to
modulate albumin secretion, cytoskeletal organization, and ECM
deposition by addition of a monoclonal antibody against LRP.
Furthermore, the authors discount extracellular matrix as potential
ligand for LRP due to the inability of anti-LRP antibody to modulate
cell adhesion to immobilized ECM. In addition, this inhibitory effect
was produced only upon addition of the antibody early in culture. The
authors suggest that this time dependence supports the role of
cell–cell contact in the coculture effect due to the indirect evidence
that establishment of cell–cell contacts occurs during the same time
frame in culture. Finally, it seems that LRP is almost certainly not
the whole story; although some cell types that induce liver-specific
functions in hepatocytes stained positive for LRP (sinusoidal cells and
Ito cells), other cell types did not (vascular endothelia, biliary
ductal cells) (89)
. Therefore, although the presence of
LRP may modulate hepatocyte function in epithelial coculture, the
absence of LRP in coculture with other cell types does not seem to
prevent induction of liver-specific functions.
Other modes of direct contact such as gap junctional communication may
also play a role in cell signaling. In one study, levels of connexin 43
expressed by fat-storing cell subclones correlated with albumin mRNA
levels in cocultured hepatocytes. Functional heterotypic gap junctions
were observed as a result of connexin 43 protein synthesis
(49)
. This mode of cell signaling may be particularly
important in hepatocyte interaction with Ito cells compared to other
cell types due to the potential relevance of this signaling mechanism
in vivo I(90)
. In addition, communication
between cells has also been implicated in transport of reactive
intermediates (91)
.
Due to the relationships described between dedifferentiation in tumors
and decrease in gap junctions, studies were also done to assess the
necessity of homotypic gap junctional communication for the
stabilization of differentiated functions. Traiser et al.
(92)
found that gap junction intercellular communication
could be effectively blocked with minimal effects on the stabilization
of xenobiotic metabolic enzyme activities (another liver-specific
marker), suggesting they may be unimportant in preservation of the
hepatocyte phenotype. However, the results of this study may not be
conclusive due to the potential effects of the compounds used for
interfering with gap junctional communication on induction of P-450
enzymes. The notion that hepatocyte gap junctions may be decoupled from
liver-specific functions is also supported by the lack of observable
gap junctions in well-established hepatic culture systems after 24 h (73
, 93)
.
In summary, despite the substantial data existing on potential mediators of cell communication in cocultures (receptors, gap junctions, cytokines, ECM), the mechanisms by which coculture of hepatocytes with other cell types induce and stabilize liver-specific function and viability are undefined. Indeed, many distinct mechanisms may operate in concert, each modulating a subset of hepatospecific functions. For example, expression of glutamine synthetase, albumin, and connexin 32 may each be independently regulated both in the time course of expression and rates of secretion. The difficulty with which homotypic and heterotypic interactions can be experimentally uncoupled, however, has made their role in these processes difficult to assess. Here we review methods that have been used to study the role of cell–cell interactions, including a `micropattern'-based technique that has recently facilitated some novel insights in this area.
|
METHODS TO EXAMINE INFLUENCE OF CELL–CELL INTERACTIONS ON LIVER-SPECIFIC FUNCTIONS |
|---|
|
TOPABSTRACTINTRODUCTIONCELL-CELL INTERACTIONS IN THE...COCULTIVATION OF HEPATOCYTES AND...EFFECT OF COCULTURE ON...MECHANISM OF INDUCTION OF...METHODS TO EXAMINE INFLUENCE...MICROFABRICATED COCULTURES:...Role of the Extent...Role of Heterotypic Interactions...Role of...CLINICAL APPLICATIONS OF...SUMMARYREFERENCES |
|---|
. Control of cell–cell interactions fall into two general categories:
1) prevention of contact or 2) modulation of the
degree of contact. Prevention of contact has been achieved by
cocultures with porous filter inserts, insertion of crude spacers, or
conditioned media experimentation. As detailed above, these culture
configurations have led to some novel insights but have also been
limited by variations in media sampling, storage, filter material, and
cell seeding densities. In addition, absolute lack of contact is
difficult to ensure; for example, transfer of detached cells in
conditioned media or protrusion of cell processes through the porous
filter is difficult to completely rule out.
|
Another approach at prevention of contact was reported by Shrode et al.
(42)
. Creation of a cell-free annulus was achieved through
the addition of a polymer spacer to a culture dish by use of rubber
cement adhesive. This spacer was then removed, resulting in a defined,
relatively large (~ mm) cell-free annulus between the cell
populations. Cell populations then grew together, allowing study of the
role of local cell contact in zonation of the liver. Although this
method did show that hepatocyte populations can undergo induction
locally, the method is limited by the undefined underlying substrate
(residual adhesive) and relegation to relatively large dimensions of
annuli (spacers must be large enough to manipulate manually).
In addition to control of cell–cell interactions by prevention of
contact, modulation of the degree of cell contact has also been
attempted. Both conventional techniques (variations in seeding density)
as well as more specialized systems (addition of confluent coverslips
to confluent cultures) have been used. Variations in seeding density
were used by Guguen-Guillouzo (94)
to study the effects of
cell contact on hepatocyte differentiation. They examined effects of
lower seeding densities by seeding the same cell numbers in a two
different size flasks. This method is simple and reproducible, but
heterotypic cell contacts occur due to random events such as attachment
during cell seeding. In addition, a confounding factor in these
experiments may be the ability of the nonparenchymal cell type to
divide: lower seeding densities may permit increases in the
nonparenchymal population and the accompanying soluble factors
synthesized by these cells.
Another study examined the role of cell contact by addition of
confluent cultures of hepatocytes on a coverslip to the center of
confluent cultures of either fibroblasts or fibroblast/hepatocyte
cocultures (58)
. This technique also attempts to examine
the role of local contact, and these studies succeeded in probing the
role of soluble factors in a novel way; however, it is likely that the
results were confounded by cell death underlying the coverslip and the
significant topological variations in the culture (height of a
coverslip is typically 100–300 µ). Another similar study using
coverslip inserts examined the role of local stellate cell–hepatocyte
interactions (48)
. This study demonstrated a localized
`paracrine' signal ~10 cell widths from the heterotypic interface,
providing valuable insight into the potential mediators of this
signaling process. However, this technique is also relegated to
relatively large dimensions and significant potential for artifacts
secondary to local cell damage.
Although some of these models have successfully examined the
outcome associated with a complete lack of heterotypic cell contact, no
existing experimental techniques have conferred the ability to
systematically and uniformly vary the degree of local heterotypic cell
interaction. Rather, cell–cell interaction has been typically dictated
by poorly controlled parameters such as cell attachment, aggregation,
and migration or by gross manipulations of culture configurations.
Recently, a method was reported that significantly advance the current
state-of-the-art reviewed here (85)
. These techniques
allow control over the spatial distribution of two cell types in planar
cultures and systematic investigation of the effects of cell–cell
interactions on tissue function.
|
MICROFABRICATED COCULTURES: CONTROL OF HOMOTYPIC AND HETEROTYPIC CELL INTERACTIONS |
|---|
|
TOPABSTRACTINTRODUCTIONCELL-CELL INTERACTIONS IN THE...COCULTIVATION OF HEPATOCYTES AND...EFFECT OF COCULTURE ON...MECHANISM OF INDUCTION OF...METHODS TO EXAMINE INFLUENCE...MICROFABRICATED COCULTURES:...Role of the Extent...Role of Heterotypic Interactions...Role of...CLINICAL APPLICATIONS OF...SUMMARYREFERENCES |
|---|
, 75
, 85
, 95
, 96)
. Indeed, cellular `micropatterning' has already
been useful in the study of many diverse biological phenomena. Spatial
control over cell attachment and spreading has facilitated an
unprecedented level of sophistication in the investigation of
mechanisms of nerve growth cone guidance, influence of cell shape on
growth and apoptosis, and growth and orientation of fungal pathogens
(97
98
99
100
101
102)
. Improvements on existing micropatterning
technology recently allowed control over the degree of interaction
between two different cell populations. These techniques were used to
explore the effects of homotypic and heterotypic cell interaction on
tissue function.
The photolithographic technique developed for the micropatterning of
cells to allow spatial control over two distinct cell populations is
depicted in Fig. 5
. Borosilicate wafers were patterned with photoresist (a polymer that
has variable solubility with exposure to ultraviolet light) by exposure
to light through a prefabricated chrome mask (Fig. 5A
).
Patterned substrates were used to control subsequent immobilization of
collagen I (103
, 104)
(Fig. 5B
). The
localization of adhesive extracellular matrix (here, collagen I)
allowed for patterning of the first cell type, primary hepatocytes
(Fig. 5C
). Hepatocytes exhibited a well-spaced morphology
with distinct nuclei and bright intercellular borders. Subsequent
deposition of a nonparenchymal cell type (here, 3T3-J2 fibroblasts)
allowed for spatial control over heterotypic cell interactions in the
cellular microenvironment (Fig. 5D
). This technique offers
the ability to present different adhesive ligands to each population
within a single culture (here, collagen I to hepatocytes, and
serum-adsorbed proteins to fibroblasts), which cannot be achieved with
many conventional methods. The versatility of this technique is derived
from the ability to alter cell–cell interactions with ease via use of
different chrome masks; therefore, the size of each cell subpopulation
may be maintained while allowing variation in the extent of heterotypic
interaction. Conversely, the level of heterotypic interaction may be
held constant while allowing variation of the number of cells in each
subpopulation.
|
Although the techniques described above enable the investigation of
complex interactions between two cell types, the existing technology
suffers from a number of limitations. First, although cells are
patterned initially, tissue morphogenesis is not restricted in these
cultures. Reorganization through cell motility was observed to be
dependent both on hepatocyte island diameter as well as
center-to-center spacing, with islands greater than or equal to 490
µM retaining an observable pattern for at least 2 wk whereas smaller
patterns reorganized into cord-like structures on the order of days.
Second, the success of this technique is dependent on the relative
cell–cell and cell–substrate adhesiveness of each cell type, i.e.,
the relative preference of nonparenchymal adhesion to the substrate
rather than the preseeded hepatocyte surface. This aspect of a
micropatterned coculture could be studied using labeling with
fluorescent vital dyes and/or confocal microscopy. Finally, the method
depicted by Fig. 5
is limited to confluent cocultures on glass
substrates; however, many applications may require separation of cell
populations and/or the flexibility to use a variety of underlying
substrates. Recent reports on use of polymeric microchannels to direct
protein immobilization provide methods to pattern cells on a variety of
surfaces from polystyrene to thin metal films (105
106
107)
,
thus broadening the potential of the techniques described in this
review.
|
Role of the Extent of Heterotypic Interactions in Induction of Liver-Specific Function |
|---|
|
TOPABSTRACTINTRODUCTIONCELL-CELL INTERACTIONS IN THE...COCULTIVATION OF HEPATOCYTES AND...EFFECT OF COCULTURE ON...MECHANISM OF INDUCTION OF...METHODS TO EXAMINE INFLUENCE...MICROFABRICATED COCULTURES:...Role of the Extent...Role of Heterotypic Interactions...Role of...CLINICAL APPLICATIONS OF...SUMMARYREFERENCES |
|---|
A–D). Micropatterned hepatocytes visualized with a
fluorescent vital dye and were found to adhere predominantly to
collagen-modified islands (Fig. 6E-H
). Addition of 3T3-J2
fibroblasts to micropatterned hepatocytes resulted in micropatterned
cocultures with marked alterations in initial heterotypic interface
despite similar numbers of fibroblasts and hepatocytes across
conditions, as depicted in Fig. 6
(I–L).
|
To determine the effect of modulation of heterotypic cell interactions
on liver-specific function, cultures were characterized for expression
of liver-specific biochemical (urea and albumin secretion) and
immunohistochemical markers (intracellular albumin staining) and
overall DNA content. We measured these markers in five different
micropatterned cocultures with varying degrees of heterotypic
interaction, each with a matched control of micropatterned hepatocytes
in the identical configuration (i.e., no heterotypic interaction). In
all micropatterned cocultures, urea synthesis was found to be
significantly increased by 2.5- to 6-fold over micropatterned
hepatocyte (only) controls on day 11, indicating that the induction of
urea synthesis in hepatocytes was due to cocultivation with fibroblasts
(Fig. 7
A). The degree of improved function over control
micropatterned hepatocyte (only) cultures varied with the degree of
heterotypic interaction. Two patterns of up-regulation of this
liver-specific marker emerged: 1) the three smallest island
configurations (36,100, 490 µM, with relatively increased heterotypic
interaction) showed up-regulation of urea synthesis to similar levels,
whereas 2) the two larger island configurations (6.8, 17.8
mm) showed relatively little up-regulation (~50% of cultures with
greater heterotypic interaction). Therefore, a statistically
significant increase in urea synthesis production was achieved in
certain pattern configurations by modulation of the initial cellular
microenvironment despite similar cellular constituents.
|
Similarly, all micropatterned cocultures had marked
induction of albumin secretion when compared to micropatterned
hepatocyte (alone) controls (Fig. 7B
). By day 11, all
micropatterned hepatocyte (alone) conditions had negligible levels of
albumin secretion. In contrast, cocultivation with fibroblasts produced
variations in the degree of up-regulation of this marker with the
degree of heterotypic interaction. Again, two patterns emerged:
1) dramatic up-regulation to similar levels of albumin
secretion in the three smallest island configurations (with relatively
increased heterotypic interactions) and 2) relatively modest
up-regulation (~30% of cultures with greater heterotypic
interaction) in the two larger island configurations. Therefore, a
substantial increase in albumin production was achieved in certain
pattern configurations by modulation of the initial heterotypic
cellular microenvironment.
Thus, variation of initial heterotypic cell–cell interactions was
found to modulate long-term bulk tissue function for at least two
liver-specific functions. The kinetics of this response are described
in detail elsewhere (75)
. Briefly, micropatterned
cocultures demonstrated increased albumin synthesis rates until
stabilization at day 9 for all configurations, whereas urea synthesis
was either stabilized or increased to a plateau by day 3. Thus, despite
the similarity in long-term effects of heterotypic interaction on two
different markers of liver-specific function, the kinetic response of
this induction varied. This finding is consistent with known
differences in the patterns of recovery for various liver-specific
functions in other hepatocyte culture systems (73)
. In
addition, randomly distributed cocultures (i.e., not micropatterned) in
the same model system had similar kinetics for induction of albumin
secretion, but induction of urea synthesis was delayed until
stabilization at day 7–10 (74)
. Differences in the
kinetic response of randomly distributed cocultures from micropatterned
cocultures may be due to reorganization of cell populations over time,
artifactual due to differences in culture conditions (i.e., fibroblast
adhesion to collagen I vs. serum-adsorbed proteins), or reflective of a
time delay in signal propagation through randomly distributed
cocultures.
As previously mentioned, extensive studies of the effect of initial
cellular microenvironment on liver-specific function in cocultures are
scant due to the limitations of existing experimental methods. One
study attempted to examine the effect of local microenvironment by
variation in size of culture plate (94)
. This study of
human hepatocytes cocultured with rat liver epithelial cells (RLEC)
used the same numbers of cells in 25 cm2 and 75
cm2 dishes. Heterotypic cell interactions were
largely dictated by seeding density, plate size, and random cell
aggregation. Their results suggest twofold higher albumin secretion in
sparser cultures; however, this result may have been affected by RLEC
number and associated cellular products (due to potential for increased
RLEC growth on larger plate), differences in nutrient supply (oxygen,
glucose, essential amino acids due to differences in amount of media),
increased heterotypic interactions, or some combination
thereof. In contrast, these novel microfabrication techniques allowed
the demonstration of a threefold increase in albumin secretion due
solely to variations in initial heterotypic cell–cell interaction.
Thus, it seems that the local cellular microenvironment has been
definitively isolated as an important modulator of liver-specific
function.
|
Role of Heterotypic Interactions in Localized Induction of Liver-Specific Function |
|---|
|
TOPABSTRACTINTRODUCTIONCELL-CELL INTERACTIONS IN THE...COCULTIVATION OF HEPATOCYTES AND...EFFECT OF COCULTURE ON...MECHANISM OF INDUCTION OF...METHODS TO EXAMINE INFLUENCE...MICROFABRICATED COCULTURES:...Role of the Extent...Role of Heterotypic Interactions...Role of...CLINICAL APPLICATIONS OF...SUMMARYREFERENCES |
|---|
