Vol.
4, 84 1 -847,
October
1993
Cell Growth
& Differentiation
841
Immortal
Cell Lines Isolated
from Heart Differentiate
to an Endothelial
Cell Lineage
in the Presence
of Retinoic
Acid’
Ala-Eddin
Al Moustafa
Bloomfield
Center
for
Research, Sir Mortimer
Quebec, Canada H3T
Immunology,
Montreal,
McGill
in Aging,
E. Chalifour2
Lady
Davis
Institute
for
small (22 Kd) tumor
a primary
transcript
Medical
Canada
for
Studies
in Aging,
McGill
University,
the mouse
H3A 2B4
genic
cardiomyocyte
tricles.
The
endothelium
6/7/93;
study
revised
was supported
Canada
and the Heart
cheur boursier
award
2
To whom
requests
and
from
3
The abbreviations
accepted
by funding
saic
of different
biochemical
siderable
from
and
virus of the
(56 Kd), and
should
McGill
Canada.
be addressed,
Center
for Studies
Council
cytoplasmic
RA-binding
3-acetate;
DMSO, dimethyl
lipoproteins;
dil-Ac-LDL,
acetylated
1 ,3,3,3,3-tetramethyl-indocarbocyanine
DMEM,
Dulbecco’s
modified
Eagle’s
line.
at Department
in Aging,
of
of Mi-
McGill
PV, polyomavirus;
Uni-
PVLT, poly-
1 ; TGF-f31,
transforming
receptor;
RXR, retinoid
protein;
sulfoxide;
LDL
labeled
perchlorate;
medium;
PBS,
TPA,
LDL,
involving
atria
and
yen-
of cardiomyocytes
blood
vessels.
proliferate
only in fetal life,
dependenton
cell enlargement
have indicated
that adult car-
biological
cell
layer
of squamous
epi-
blood and lymph vessels. The endoof the microvascular
system is a mo-
endothehial
markers
rat (18,
cells
with
varying
(1 2). Endothelium
importance
shapes
and
is an organ of con-
as it serves
as a protective
19) tissues.
Cultured
endothelial
cells,
when
X
12-O-tetralow density
with
1 ,1 -dioctadecyl
HS,
horse
serum;
phosphate-buffered
sa-
(23-25).
RA can induce differentiation
of mouse embryonal
carcinoma cells in vitro and suppress malignant
transformation
both in vitro and in vivo (26-28).
RA was shown to suppress
in vitro transformation
by PV middle T antigen (29). RA interactswith two distinctclasses of intracellularproteins: the
RA receptors (RAR-a, RAR-/3, and RAR-’y) or the retinoid
X
receptors
used are: Kd, kilodalton(s);
large T antigen;
MT, metallothionein
factor
13i; RA, retinoic
acid;
RAR, RA
receptor;
CRABP,
decanoylphorbol-1
Research
a uni-
grown under appropriate
conditions,
will form organized
structures
resembling
capillary
blood vessels, an activity
which has been described
as “angiogenesis
in vitro”
(20).
Culture of mouse endothehial
cells is somewhat
more problematic (2i ) but transformed
cell lines have been obtained
from hemangiomas
produced
in transgenic
mice expressing
polyomavirus
middle T antigen (22). Endothelial
cells have
been transformed
by SV4O, BK, and JVC virus with detectable T antigen expression
and DNA replication,
and by
8/2/93.
the Medical
trans-
displaying
and antithrombolytic
layer and performs transport functions
for molecules
entering
and exiting
the bloodstream
(reviewed
in Ref. 1 3). During development,
the endothelial
lineage is thought
to originate
from the hemangioblast,
a
precursor to both endothehial
and hematopoietic
cells (14).
Endothelial
cells have been successfully
cultured
in vitro
from diverse sources including
bovine (i 5, 16), human (1 7),
Stroke Foundation
of Quebec.
L. E. C. has a cherthe Fonds de Ia Recherche
en Sante du Quebec.
for reprints
crobiology
and Immunology,
versity,
Montreal,
Quebec,
omavirus
growth
7/30/93;
gene (2-4).
PVLT, linked to
six MT-PVLT
heart
in the nuclei
is a single
thelium
which
lines
thelial cell population
v-myc/mil
1 This
hypertrophy
PVLT was expressed
ofthe
of
diomyocytes
have a limited
capacity
to re-enter the cell
cycle traverse, and, indeed, tumors of cardiomyocytes
are
extremely
rare (8-1 0). Recently,
rabbit cardiac
fibroblast
cells were shown to convert to a myocyte
phenotype
after
incubation
with TGF-f31 (ii).
Introdudion
Received
(5, 6). One
an enlarged
and in cells lining
coronary
In mammals,
cardiomyocytes
with further
postnatal
growth
(7). Numerous
experiments
Previously, we isolated a single line of transgenic mice
which develop an enlarged heart due to the expression
of the immortalizing
gene, polyomavirus
large T antigen.
Immortal
cell lines were isolated from adult transgenic
but not from nontransgenic
hearts. All of the 24 cell
lines expressed vimentin and fibronedin
but not desmin
or myosin heavy chain. We conclude that the cell lines
are of non-muscle
origin. Six cell lines were chosen
for further study. All six cell lines demonstrate
profound
morphological
and biochemical
effects when incubated
with 1 O M to 1 0
M retinoic
acid. The retinoic
acid-treated
cell lines showed arrested cellular
proliferation
and aligned to form rows and vesicle-hike
strudures.
Cycloheximide
inhibited these retinoic
acid-induced
changes, indicating a need for continued
protein synthesis. Retinoic acid-treated,
but not
untreated,
cells lost expression of vimentin and
fibronedin,
gained the ability to incorporate
acetylated
low density lipoprotein,
and expressed Fador
VIII-related
antigen. Retinoic acid did not induce
expression of desmin or myosin heavy chain. Incubation
of the cell lines with transforming
growth factor fi,
dimethyl sulfoxide, or phorbol esters had no biochemical
or morphological
effed. We conclude that these cell
lines differentiate
to an endothehial lineage in the
presence of retinoic acid.
is a double-stranded
DNA tumor
family. The large (1 00 Kd3), middle
MT promoter
lines developed
form
Abstrad
Polyomavirus
papovavirus
antigens
arise by differential
splicing
(1 ). PVLT is a nuclear
phosphoprotein
that has been classified
as an immortalizing
Transgenic
mice were isolated which harbor
B. Davis-Jewish
General Hospital,
Montreal,
1E2, and Department
of Microbiology
and
Center
Quebec,
and Lorraine
Research
(RXR-a,
RXR-3,
and
RXR-’y),
all
members
of the
steroid/thyroid
receptor superfamily
(30-35),
and CRABPs
(33, 36). TGF-31 can act as a mitogen and growth factor for
various cultured cells or induce differentiation
of other cells
(37,
38).
DMSO
has induced
muscle-specific
gene
expres-
sion in P19 embryo carcinoma
cells (39).
We have isolated immortal
cell lines from transgenic
but
not nontransgen
ic hearts. These transgen ic cells were treated
842
Retinoic
Acid
and
Endothelial
Differentiation
with
transforming
growth
factor
f3, dimethyl
sulfoxide,
phorbol
esters, and retinoic
acid. Treatment
with retinoic
acid induced
a profound
biochemical
and morphological
change
consistent
with terminal
differentiation
to an endothelial
lineage.
b’
;
-
...
Results
Cells
from
MT-PVLT
Line 8
#{149}:
- .s
Transgenic
mice which
harbor the immortalizing
gene from
PVLT linked
to the mouse
MT promoter
were isolated
(5).
Tissue surveys of the transgenic
mice revealed
PVLT expression in heart and testes, but no expression
was found in other
tissues,
including
kidney.
Primary
cells were isolated
from eight transgenic
and five
age-matched
nontransgenic
mice. Hearts and kidneys
from
adult animals,
ranging
in age from 64 to 192 days of age,
were cultured.
Cells from nontransgenic
mice did not survive past the fifth passage, whereas
some cells from all transgenic mice did become
immortal.
No difference
was found
between
the ability
to isolate
cell lines from young
versus
older transgenic
mice.
We isolated
a total of 24 cell lines
from hearts of the transgenic
mice and have characterized
6 of these cell lines. They have been in culture
for greater
than 40 passages
and can now be considered
immortal.
Growth
.
.
Isolation
of Immortal
Transgenic
Mice
Characteristics
of Immortal
Transgenic
cell lines TC1,
Mop-8
(PV-transformed
cell
in media containing
normal
Cells
TC3, TC4, TC5, TC6,
line), and 3T3 cells were
growth
media or growth
containingO.i-fold sera.All ofthecellsgrewwell
serum concentrations
shown).
TC cell lines,
TC1O,
plated
media
independent
of anchorage.
Mop-8
cells formed
large cobnies in soft agar after less than 3 weeks, but, although
the 3T3
and all transgenic
cell lines survived,
they did notform
cobnies. The TC cell lines are thus immortal
but not transformed.
Biochemical
Charaderistics
,
of Immortal
Cells
TC3, TC4, TC5, TC6, TC1 0, MOP-8,
-
.
.
.
.
1C
-f8.i
‘A:’J’
- r’i;’sT
;
-
-
0
I\*1,
“
I
i4’1
:
ftt
.
.
,#{149}
frE
Fig. 1. \/imentin
but not desmin
expression
in immortal
cardiac
cultures.
Transgenic
TC1O cells, 3T3 cells, and G7 myoblasts
were fixed,
incubated
with anti-desmin
(a and b( or anti-vimentin
(dand
e), followed
by fluorescenttagged
second
antibody
and Hoechst
33258
staining.
Desmin
expression
(a
and b): a, G7 myoblasts;
b, transgenic
cells; and c, nuclei
of the transgenic
cells stained with Hoechst
33258.
Vimentin
expression
(dand
e): d, 3T3 cells;
e, transgenic
cells; and 1, nuclei
of the transgenic
cells stained
with Hoechst.
All photographs,
x 200.
in normal
but poorly
in reduced
serum (data not
MOP-8
cells, and 3T3 cells were cul-
tured in agar-containing media to testfor the abilityto grow
Cell IinesTCi
-‘
3T3, and
G7 (fetal-derived
myoblasts)
cells were examined
for vimentin, fibronectin,
actin, desmin,
myosin,
and PVLT using specific antibodies
and indirect
immunofluorescence
(Fig. 1).
All transgenic
cell lines and MOP-8
cells, but not 3T3 or
G7 cells,
demonstrated
nuclear
staining
after incubation
with
rat anti-polyomavirus
sera, indicating
expression
of
PVLT (data not shown).
The presence
of PVLT in all clones
tested and the inability
to isolate
immortal
cell lines from
nontransgenic
mice point to the expression
of PVLT as essential
for immortalization.
Transfection
experiments
found
the transgenic
PVLT to participate
in the DNA replication
of
Pv origin containing
plasmids
(5).
Immortal
cells were isolated
from adult hearts and thus
could
potentially
be of muscle
origin.
Staining
with antibodies specific
for the muscle-specific
markers
desmin
and
myosin
would
establish
these cells as immortal
cardiomyocytes. A lack ofstaining
with desmin
and myosin
but staining
with the non-muscle
markers
fibronectin
or vimentin
would
establish
the immortal
cell lines as of non-muscle
origin.
Fig.
1 , a-f, shows
the results of these experiments
using TCS as
a typical
transgenic
cell line, G7 cells as a positive
control
for desmin
staining,
and 3T3 cells as a positive
control
for
vimentin
staining.
G7 cells (Fig. 1 a) and TC5 cells (Fig. i b)
were stained with anti-desmin
antibody;
the Hoescht
33258
stain of the transgenic
cells is shown
in Fig. 1 c. G7 cells, a
skeletal
myoblast
cell line, show intense
cytoplasmic
staining of desmin
protein,
but no reaction
was seen in TC5 cells
or any other of the TC cell lines. The same result was obtamed
using anti-myosin
sera (data not shown).
3T3 cells
(Fig. ld) and TC5 cells (Fig. le) were stained
with
antivimentin
sera; the Hoechst
33258
stain of the transgenic
cells is shown
in Fig. 1 1. In Fig. 1 e, TC5 cells demonstrate
cytoplasmic
staining
similar
in intensity
and pattern
to the
vimentin
staining
of 313 mouse fibroblasts
shown
in Fig. 1 d.
The same result was obtained
using anti-fibronectin
sera
(data not shown).
Thus, we conclude
from the lack of staining with antibodies
for muscle-specific
markers,
and presence of staining
with antibodies
for fibrobbast-specific
markers, that all of the cell clones
isolated
from the transgenic
hearts were not immortal
cardiomyocytes
but are of a nonmuscle
lineage.
Indudion
of Differentiation
Morphological
Changes. We treated
six transgenic
cell
lines with RA, TGF-/31,
TPA, or DMSO
(Fig. 2; and data not
shown).
In the absence
of an inducer,
the transgenic
cells
displayed
a fibroblast-like
morphology,
but about 5-i 0% of
the cells,
varying
between
the individual
cell lines,
resembled
endothelial
cells.
Continuous
treatment
for up to 4 days with 10-20
ng/ml
TGF-f31 had no effect on cell morphology
but decreased
cell
proliferation
in a dose-dependent
manner.
Neither
TPA
(200-800
ng/mb) nor i-S0!0
DMSO
induced
any morphological
change.
TPA slightly
increased
cell proliferation
compared
to untreated
cultures,
and DMSO
inhibited
cell
proliferation
in a dose-dependent
manner.
Media containing
Cell
Growth
& Differentiation
-0---
-.--
i06
-.--0-.-
SC
0
-.--
CI
C.)
4
Time
Fig. 2.
Growth
of retinoic
acid.
incubated
with
and counted
in
curve
are the
(Days)
curve of transgenic
cells incubated
with increasing
amounts
Transgenic
TC1O cells were plated
at 1 x 10 cells/mI
and
increasing
amounts
of retinoic
acid. Cells were trypsinized
duplicate
every
day for 6 days. Numbers
identifying
each
molar
concentrations
of retinoic
acid.
greater than 5% DMSO
or greater than 1 O’ M RA was lethal
to the cells, but no increased
mortality
was noted with the
other inducers.
Treatment
with 1 0” to 1 o vs RA bed to a dose-dependent
decrease
in cell proliferation
(Fig. 2) and led to a dramatic
morphological
change
in the transgenic
lines regardless
of
the amount
of retinoic
acid. Fig. 3 shows the morphological
changes
found after treatment
of TC5 cells. After treatment
for 3 days, about 80%, and after 4 days, virtually
i 00%,
of
the cells greatly
resembled
terminally
differentiated
endothelial cells. RA-treated
cells, although
originally
evenly dispersed throughoutthe
culture
dish, aligned
to form rows and
lines composed
of numerous
cells (Fig. 3b). In other areas of
the culture
dishes,
some,
but not all, cell lines formed
vesicle-li
ke structures
also composed
of nu merous cells (Fig.
3a). These rows, lines, and vesicle-like
structures
were never
seen in untreated
cultures.
Concomitant
with these morphological
changes
was the loss of vimentin
and fibronectin
expression.
Initially,
90-95%
of the cells were stained
with
vimentin
and fibronectin
antibodies,
but staining
for both
proteins
progressively
decreased
until,
after 4 days of exposureto
RA, only 5-7% ofthe cells expressed
these proteins
(data not shown).
Continued
exposure
to RA was necessary,
as incubation
for a single 24-h period
followed
by a 3-day
period
without
changes.
RA did
of new
not
result
in any
morphological
was required,
as inclusion
of cycboheximide
at i O,
i 0_6, and 1 O M concentrations
in the RA-containing
media
abolished
any morphological
change
Synthesis
protein
(data not shown).
RA treatment
of the mouse fibroblast
PV-expressingfibroblast
cell line MOP-8
cell line 3T3
did not result
or the
in any
morphological
were isolated
change (data not shown). Immortal
cell lines
from the kidneys at the same time and from the
same animals
immortal
cell
as the transgenic
lines from kidney
colonies
cell lines
expressed
from heart. These
PVLT, did not form
in soft agar, and had a fibroblast-like
appearance.
RA treatment
did not induce
any morphological
the kidney
fibroblasts
(data not shown).
The
changes
morphological
changes
to the cells iso-
bated from
induced
by RA were thus restricted
in
mation
oflines.
cell lines after
media.
Changes. RA treatment
induced
profound
changes such that virtually
100% ofthe cells
x 200. Similar
results were obtained
growth
in 1 x 10_6 M or 1 X 10
.i
cells.
4 days.
washed
vesicleculture
the for-
using all ofthetransgenic
retinoic
acid-containing
greatly
resembled
endotheliab
cells. Cells ofendothebiab
and
macrophage
lineages
endocytose
dib-Ac-LDL
(20). Dib-AcLDL was incorporated
by all six transgenic
cell lines tested
(data not shown).
Fig. 4 shows dil-Ac-LDL
incorporation
in
untreated
TC5 cells (a) and RA-treated
TC5 cells (b), with rat
lung primary
endothelial
cells (c) as the positive
control.
Both RA-treated
TC5 cells and the rat endothebiab
cells, but
not the untreated
TC-5 cells, show cytopbasmic
staining
indicating
incorporation
ofthe fluorescent-tagged
LDL. Endothelial
cells, but not macrophage
cells, synthesize
the anticoagulant
factor,
Factor VIII-related
antigen
(21). All RAtreated
but not untreated
TC lines synthesized
Factor VIII-
related antigen.
Fig. 4 shows Factor VIll-related
antigen
expression
in untreated
TC-5 cells (d) and RA-treated
TCS
cells (e), with rat lung primary
endothelial
cells ( I) as the
positive
control.
In both RA-treated
TC5 cells and rat endothelial cells, but not the untreated
TC5 cells, the presence
of
cytosobic
staining
indicates
synthesis
of Factor
VIII-related
antigen.
heart.
Biochemical
morphological
Fig. 3.
Morphological
effect of retinoic
acid treatment
on transgenic
Transgenic
TC1 0 cells were incubated
with 1 X 1 o
M retinoic
acid for
a, incorporation
of dil-Ac-LDL
proceeded
for 4 h; then the excess was
away, and the cells were fixed. The retinoic
acid-treated
cells formed
like structures
evident
at X 1 00 magnification.
b, retinoic
acid-treated
was stained
with
Hoechst
33258.
The stained
nuclei
demonstrate
Initially,
Factor
VIII-related
antigen
by less than iO% of cells, but expression
25% of cells after 1 day, 40-45%
of cells
was expressed
increased
to 20after 2 days, 75-
843
844
Retinoic
Acid
r
and
Endothelial
Differentiation
‘
I
.
.
.4
.
.i
Fig. 4.
Retinoic
acid-treated
but
not untreated
transgenic
cells incorporate
dil-Ac-LDL
and synthesize Factor
VIll-related
antigen.
Transgenic
TC1 0 cells were grown
in the absence
or presence
of 1 x
10-i M retinoic
acid, and incorporation
of fluorescent-tagged
LDL
(a-c) and synthesis
of Factor VIIIrelated
antigen
(d-((
were exammed. Incorporation
of dil-Ac-LDL
was examined
after a 4-h incubation. a, TC cells incubated
without
retinoic acid; b, TC cells incubated
in the presence
of retinoic
acid;
and c, rat primary
endothelial
cells
4.’
;4
.‘.(
&
t
‘
..4:.
.
c3’
t
L’
&.‘c
p.
as the positive
.
control.
Synthesis
of
Factor VIII-related
antigen
was examined
after fixation
of the cells,
incubation
with
anti-Factor
VIIIrelated
antigen
antibody,
and reaction
with
a fluorescent-tagged
second
antibody.
d, TC1O cells
grown
without
retinoic
acid;
e,
TC1O
cells
grown
with
retinoic
acid; and (, rat primary
endothelial
cells as a positive
control.
All panels, x 400. Similar
results were obtamed
using all of the transgenic
cell lines after growth
in 1 x 10_6
M or
1 X 10-p M retinoic
acidcontaining
80%
after 3 days, and 90-95%
of the cells after 4 days of RA
treatment
(data not shown).
Thus, RA-treated
cells progressively
lost vimentin
and fibronectin
staining,
while
gaining
Factor VIII-related
antigen
staining
(data not shown).
Transgenic cells treated with TGF-j31,
TPA, or DMSO
were negative for dil-Ac-LDL
incorporation
and Factor VIll-related
antigen synthesis
(data not shown).
Thus, morphologically
and
biochemically,
RA-treated
cells show the characteristics
of
terminally
differentiated
endothelial
cells.
None of the RA-, TGF-f31-,
TPA-, or DMSO-treated
cell
lines expressed
myosin
or desmin.
Discussion
The inclusion
of RA in the growth
media
had a profound
biochemical
and morphological
effect on all ofthe cell lines
isolated
from the hearts ofthe transgenic
mice. RA treatment
had no effect on 313 cells, MOP-8
cells, or immortal
cells
isolated
from the kidneys
of the transgenic
mice. The RAtreated
transgenic
cells progressively
lost vimentin
and fibronectin
expression,
but increasing
numbers
expressed
Factor
VIII-related
antigen
and incorporated
dil-Ac-LDL.
Both
incorporation
of LDL and synthesis
of Factor
VIIIrelated
antigen
are hallmark
features
of fully differentiated
endothelial
cells (40, 41 ). In addition,
the RA-treated
cells
media.
aligned
to form multicellular
rows and lines, and some, but
not all, cell lines developed
vesicle-like
structures
reminiscent of “angiogenesis
in vitro.”
PVLT
expression
does not
confer
sensitivity
to RA-mediated
differentiation,
as MOP-8
cells, which
express
all three PV early antigens,
or kidney
fibroblasts
from
transgenic
mice
which
express
PVLT
showed
no effect upon RA treatment.
Endothebial
cells from
different
organs or tissue locations
are not uniform
in their
molecular
marker
presentation,
although
they share common properties.
Unique
cell surface
profiles
in endothelial
cells from different
tissues are in part thought
to determine
the arrest and invasion
patterns
of tumor
cells and form an
important
aspect of the “seed and soil” hypothesis
of Paget
(1 3). Thus,
the endothelial-bike
cells isolated
from the transgenic
hearts may be part of a distinct
population
not represented
in the endothelial
cell compartment
present
in
transgenic
kidney.
This heterogeneity
ofthe
endothebial
cell
would
explain
the ability
of RA to mediate
differentiation
in
cells isolated
from heart but not from kidney.
On the other
hand, there is no evidence
to indicate
that the cells from
kidney
originated
from
endothelial
cells.
RA-induced
changes
are not a property
of mouse fibroblasts,
because
3T3
cells incubated
with RA did not show any morphological
or
biochemical
changes.
Cell Growth
We initially
characterized
the transgenic
cells
as
fibroblast-like
because of their morphology
and their fibronectin and vimentin
expression.
In general, differentiated
endothelial
cells do not express fibronectin
(42) but can express vimentin
(43). However, the presence of PVLT staining
in the intact cardiac vasculature
of the transgenic
mice, the
presence of 5-1 0% of cells with an endothehial
cell morphology
in untreated
cultures, the induction
of endothelial
cell morphology
and biochemical
markers, and the loss of
fibroblast
markers after RA treatment
led us to term the immortal cells “endothehial-hike
cells.” As the precursors
to
endothelial
cells are relatively
uncharacterized,
and in the
absence of reliable markers, it is currently
premature to label
the cells as truly endothehial
precursor cells. We do not think
dedifferentiation
of mature endothehial
cells occurred due to
in vitro culture conditions,
because control experiments
did
not isolate immortal
cell lines from the nontransgenic
mice.
It is more likely thatthe transgenic cells represent endothelial
cells dedifferentiated
in vitro because of immortalization
by
PVLT. They are thus endothelial-hike
cells which, although
they do not express endothehial
cell markers, are committed
to an endothehial
cell lineage. Retinoic acid treatment
tiates the differentiation
of these committed
cells to cells
capable of such endothehial
specific activities as LDL incorporation
and Factor VIII-related
antigen expression.
Treatment with other inducers, such as TGF-31, DMSO, or TPA,
did not cause differentiation
to endothehial
cells or indeed to
any other cell type.
In chicks upto Stage 1 5, the heart is composed
of only two
cell types: myocytes
and endothehial
cells (44). Retroviral
tagging ofchick embryos demonstrated
that endothehial
cells
forming
coronary
vessel walls arose from in situ vasculogenesis and not from the root ofthe aorta (45). In this model,
numerous
endothelial
cells must originate
vessel formation
from multiple
locations.
These data perhaps point to an Undifferentiated
endothehial
precursor
cell as responsible
for
the initial population
of the growing vessel. Terminal differentiation
of the precursor
endothelial
cell would then be
accomplished
later. It was also demonstrated
that endothehal cells did not derive from the same precursor
as cardiomyocytes,
vascular smooth muscle cells, or perivascular
fibroblasts. The cells isolated from the transgenic
heart may
represent
this precommitted,
but as yet undifferentiated,
endothehial
cell. In this respect, it may be significant
that the
cell lines were isolated from hearts which were undergoing
hypertrophy
and perhaps expanding
and remodeling
their
coronary
vasculature
to accommodate
the enlarging size of
the heart.
If retinoic acid in vivo affects the differentiation
of endothelial cells as in this in vitro study, then retinoids may have
a role in vascular development
and angiogenesis.
The presence of retinoids
is necessary for normal development:
vitamin A-depleted
chick embryos
lacked a developing
extraembryonic
circulation
system and were nonviable
(46).
Retinoid
supplementation
has teratogenic
effects on heart
development
in mammals
(44, 47-49).
The teratogenic
effect is perhaps related to RA-mediated
disruption
of interactions between precardiac
cells and the cell matrix and to
inhibition
of cell migration
during formation
of the heart.
Retinol treatment
of bovine endothelial
cells in vitro was
shown to induce a change in the shape of the treated cells
so they occupied
a smaller cellular
area than nontreated
cells, and similar to the results presented here, cellular proliferation was inhibited
(SO). In contrast, another study found
that RA treatment
stimulated
the growth of bovine endothe-
mi-
& Differentiation
845
hal cells (51). RA interacts with nuclear and cytoplasmic
receptors.
RXR-’y is highly expressed
in heart and functions
to up-regulate
CRABP-2 through a specific interaction
with
its gene
promoter
(35).
CRABP-2
is induced
as much
as 50-
fold upon RA treatment
of F9 teratocarcinoma
cells (33).
There are no data on the presence or absence of RARs, RXR,
or CRABPs in endothehial
cells.
In this study, it proved impossible
to culture cardiomyocytes which retain cardiomyocyte-specific
gene expression
from the transgenic
mice, even though PVLT was expressed
in cardiomyocytes.
None of the 24 cell lines isolated from
a total of 8 transgenic
mice displayed
any myosin or desmin
reactivity,
nor
any
contractility.
Cardiomyocytes
formed
a
significant
proportion
of the total cells in our primary cubtures; yet, similar to previous studies, these cardiomyocytes
were not capable of continued
proliferation.
Transfection
of
first and second passage primary
heart cultures with DNA
containing
a polyomavirus-specific
origin of DNA replication indicated
that transgenic
PVLT was present and active
in initiation
of polyomavirus-specific
DNA
replication
(5).
As more PVLT is required for DNA replication
than immortalization,
sufficient
quantities
of PVLT to immortalize
cells
were present in the cardiomyocytes
(52). Incubation
of cardiac fibroblasts
isolated from rabbit heart with TGF-31 led to
the induction
ofthe myocyte-specific
marker sarcomeric
actin and loss of the fibroblast
marker vimentin
(11). In our
studies, similar and even higher concentrations
of TGF-31
included
in the media had no effect on muscle-specific
gene
expression,
nor did they induce any morphological
change.
In conclusion,
we have
isolated
immortal
endothehial-like
cells from hearts oftransgenic
mice. These cell lines express
vimentin
and fibronectin,
but not myosin or desmin.
Upon
incubation
with retinoic acid, they underwent
morphological and biochemical
changes consistent
with terminal
differentiation
to endothehial
cells.
Materials
and Methods
Cell Lines. MT-PVLT line 8 transgenic
mice were isolated
and characterized
previously
(5). Rat-i cells transformed
with polyomavirus
(Rat-i -PV) were described
previously
(53). The mouse fibroblast
line 3T3 (CCL 92), MOP-8 cells
(CRL 1 709), which express polyomavirus
large, middle, and
small T antigens, and G7 (CRL 1447), which are fetal-derived
muscle myoblast
cells, were purchased
from the American
Type Culture Collection.
Antibodies.
Mouse monocbonal
anti-desmi n, anti-actin,
anti-vimentin,
and fluorescein
isothiocyanate
or rhodamineconjugated
anti-mouse
or anti-rabbit
sera were purchased
from ICN Biomedicals.
Goat antiserum to human Factor VIIIrelated
antigen
was purchased
from
ICN Biomedicals.
Mouse monoclonal
anti-human
fibronectin
clone FN-l 5
was obtained
from Sigma. Mouse monoclonal
anti-myosin
MF-20 cells and rat anti-polyomavirus
antisera were gifts
from Dr. Michael
Rudnicki
and Dr. John Hassell, McMaster
University,
Hamilton,
Ontario,
Canada,
respectively.
Primary rat lung endothelial
cells were a gift from Dr. David
Langleben,
Lady Davis Institute. Dil-Ac-LDL
was purchased
from ICN Biomedicals.
Inducers. TGF-f31 was purchased
from Austral Biobogicals, San Ramon, CA. RA, TPA, and cell culture grade DMSO
were purchased
from Sigma. Cell culture media, enzymes,
antibiotics,
and general chemicals
were purchased
from
Sigma. Fetal bovine sera and HS were purchased
from ICN
Biomedicals
or lmmunocorp,
Montreal,
Quebec,
Canada.
846
Retinoic
Acid
and Endothelial
Differentiation
Isolation of Immortal Cells. Transgenic and age-matched
nontransgenic
animals were sacrificed
by cervical dislocation, and the hearts were removed.
Hearts were rinsed to
remove blood in ice-cold basal media (DMEM plus 50 ig/ml
gentamicin
and 2.5 ig/ml
fungizone)
and minced.
The
pieces were placed in digestion
media 1 (basal media plus
50 pg/mI trypsin and 20 units/mI collagenase),
and digestion
was allowed to proceed for 3 h at 37#{176}C.
The digested pieces
were gently pipeted up and down to free individual
cells,
washed,
resuspended
in growth media 1 (basal media plus
1 0% HS), and plated
on gelatin-coated
dishes.
Cultures
were
passed 1 :2 every 7-1 0 days until nofurther
growth occurred.
Kidneys from the same animals were minced and then
placed in digestion media 2 (50% DMEM and 50% verasine
plus 2.4 mg/mI trypsin and antibiotics)
and incubated
overnight at 4#{176}C.
After a 30-mm incubation
at 37#{176}C,
the cells
were washed free of digestion enzymes and resuspended
in
growth media 2 (DMEM plus 5% fetal bovine sera plus antibiotics)
and plated onto dishes.
Individual
areas of continued
cell growth were collected
on trypsmn-soaked
papers and plated in single wells of 24well dishes. As these cultures continued
to grow, the cells
were transferred
to increasingly
larger Petri dishes. A total of
24 clones were isolated from 8 transgenic
heart samples.
Analysis of Growth Charaderistics.
Cells, i X 1 04/mI,
were plated in 24-well
dishes in media containing
either
1 0% HS or 1 % HS. Cellsfrom
by trypsinization,
mocytometer.
washed,
Transgenic
duplicate
and counted
cells,
3T3 cells,
wells
were
daily
collected
using
a he-
and polyomavirus-
transformed
Rat-i cells, at 1 x l0,
1 x i0, and 1 x i0
cells/culture,
were placed in media containing
0.3% agar
and plated over a layer of media containing
0.6% agar. The
cultures were examined
every 3-4 days for 3-4 weeks for
growth of individual
colonies.
Cultures were layered with
fresh media containing
agar once per week. MOP-8 cells
served as a positive control, and 3T3 cells served as a negative control.
Immunofluorescence
Staining. Cells were grown
on
LabTek slides, washed
with PBS, and fixed in acetone:
methanol
(2:1 ) at -20#{176}C
for 10 mm or fixed in 3.7% formaldehyde
in PBS for 5 mm followed
by washing
in PBS.
Fixed cells were hydrated
in PBS, and indirect
immunofluorescence
was performed
by standard
techniques
(54).
Rabbit anti-human
Factor VIlI-related
antigen
sera (1 :4),
mouse anti-vimentin
(1 :75), mouse anti-fibronectin
(1 :100),
mouse anti-desmin
(1 :1 0), and mouse anti-actin
(1 :1 0) were
diluted in PBS, and mouse MF-20 medium was used undiluted. Antibodies
were incubated
at room temperature
for 1
h prior
to incubation
with
the
secondary
antibodies.
After
staining with antibodies, the cellswere dipped for 1 mm in
Hoechst 33258 (1 0 pg/mI PBS), washed,
and examined
with
an Olympus
BH2 microscope
equipped
with epifluorescence using a Pboempak filter for Hoechst stain, a Filter 12
for fluorescein,
and a Filter N2 for rhodamine,
or with a
Bio-Rad MRC-600
confocal
microscope.
Cells were incubated
with i 0 jg/mI dil-Ac-LDL
for 4 h at
37#{176}C
in normal media. The cells were then washed free of
unendocytosed
material,
fixed
hyde, and examined
as above.
Indudion
of Differentiation
ethanol
and
added
to the growth
in 3.7%
buffered
RA was dissolved
media
formalde-
in 95%
at concentrations
varying from i 0 to 1 0
M. TGF-/31
was dissolved
in 95%
ethanol and added to the growth media at concentrations
varying from 10 to 20 ng/ml. TPA was dissolved
in DMSO
and added to the growth media at concentrations
varying
from
200
to 800
ng/mI.
Cell
culture
grade
DM50
was
in-
cluded in the culture media at concentrations
varying from
1 % to 5%. Control
cultures were treated with equivalent
volumes of 95% ethanol or DMSO. Cells were examined
for
morphological
or biochemical
changes 1-4 days after treatment.
Acknowledgments
We thank the members ofour
We gratefully
acknowledge
laboratory
for generous
the excellent
technical
and helpful comments.
assistance
of Tracey
Dorrance.
References
J. DNATumor
1 . Tooze,
Laboratory
Press,
Viruses.
Cold Spring Harbor,
NY: Cold Spring Harbor
1981.
2. Asselin, C., and Bastin, M. Sequences from polyomavirus
and simian virus
40 large T genes capable of immortalizing
primary rat embryo fibroblasts.
I.
Virol.,
56:958-968,
3. Cuzin,
distinct
1985.
F. The polyomavirus
proteins
Biophys.
Acta,
4. Glaichenhaus,
N., Mougneau,
Cuzin,
F. Co-operation
blasts:
an experimental
291-315,
oncogenes:
in the transformation
781: 193-204,
1984.
co-ordinated
of rodent
E., Connan,
between
multiple
model
of tumor
cells
function
of three
in culture.
Biochim.
G., Rassoulzadegan,
oncogenes
in rodent
progression.
M., and
embryo
Adv. Cancer
fibro-
Res., 45:
1985.
5. Chalifour,
L. E., Gomes, M. L., Wang, N-S., and Mes-Masson,
omavirus
large T-antigen expression
in heart of transgenic
mice
diomyopathy.
Oncogene,
5: 1 71 9-1 726, 1990.
A-M. Polycauses car-
6. Chalifour,
L. E., Mes-Masson,
A-M., Gomes, M. L., and Wang, N-S. Testicular adenoma
and seminal vesicle engorgement
in polyomavirus
large
T-antigen transgenic
mice. Mol. Carcinog.,
5: 187-189,
1992.
7.
Zak,
R. Growth
of the
Heart
in Health
and
Disease.
New
York:
Raven,
1984.
8. Morgan, H. E., Gordon, E. E., Kira, Y., Chua, B. H. L., Russo, L. A., Peterson,
C. J., McDermott,
P. J., and Watson, P. A. Biochemical
mechanisms
of cardiac
hypertrophy.
Annu. Rev. Physiol., 49: 533-543,
1987.
9. Piper, H. M., Probst,
In: A. Pinson
10.
ley,
I., Schwartz,
(ed), The Heart
P., Pahr, R. S., and Spieckermann,
P. G.
Boca Raton, FL: CRC Press, 1987.
Cell in Culture.
Hulland,
T. J. ln:J. E. Moulton
CA: University
of California
led),
Press,
Tumors
1978.
in Domestic
Animals.
Berke-
11 . Eghbali, M., Tomek, R., Woods, C., and Bhambi,
B. Cardiac fibroblasts
are predisposed
to convert into myocyte phenotype:
specific effect of transforming growth factor . Proc. NatI. Acad. Sci. USA, 88: 795-799,
1991.
1 2.
Rupnick,
M. A.,
cultured
cerebral
435-444,
1988.
1 3.
Pauli,
Carey,
A.,
B. U., Augustin-Voss,
Hammer,
adhesion
William,
S. K. Phenotypic
diversion
cells. In VitroCell
H. G., El-Sabban,
15. Folkman,
cell embryology
M. E., Johnson,
J., Haundenchild,
endothelial
cells.
16. Gitlin, J., and D’Amore,
growth media. Microvasc.
and growth.
C., and Zetter,
Proc.
NatI.
Acad.
in
Dev. Biol., 24:
R. C., and
D. A. Organ preference
of metastasis: the role of endothelial
molecules.
Cancer Metastasis
Rev., 9: 1 75-189,
1990.
14. Wagner, R. C. Endothelial
9:45-75,
1980.
capillary
and
microvascularendothelial
cell
Adv. Microcirc.,
B. R. Long term culture of
76:521
7-5221
, 1979.
Sci. USA,
P. Culture of retinal capillary
Res., 26: 74-80,
1983.
cells using selective
1 7. Davison, P. M., and Karasek, M. A. Human dermal microvascular
endothelial cells in vitro: effect of cyclic AMP on cellular
morphology
and proliferation
rate. J. Cell. Physiol.,
106: 253-258,
1980.
1 8.
Bowman,
P. D., Bertz,
R. R., and Goldstein,
InVitro,
17:353-362,
A. L., Ar, D., Wolinsky,
G. W. Primary
culture
J. S., Penny,
of endothelium
J. B., Shivers,
from
rat brain.
1981.
19. Diglio, C., Grammas,
P., Giaccamelli,
of rat cerebral microvascular
endothelial
1982.
F., and Weiner, J. Primary culture
cells. Lab. Invest., 46: 554-563,
20. lngber, D. E., and Folkman,
J. Mechano-chemical
switching
between
growth and differentiation
during fibroblast
growth factor-stimulated
angiogenesis in vitro: role of the extracellular
matrix. J. Cell. Biol., 109:317-330,
1989.
.
DeBault, L., Henriquez,
E., Hart, M., and Cancilla,
P. Cerebral microvessels and derived cells in tissue culture. In Vitro, 17: 480-494,
1981.
21
22.
Dubois,
V. L. Isolation
N. A.,
Kolpack,
L. C., Wang,
and characterization
R., Azizkhan,
of an established
R. G.,
endothelial
and
Bautch,
cell line,
Cell Growth
from transgenic
mice
hemangiomas.
Exp. Cell Res., 196: 302-31
3, 1991.
38. Massagu#{233},J. The transforming
23. Al Moustafa,
A-E., Quatannens,
B., Dieterlen-Lievre,
F., and Saule, S.
Tumorigenic
effect mediated in the avian embryo by one or more oncogenes
associated
24.
with
Ressetar,
v-myc.
Oncogene,
H. G., Webster,
cells express JC virus
cyclophosphamide-treated
8174, 1992.
25. Harder,
R., Uhlig,
C.,
H.
Thiele,
endothelial
thelial
cell
cell
lines:
dependent
26.
H., and
and
Stoner,
1992.
G. L. Brain
vascular
endothelial
6:597-641,
A., Schutt,
Hamann,
interaction
A.
mechanism
mechanisms
of
A., Butcher,
murine
by SV4O-transformed
basal
mouse
adhesion.
Exp. Cell
and their
receptors
in differentiation,
FASEB
mediating
J., 5: 2924-2933,
E.
lymphocyte-
binding,
Res., 197:
L. G. Retinoids
neoplasia.
B., Duijvestin,
Dissection
endo-
and a4-integrin259-267,
1991.
embryo-
factor-a
family.
Annu.
Rev. Cell
1990.
M. K. S., and
in a cultured
40.
cells
Stein,
0.,
and
macrophage-like
Stein,
Y.
properties
Biochim.
Bovine
towards
Biophys.
aortic
acetylated
Acta,
620:
1982.
endothelial
El-i 251-labelled
631-635,
display
low density
1980.
41 . Belloni,
P. N., and Tressler, R. J. Microvascular
endothelial
cell heterogeneity:
interactions
with leukocytes
and tumor cells. Cancer
Metastasis
Rev.,
8:353-389,
42.
1990.
Robinson,
clonal
brain
1991.
growth
847
39. McBurney,
M. W., Jones-Villeneuve,
E. M. V., Edwards,
Anderson,
P. J. Control of muscle and neuronal
differentiation
embryonal
carcinoma
cell line. Nature (Lond.), 299: 165-167,
lipoprotein.
H., Kashani,
and
novel
670,
large tumour
antigen
in immunocompetent
and
hamsters. Proc. NatI. Acad. Sci. USA, 89:8170-
cytokine-induced
De Luca,
genesis
G.,
7: 1 667-1
Biol.,
& Differentiation
A. P., White,
antibody
which
capillaries.
T. M., and
Mason,
D. W. MRC
reacts with
all vascular
57: 231-237,
Immunology,
Ox-43:
endothelium
1986.
a mono-
in the rat except
27. Merriman,
R., and Bertman,
J. Reversible
inhibition
by retinoids
of
3-methylcholanthrene-induced
neoplastic
transformation
in C3H/1 OT-clone
8 cells. Cancer Res., 39: 1661-1666,
1979.
43. Abbott, N. J., Hughes, C. C. W., Revest, P. A., and Greenwood,
J. Development
and characterization
of a rat brain capillary
endothelial
culture:
towards an in vitro blood-brain
barrier.
J. Cell Sci., 103: 23-37,
1992.
28. Lotan,
R. Effects of vitamin
neoplastic
cells.
44. Osmond,
M. K., Burler, A. J., Voon, F. C. T., and Bellairs,
of retinoic acid on heart formation
in the early chick embryo.
(Camb.),
1 13: 1 405-1 41 7, 1991.
29. Talmage,
Biochim.
A and its analogs (retinoids)
on normal
Acta, 605: 33-91,
1980.
D. A., and Lackey,
R. S. Retinoic
c-fos
acid receptor
polyomavirus
transformation
gene, 7: 1837-1845,
1992.
and
30. Benbrook,
D., Lernhardt,
E., and Pfahl, M. A new retinoic
identified
1988.
a hepatocellular
from
expression
carcinoma.
a suppresses
in rat fibroblasts.
Nature
(Lond.),
333:
669-672,
45. Mikawa,
T., and
genesis:
discontinuous
USA, 89:9504-9508,
46. Thompson,
32. Krust, A., Kastner,
human
retinoic
acid
P., Petkovich,
receptor,
M., Zelent,
hRAR-y.
Proc.
NatI.
P. Cloning
y predomi-
Sci. USA,
P. A third
86:5310-
33. Giguere, V., Lyn, S., Yip, P., Siu, C-H., and Amin, S. Molecular
cloning
ofcDNAencoding
a second cellular retinoic acid-binding
protein. Proc. NatI.
Acad. Sci. USA, 87: 6233-6237,
1990.
A., Tiollais,
sion
and
ofthe
J.,
8:429-433,
ligand
regulation
retinoic
P., and Dejean,
acid
receptor
A. Differential
a and
f3 genes.
J. N., Howell,
activity
A deficiency
formations
expresEMBO
49.
Lammer,
J. T., Curry,
35. Mangelsdorf,
D. J., Borgmeyer,
U., Heyman,
R. A., Zhou, J. Y., Ong, E.
S., Oro, A. E., Kakizuks, A., and Evans, R. M. Characterization
of three RXR
genes that mediate the action of 9-cis retinoic acid. Genes & Dev., 6: 329344, 1992.
E. J., Chen,
C. J., Fernhoff,
50.
Braunhut,
52.
Ogris,
large
cogene,
53. Mes-Masson,
velopment
organogenesis.
De-
1990.
37. Zentella, A., and Massague, J. Transforming
growth factor
induces myoblast differentiation
in the presence of mitogens.
Proc. NatI. Acad. Sci. USA,
89:5176-5180,
1992.
embryo.
Br. J. Nutr.,
C. I. The
fowl and effects of vitamin
23: 471-490,
1969.
acid-induced
heart
D. T., Hoar, R. M., Agnish,
P. M.,
Grix,
acid embryopathy.
A. W.,
Lott,
8: 1277-1283,
mal-
1981.
N. D., Benke,
cushion
P. J., Braun,
I. T., Richard,
N. EngI. J. Med.,
313:
J. M.,
837-841
M. Modulation
of endothelial
Res., 41: 47-62,
1991.
E., Mudrak,
I., and Winterberger,
E. Distinct
T-antigen
are required
for different
functions
their
mouse
in the domestic
51 . Melnykovych,
M., and Clowes, K. K. Growth stimulation
thelial cells by vitamin A. J. Cell. Physiol.,
109: 265-270,
virus
during
acid
S. J., and Palomares,
by retinoids.
Microvasc.
36. Dolle, P., Rubert, E., Leroy, P., Morriss-Kay,
G., and Chambon,
P. Retinoic
acid receptors and cellular retinoid binding protein. 1 . A systematic study of
pattern
of transcription
(Camb.),
110: 1133-1151,
J. M., Pitt, G. S. J., and McLaughlin,
of retinoic
on the chick
D. A. Retroviral
analysis
of cardiac
morphoof coronary
vessels.
Proc. NatI. Acad.
Sci.
I. M., Wiley,
M. J., and Agur, A. Retinoic
in the hamster.
Teratology,
21: 193-197,
Sun, S. C. Retinoic
and growth
1989.
differential
Fischman,
formation
1992.
48. Davis, L. A., and Sadler, T. W. Effects of vitamin A on endocardial
development
in the mouse heart. Teratology,
24: 1 39-1 48, 1981.
1989.
34. de The, H., Marchio,
biological
47. Taylor,
A., and Chambon,
Acad.
R. The effects
Development
Onco-
acid receptor
31 . Zelent, A., Krust, A., Petkovich,
M., Kastner, P., and Chambon,
of murine a and f3 retinoic acid receptors and a novel receptor
nantly expressed in skin. Nature
(Lond.), 339: 714-71 7, 1989.
5314,
and
Biophys.
,
cell shape
ofbovine
1981.
amounts
of the
and
1985.
endo-
of polyomaprotein.
On-
1993.
A. M., Masson, S., tsanville,
D., and Chalifour,
L. E. Expression
of oncomodulin
does not lead to the transformation
or immortalization of mammalian
cells in vitro. J. Cell Sci., 94: 517-525,
1989.
54.
Freshney,
New York: Alan
R. I. Culture
of Animal
R. Liss, 1987.
Cells:
A Manual
of Basic
Technique.