.JPG?ad=18503&adview=true)
[Cancer Research 60, 5862-5869, October 15, 2000]
© 2000 American Association for Cancer Research
B16 Melanoma Cell Arrest in the Mouse Liver Induces Nitric Oxide Release and Sinusoidal Cytotoxicity: A Natural Hepatic Defense against Metastasis1
Hui Helen Wang,
Alan R. McIntosh,
Brian B. Hasinoff,
Edward S. Rector,
Naeem Ahmed,
Dwight M. Nance and
F. William Orr2
Department of Pathology, Faculty of Medicine, University of Manitoba, Winnipeg, Manitoba, Canada R3E OW3 [H. H. W., E. S. R., N. A., D. M. N., F. W. O.], and Faculty of Pharmacy, University of Manitoba, Winnipeg, Manitoba, Canada R3T 2N2 [A. R. M., B. B. H.]
 |
ABSTRACT
|
|---|
The formation of liver metastases involves interactions between
intravascular cancer cells and the hepatic microvasculature. Here we
provide evidence that the arrest of intravascular B16F1 melanoma cells
in the liver induces a rapid local release of nitric oxide (NO) that
causes apoptosis of the melanoma cells and inhibits their subsequent
development into hepatic metastases. B16F1 melanoma cells (5 x 105) labeled with fluorescent microspheres were
injected into the portal circulation of C57BL/6 mice. The production of
NO in vivo was detected by electron paramagnetic
resonance spectroscopy ex vivo using an exogenous
NO-trapping agent. A burst of NO was observed in liver samples examined
immediately after tumor cell injection. The relative electron
paramagnetic resonance signal intensity was 667 ± 143
units in mice injected with tumor cells versus
28 ± 5 units after saline injection
(P < 0.001). Two-thirds of cells
arrested in the sinusoids compared with the terminal portal venules
(TPVs). By double labeling of B16F1 cells with fluorescent microspheres
and a TdT-mediated UTP end labeling assay, we determined that the
melanoma cells underwent apoptosis from 4–24 h after arrest. The mean
rate of apoptosis was 2-fold greater in the sinusoids than in the TPVs
at 4, 8, and 24 h after injection (P < 0.05–0.01). Apoptotic cells accounted for 15.9 ± 0.8% of tumor cells located in the sinusoids and 7.1 ± 0.9% of tumor cells in the TPVs. The NO synthase inhibitor
NG-nitro-L-arginine methyl ester completely
blocked the NO burst (P < 0.001) and
inhibited the apoptosis of B16F1 cells in the sinusoids by 77%.
However, the rate of tumor cell apoptosis in the TPVs was not changed.
There were 5-fold more metastatic nodules in the livers of
NG-nitro-L-arginine methyl ester-treated mice
(P < 0.05). The inactive enantiomer
NG-nitro-D-arginine methyl ester had no effect
on the initial NO burst or on apoptosis of tumor cells in
vivo. Both annexin V phosphatidylserine plasma membrane
labeling and DNA end labeling of apoptotic cells were demonstrated
after a 5-min exposure (a time equivalent to the initial transient NO
induction in vivo) of B16F1 cells to a NO donor
in vitro. These results identify the existence of a
natural defense mechanism against cancer metastasis whereby the arrest
of tumor cells in the liver induces endogenous NO release, leading to
sinusoidal tumor cell killing and reduced hepatic metastasis formation.
|
INTRODUCTION
|
|---|
The formation of liver metastases involves interactions
between intravascular cancer cells and the hepatic microvasculature
(1)
. Hepatic microvascular functions of potential
importance in tumor cell-endothelial interactions include the
expression of adhesion molecules and release of reactive oxygen species
(e.g.,
NO,3
O
, and H2O2) under
physiological and pathological conditions (1, 2, 3)
. Evidence
from in vitro studies has shown that reactive oxygen and
nitrogen species may be cytotoxic to neoplastic cells that are adherent
to the hepatic sinusoidal endothelium derived from ischemic or
IL-1-stimulated mouse livers (4
, 5)
. However, little
in vivo data are available on the involvement of
endothelium-derived NO in the mechanism of hepatic metastasis
(6)
.
Using a murine model of liver metastasis, our previous work has shown
that intravascular B16F1 melanoma cells are trapped within the
sinusoids by size restriction in control mice, whereas they arrest in
the TPV region in mice that have been pretreated with IL-1
. Even so,
early metastases developed in the portal tract areas in both control
and IL-1-prestimulated groups (7)
. This selective
growth of liver metastases in acinar zone 1 has also been demonstrated
in rats and mice by others (8)
. Early development of
metastases has not been observed in the sinusoidal region (7
, 8)
, suggesting that the sinusoids constitute a cytotoxic zone to
cancer cells. Because microvascular NO can be toxic to cancer cells
(9, 10, 11)
and reduces their adhesion to post-capillary
venules (12)
, we have tested the hypothesis that the
hepatic microvasculature regulates the arrest and fate of metastasizing
cancer cells in vivo through release of NO. To our
knowledge, this is the first demonstration that NO is released from the
liver after arrest of tumor cells and that NO has a direct effect on
sinusoidal killing of cancer cells and their metastatic outcome. The
present study provides further evidence to support the development of
metastases in the portal tract areas, with sinusoids being a more
cytotoxic zone to tumor cells.
|
MATERIALS AND METHODS
|
|---|
Animals and Materials
Female C57BL/6 mice weighing 20–22 grams were purchased from
Charles River (Montreal, Quebec, Canada) and housed according to the
University of Manitoba guidelines. The murine B16F1 melanoma cell line
was obtained from the American Type Culture Collection (Manassas, VA).
Fluoresbrite carboxylate YG 0.05-µm microspheres were purchased from
Polysciences, Inc. (Warrington, PA). -MEM, Opti-MEM,
penicillin/streptomycin, and trypsin-EDTA were purchased from Life
Technologies, Inc. (Burlington, Ontario, Canada). Avertin
(2,2,2-tribromoethanol) was from Caledon Laboratories Ltd. (Georgetown,
Ontario, Canada). DETC was purchased from Sigma-Aldrich Canada Ltd.
(Oakville, Ontario, Canada). L-NAME, D-NAME, and SNAP were purchased
from Sigma. The ApopTag Peroxidase Kit was from Intergen Company
(Purchase, NY). Fluorescent polystyrene microspheres (15 µm in
diameter) were kindly provided by Dr. D. Goetz (University of
Ohio).
Experimental Design
In Vivo Experiments.
Fluorescent microsphere-labeled B16F1 melanoma cells were injected into
the mesenteric veins of C57BL/6 mice, and the livers were removed
between 0 and 24 h after injection. Data for each observation were
obtained from three separate animals. In accordance with principles for
the ethical use of animals, we used the smallest number required to
obtain statistically meaningful data. Because data from three mice
generally showed a clear statistical distribution, no more animals were
added for most groups. Each liver sample was divided for the
measurement of NO production, cell localization, and tumor cell
apoptosis. For metastasis studies, unlabeled B16F1 cells were injected,
and the livers were collected between days 3 and 7 (animal survival
time) for surface metastatic nodule and histomorphometric analysis. A
total of 83 mice (3–10 mice/condition) were used in the study.
In Vitro Experiments.
In vitro cultured B16F1 melanoma cells were exposed to NO
donor SNAP, and apoptosis was measured. All analyses were performed
blindly.
|
Cell Culture and Fluorescent Labeling
|
|---|
B16F1 cells were cultured in -MEM supplemented with 10%
fetal bovine serum and 1% penicillin/streptomycin in a T75 culture
flask (Corning) to 80% confluence at 37°C and 5%
CO2. The culture medium was then aspirated.
Opti-MEM serum-reduced medium and fluoresbrite carboxylate microspheres
were added to the flask in a ratio of 150 µl of microspheres:7.5 ml
of Opti-MEM (microsphere stock, 2.5% solids-latex). The flask was
gently mixed and put back in the incubator for 2.5 h to label the
cells with occasional rocking of the flask. At the end of the labeling,
the supernatant was removed. The cells were rinsed three times with
Opti-MEM to wash off the unincorporated beads and left in -MEM
culture overnight to reduce cell aggregation on detaching. The labeled
cells were detached on the following day with trypsin-EDTA-PBS (2:3
dilution) at 37°C for 5 min with frequent rocking of the
flask. The cell suspension was transferred into a centrifuge tube with
an additional 5 ml of culture medium to stop the trypsin reaction and
centrifuged at 170 x g at room temperature
for 3 min in a bench top centrifuge. The cell pellet was resuspended in
saline and kept on ice before injection. The injecting cell viability
measured by trypan blue was 95.0 ± 3.1% (SD) in the
study.
|
B16F1 Injection
|
|---|
Surgery was performed under aseptic conditions between 11 a.m. and 6 p.m. Instruments were soaked in 70% ethanol
before the operation. The mouse was not stimulated by any cytokines or
bacterial endotoxin before cell injection. The mouse was anesthetized
with Avertin (30 mg/ml; 0.2–0.3 ml/mouse, i.p). The abdomen was
cleaned with 70% ethanol and opened through a small midline incision
(
10 mm). The intestine was gently pulled out to expose the
mesenteric blood vessels. One branch of the superior mesenteric vein
was isolated by separating the connective tissue around the vein, using
forceps. A fluorescence-labeled cell suspension of 5 x 105 B16F1 melanoma cells in a volume of 150 µl
in saline was injected into the isolated mesenteric vein over a
5–10-min period using a 30 G1/2 needle and a 1-ml syringe. A
small piece of Gelfoam Sterile Sponge (12–7 mm; Upjohn) was used to
stop the bleeding at the end of the injection. The incision was sutured
by muscle and skin two layers using synthetic absorbable sutures
(3–0 Vicryl; Ethicon, Johnson & Johnson, Peterborough, Ontario,
Canada). The mouse was allowed to recover with a plastic catheter
inserted in the mouth to facilitate breathing. The liver was collected
under anesthesia at specific times (0–24 h) after injection. Samples
with macroscopic infarction were excluded, and no microscopic
infarction was found in the samples analyzed. For 0 h samples, the
liver tissue was removed immediately at the end of the injection
(within 30 s to 1 min) without suturing the abdomen. A group of
mice received injection of 15 µm of fluorescent microspheres
(5 x 105 beads/150 µl of
saline) following the same procedure, and liver samples were removed
immediately after injection (0 h).
|
L-NAME Administration
|
|---|
The competitive NO synthase inhibitor L-NAME was administered in
the doses and routes shown in Table 1
, to inhibit NO production and test its effects on cell arrest,
apoptosis, and metastasis formation. The inactive enantiomer of L-NAME,
D-NAME, was administered by regimens A and D as described for L-NAME
(see "Results").
|
EPR Measurement of NO Production
|
|---|
The method of Doi et al. (13)
was
modified. The NO-trapping agents DETC (400
mg/kg-1 in saline,
administered i.p.) and FeSO4/sodium citrate
(trisodium salt; formula weight 294.1) (40
mg/kg-1 200
mg/kg-1 mixed in
water; administered s.c. on the back) were given to each mouse
30 min before obtaining the liver sample. The mouse was anesthetized 10
min before sacrifice. The abdomen was opened, and constant portions of
the left lateral, medial left and right, and right lateral lobes were
removed quickly (between 30 s and 1 min) and placed onto a Petri
dish precooled on ice. The mouse was sacrificed by opening the
diaphragm and heart. The liver was sliced quickly into smaller pieces,
placed into a precooled 1-ml syringe, and transferred into a suprasil
synthetic quartz tube (2.4 mm, inner diameter; Heraeus Amersil,
Atlanta, GA) by pushing the tissue through an 18-gauge needle
from the syringe. An aliquot of liver weighing 0.27 ± 0.04 gram/mouse was transferred into each quartz tube (filled to 6 cm
in height per tube), which was then immediately placed into liquid
nitrogen until NO was measured by EPR spectroscopy. The remaining liver
tissue was removed from the mouse and fixed in 1% paraformaldehyde
(BDH Laboratory Supplies, Poole, United Kingdom) for 24–48 h at 4°C.
To assay the production of NO in cultured cells (control), fluorescent
bead-labeled B16F1 cells (5 x 105
cells/150 µl of saline) were supplemented with DETC and
FeSO4/sodium citrate at a final dilution of 1/15
(to approximate the final in vivo concentration),
transferred into a quartz tube, and stored in liquid nitrogen before
measurement.
EPR measurement of the trapped
NO-Fe2+-(DETC)2 complex was
carried out at 115 K (-158°C; temperature controller Model ER-4111;
Bruker Instruments). The spectra were measured with a Varian Associates
Model E-12 EPR spectrometer operating at 9.025 GHz with 100 KHz
modulation that was interfaced with a Nicolet Instruments Model 1180
computer and Model 2090 digital oscilloscope. The instrument settings
were as follows: (a) microwave power, 5 mW; (b)
modulation amplitude, 5 G; and (c) scan range, 505 or 1000
G. The concentration of the
NO-Fe2+-(DETC)2 complex in
each sample was assumed to be proportional to the signal amplitude
(peak-to-peak) of the triplet-hyperfine structure (hyperfine splitting
of 13 G) observed at g = 2.04 and expressed
as relative EPR signal intensity (arbitrary units) after subtracting
the Cu2+-(DETC)2 complex
signal observed in all samples (14
, 15)
.
|
Analysis of Cell Arrest
|
|---|
For 0 h samples, 15-µm frozen sections from
constant sites at three different planes in the left lateral and medial
and right lateral lobes per liver were cut using a freezing microtome
(American Optical Company, Buffalo, NY). Ten microscopic fields (x100
magnification) with the highest cell numbers were counted from each
lobe to determine the number of fluorescence-labeled B16F1 cells in
both the TPV and sinusoidal regions. Data were expressed as the sum of
the three mean values to increase the area of sampling. Each mean was
determined by counting 10 fields in each lobe. The percentage of cells
per region (sinusoids or TPV) was calculated by the following formula:
total cell number (sum of three means) per region/total cell number in
both regions (sum of six means) x 100.
|
Assessment of in Vivo B16F1 Cytotoxicity by Double
Labeling
|
|---|
Frozen sections (17 µm) were obtained from random sampling
sites in the same livers described above. In situ DNA end
labeling (TUNEL assay; ApopTag S7100 Kit) using a
digoxigenin-peroxidase detection system was performed on these
sections, which contained injected fluorescent microsphere-labeled
melanoma cells. To achieve optimal double labeling with fluorescent
microspheres and TUNEL assay, the procedures were performed according
to the manufacturer’s instructions with the following modifications:
(a) for frozen tissue sections, proteinase K digestion (20
µg/ml), which was carried out at room temperature for 15 min followed
by two washes with double-distilled H2O,
was performed before permeabilizing the cells with ethanol/acetic acid
(2:1, v/v) to facilitate the unmasking of fragmented DNA in apoptotic
tumor cells; (b) TdT incubation was performed overnight in a
humidified chamber at 37°C with a subsequent 1-h incubation with
antidigoxigenin peroxidase conjugate at room temperature in a
humidified chamber and a 12-min peroxidase substrate color development;
(c). the "counter stain specimen" and "mount
specimen" steps were omitted to prevent the quenching of fluorescence
in the microspheres by organic solvents; (d) the slides were
mounted with Gel/Mount (Biomeda Corp., Foster City, CA); and
(e) to double label fluorescent microsphere-labeled B16F1
cells in culture, the cells were fixed in 1% paraformaldehyde, with a
2-h TdT incubation and a 1-h antidigoxigenin peroxidase incubation to
optimize double labeling.
The total number of both single-fluorescent and doubly-labeled
cells (fluorescent-ApopTag DNA end labeling) and their locations on all
sections per sample were scored using a microscope with dual
fluorescent and incandescent illumination. To achieve a constant
sampling area, we quantified all tumor cells located in 5
fields/section, 10 sections/sample, under x100 magnification. The
percentage of in situ DNA end-labeled melanoma cells in the
TPV or sinusoidal region was calculated by the following formula:
in situ DNA end-labeled tumor cells/region (%) = 100 x [number of double-labeled cells]/[total
number of cells (double labeled + single labeled cells) in
the region].
|
Assessment of in Vitro B16F1 Cytotoxicity by NO
Donor-derived NO
|
|---|
Determination Using Fluorescence TUNEL Assay Dual Label and DNA
Fragmentation.
Cultured B16F1 melanoma cells were exposed for 5 min to the NO donor
SNAP. Melanoma cells were grown to 70–80% confluence in
25-cm2 flasks, labeled with fluorescent
microspheres, and cultured overnight as described previously. Stock
SNAP (100 mM in methanol) was incubated at 37°C for 10
min to initiate NO release and then added to the flask at a final
concentration of 2.5 mM. The flask was incubated for 5 min,
and the medium was removed. The monolayer of cells was washed twice
with culture medium. Fresh complete culture medium was added to the
flask, which was then put back into the 37°C incubator for 4, 8, or
24 h. Each flask was set up for one time point and handled
individually to ensure the 5-min exposure time was fulfilled. The
labeled cells without SNAP exposure (controls) were cultured for the
same time at 4, 8, and 24 h thereafter. The cells were then
detached with trypsin-EDTA and centrifuged at 170 x g for 3 min. The cell pellet was resuspended in saline. To
assess apoptosis induced by a short exposure to SNAP, part of the cells
were fixed in 1% paraformaldehyde for DNA end labeling, and the
remaining cells were used for the DNA fragmentation assay. DNA end
labeling was performed with 2 h of TdT incubation and 1 h of
antidigoxigenin peroxidase incubation. DNA fragmentation was assessed
using a Suicide Track DNA Ladder Isolation Kit (Cedarlane Laboratories
Ltd., Hornby, Ontario, Canada).
Determination by Flow Cytometry Using Annexin V Plasma Membrane
Labeling (Phosphatidylserine Binding by Annexin V-FITC in the Early
Stage of Apoptotic Cells).
Unlabeled B16F1 cells were seeded into 6-well culture plates (Costar)
at 1 x 105 cells/well (9.6
cm2). Cells were exposed to SNAP for 5 min at the
same concentrations and by same procedures as described above. The
cells were cultured for 24 h thereafter in the complete culture
medium without SNAP. The cells were then detached using trypsin-EDTA,
pelleted by centrifugation, washed twice with cold PBS, and stained
with annexin V-FITC and PI (50 µg/ml; annexin V-FITC apoptosis
detection kit I; PharMingen, San Diego, CA). Cells cultured for the
same period of time without SNAP treatment were set up as internal
controls for the background of early and late apoptosis, necrosis, and
membrane damage during trypsinization (16)
. The seeding
cell viability was 99 ± 1% by trypan blue exclusion.
Flow cytometry analysis was performed on an EPICS ALTRA cell sorter
(Beckman Coulter Inc., Miami, FL) with argon ion laser excitation at
488 nm (150 mW). Forward versus side scatter histograms were
acquired to identify and set bit map gates for single intact cells.
These gates were sufficiently wide to include both unaltered and
apoptotic cells. The fluorescence signals derived from each cell were
directed to the appropriate photomultiplier detectors by using 550 and
640 nm long-pass dichroic mirrors, with the FITC and PI fluorescence
detected through the 525 and 610 nm bandpass filters, respectively.
Bivariate histograms (log-log) relating annexin-FITC and PI
fluorescence were based on 5000 cells that satisfied the light scatter
gates. Color compensation was adjusted using cell samples stained with
annexin-FITC alone or PI alone. Data were acquired in listmode format
and subsequently analyzed using the EXPO 2 system software (Build 320)
supplied with the instrument.
|
Metastasis Studies
|
|---|
Unlabeled B16F1 cells (5 x 105 cells/150 µl of saline) were injected into
a mesenteric vein of the mouse and allowed to grow in vivo
for 3–7 days by suturing the abdomen of the mouse and allowing it to
recover in a warm environment. If any animal in an experimental group
failed to survive, all other mice in the same group were sacrificed at
the same time. Postsurgical care was given according to the guidelines
of the Central Animal Care Services at the University of Manitoba. The
mouse was sacrificed under anesthesia at the sampling time, and the
liver was fixed in 10% neutral buffered formalin. Surface metastatic
nodules 
0.5 mm on all liver lobes were counted using
a dissecting microscope (Olympus, Japan; x10). All animal
surgical procedures were approved by the Bannatyne Campus Protocol
Management and Review Committee at the University of Manitoba.
Histomorphometric analysis was performed on histological sections from
the left lateral, medial, and right lateral lobes of the same livers
using a Merz Graticule to quantify the percentage of tissue area
occupied by the metastases and the proportion of the tumor that was
located on the liver surface (7)
.
|
Statistical Analysis
|
|---|
Data from early-stage studies were analyzed using one-way and
multiway ANOVA with repeated measures, followed by planned comparisons
for pairwise comparisons. Unpaired t test and the
Mann-Whitney U test were used for metastasis studies.
|
RESULTS
|
|---|
NO Induction after B16F1 Injection.
Immediately after the injection of fluorescent
microsphere-labeled B16F1 cells into the portal circulation, there was
a burst of hepatic NO production. This was identified by using EPR
spectroscopy to detect trapped
NO-Fe2+-(DETC)2 complex in
the liver (Fig. 1A)
. This initial peak was not detected >5 min after
injection, in the livers of untreated animals, or in sham controls
injected with saline. The peak level of NO in the livers of animals
injected with B16F1 cells was estimated to be in the range of 50–100
nmol/gram of wet tissue. This was determined by using an in
vitro standard with a well-calibrated concentration of
NO-Fe2+-(DETC)2 measured by
double integration of the EPR spectrum obtained from a short 10–15-min
hydrolysis of the NO donor SNAP under conditions identical to those
used for analysis of the in vivo samples. The actual
NO-Fe2+-(DETC)2
concentration was calibrated relative to the double integration of the
EPR spectrum of a standard nitroxide at known concentration under
identical EPR operating conditions (115 K). The initial and extremely
transient production of NO was followed by a slow and much lower rise
of NO levels between 8 and 24 h after injection (Fig. 1A)
. Rapid and significant production of NO (264 ± 76 relative EPR signal intensity; n = 4; P < 0.01) was also detected in mice
injected with 5 x 105 fluorescent
microspheres (15 µm) with a diameter similar to that of B16F1 cells
(16 µm) (7)
. In these animals, the peak NO level was
significantly lower than that induced by B16F1 cell injection
(P < 0.001). The labeling of B16F1 cells
with fluorescent beads was not associated with any significant NO
production by these cells (12 ± 2 relative EPR signal
intensity; n = 3 separate cultures; EPR
conditions were the same as those used to evaluate the liver tissue
samples). Representative EPR spectra for NO detection under various
conditions are shown in Fig. 1B
.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 1. A, EPR spectroscopic measurement of NO
induction in the liver. Five x 105 B16F1
melanoma cells in 150 µl of saline were injected into the portal
circulation via a mesenteric vein. The mice were given NO-trapping
agents 30 min before sacrifice. The relative EPR signal intensity of
each NO-Fe2+-(DETC)2 complex was expressed in
arbitrary units, which were calculated by measuring the distance from
the peak to the trough of the triplet-hyperfine structure at the
g = 2.04 location after subtraction of
the Cu2+-(DETC)2 signal. NL,
normal livers without injection of cells. Saline, saline
injection only. n = 3 mice/point. Data
are expressed as the mean ± SE. ***,
P < 0.001. **,
P < 0.01. B,
representative complete EPR spectra of
NO-Fe2+-(DETC)2 complex (triplet-hyperfine
structure at g = 2.04) measured
immediately after injection of various agents
[Cu2+-(DETC)2 signal was not subtracted].
L-NAME = 2 doses of 2.5 mg/kg, given by
intramesenteric injection 10 min before cell injection and
simultaneously with the cells. The 505 G scan range is displayed for
each spectrum.
|
|
Arrest and Killing of Injected B16F1 Cells.
Immediately after injection, 65 ± 1% of cells were
found in the sinusoids, and 35 ± 1% of cells were
found in the large TPVs (Fig. 2)
. Subsequently, from 4–24 h, cells in both locations became
susceptible to DNA end labeling. However, after 4 h, a
significantly greater percentage of melanoma cells in the sinusoids
were DNA end labeled as compared with those in the TPVs (Table 2
and Fig. 3
). The location of the neoplastic cells in the sinusoids or TPVs
was clearly distinguishable. Deformation of some B16F1 cells was
observed in samples taken immediately after injection, and membrane
disruption was observed at 4 h and thereafter (data not shown).
Small fluorescent fragments or individual microspheres were not
counted.
View larger version (126K):
[in this window]
[in a new window]
|
Fig. 2. Location of B16F1 cells in the liver immediately after
injection. Fluorescent microsphere-labeled B16F1 cells (5 x 105 cells in 150 µl of saline) were injected into
the mesenteric vein of the mouse. The liver was removed immediately
thereafter, and 15-µm frozen sections were cut. Arrows
indicate melanoma cells that are located in the sinusoids
(S) and TPVs. Fluorescence microscopy; x100; FITC UV-2A
filter.
|
|
View this table:
[in this window]
[in a new window]
|
Table 2 Apoptotic cell counts in the liver after melanoma cell injection
Absolute counts of cells from which data in Fig. 3A
were
calculated. Numbers represent the median and range of the three animals
per group. Total = apoptotic + nonapoptotic cells/region.
|
|
View larger version (42K):
[in this window]
[in a new window]
|
Fig. 3. A, fluorescence/DNA end labeling of
melanoma cells in the mouse liver. Random sections of frozen liver,
17-µm thick, were cut from the same livers used in Fig. 1
A and stained with the ApopTag TUNEL assay. The total
number of single-labeled (fluorescence only) and double-labeled
(fluorescence/ApopTag) cells and their locations were scored on all
sections using dual illumination with incandescent and fluorescent
light (FITC UV-2A filter). The percentages of double-labeled cells in
the TPVs and sinusoids were calculated. The data are expressed as the
mean ± SE of the three liver samples per time point per
treatment group. S, sinusoids. B16 sinusoids
versus TPVs: 4 h, P < 0.01; 8 h, P < 0.05; and
24 h, P < 0.001. B16 TPV (4 h)
versus (0 h), P < 0.01.
B16 + L-NAME sinusoids versus B16
sinusoids: 4 h, P = 0.055; 8 h,
P < 0.05; 24 h,
P < 0.01. B16 + L-NAME TPV versus B16 TPV:
P > 0.05 for all points.  , B16,
sinusoids;  , B16, TPV;  , B16 + L-NAME, sinusoids; •,
B16 + L-NAME, TPV; n = 3
mice/point;  , B16 + DN, sinusoids
(n = 2);  , B16 + DN, TPV.
Absolute cell numbers are shown in Table 2
. B,
fluorescence/ApopTag double labeling showing DNA end-labeled
(large arrow) and nonlabeled melanoma cells
(small arrow) in the sinusoids 4 h after
injection.
|
|
The NO synthase inhibitor L-NAME was given to mice
receiving B16F1 injections to determine whether the DNA end labeling
effects could be related to the production of NO. The initial peak of
NO production was completely inhibited by L-NAME, which was
administered according to three different dose regimens
(n = 3 mice/point): (a) regimen A
(10 ± 0 relative EPR signal intensity); (b)
regime B (42 ± 17 relative EPR signal intensity; Fig. 1B
); and (c) regimen C (18 ± 3
relative EPR signal intensity). Regimens A and D also eliminated the
later rise of NO at 8 and 24 h (data not shown). The effect of
regimens B and C was not determined in the 8 and 24 h samples.
After L-NAME administration, the percentage of
DNA end-labeled B16F1 cells in the sinusoids decreased to the same
level as that seen in the TPVs (Fig. 3A)
.
L-NAME did not alter the proportion of DNA
end-labeled cells in the TPV region (Fig. 3A)
. The inactive
isomer, D-NAME, did not affect the NO burst at 0 h
after cell injection (data not shown) or the proportion of apoptotic
cells compared with these phenomena in animals receiving tumor cell
injection alone (Fig. 3A)
. Under all experimental
conditions, two-thirds of the cells (median, 66%; range, 65–78%)
arrested in the sinusoids, and one-third of the cells (median, 34%;
range, 22–35%) arrested in the TPVs.
In Vitro Cytotoxicity of NO on B16F1 Cells.
The cytotoxic effect of NO on B16F1 melanoma cells that we observed
in vivo was also demonstrated in vitro using NO
derived from the NO donor SNAP. A 5-min exposure of cultured B16F1
cells to 2.5 mM SNAP, which represented the
transient exposure of injected B16F1 cells to NO in the liver,
increased DNA end labeling at later time points with a profile (Fig. 4, A and B)
similar to that observed in vivo (Fig. 3, A and B)
. Five-min exposures to 5 and 10
mM SNAP induced an increase of DNA end labeling
at 8 h in a dose-dependent manner (data not shown). The in
vitro double labeling of apoptotic B16F1 cells at 8 h after a
5-min exposure to 5 mM SNAP is demonstrated in
Fig. 4B
. A typical apoptotic DNA ladder pattern was not
observed in 2.5 mM SNAP-treated cells at 8 or
24 h after 5-min SNAP exposure, presumably due to the low
percentage of DNA end-labeled cells in the samples (1.9–7.8%; Fig. 4A
). However, DNA ladders were generated with a longer
exposure to SNAP (24 h; Fig. 4C
). Flow cytometry analysis
showed a >20% increase of cell frequencies in the early apoptosis
region (Fig. 5C
, A4), and in the late apoptosis and dead cell
region (Fig. 5C
, A2) after a 5-min exposure of
cultured B16F1 cells to 2.5 mM SNAP in comparison
with the untreated control (Fig. 5B)
. The data further
support the hypothesis that a short exposure to NO could trigger
apoptotic cell death at a later time. Compared with the 5-min
treatment, a 24-h continuous exposure to SNAP only increased cell
frequency in the late apoptosis and dead cell region (Fig.
5D, A2) without a further increase in the early
apoptosis region (Fig. 5D, A4).
View larger version (47K):
[in this window]
[in a new window]
|
Fig. 4. Cytotoxicity of NO to cultured B16F1 cells in
vitro. A, DNA end labeling induced by a 5-min
exposure of cultured B16F1 cells to SNAP. SNAP (2.5 mM) was
added to individual flasks containing fluorescence-labeled cells and
incubated for 5 min at 37°C. The supernatant was removed, the
monolayer was rinsed twice with culture medium, and cells were
incubated in complete culture medium for the time indicated. Control
cells were labeled with fluorescent microspheres and cultured for the
same time without SNAP treatment. Cells were harvested using
trypsin-EDTA, fixed in 1% paraformaldehyde, and stained with the
ApopTag kit. Five microscopic fields (x100 magnification) per sample
were analyzed for the percentage of double-labeled cells. Data were
expressed as the mean ± SE of five fields per sample.
One flask = one time point. **,
P < 0.01 [SNAP (24 h)
versus control (24 h); SNAP (24 h) versus
SNAP (4 h)]. B, double labeling of apoptotic B16F1
cells (arrow) induced by a 5-min exposure to SNAP (5
mM) and culture in complete medium for 8 h (dual
illumination with fluorescent and incandescent light; FITC UV-2A
filter; x 400). C, DNA fragmentation of cells
treated continuously with SNAP. B16F1 cells at 80% confluence were
treated with 500 µM SNAP (Lane b) and 2
mM SNAP (Lane c) for 24 h. Cells were
pelleted by centrifugation. DNA was extracted using the Suicide Track
DNA Ladder Isolation Kit and electrophoresed on a 1.5% agarose gel. A
total of 10 µg DNA/lane was loaded. Lane a, 100-bp DNA
marker.
|
|
View larger version (39K):
[in this window]
[in a new window]
|
Fig. 5. Histograms of flow cytometry analysis in the control and
SNAP-treated B16F1 cultures. B16F1 cells (1 x 105 cells/well) were seeded (6-well plates) and cultured
for 5 min (with a subsequent 24-h culture in complete medium) or
24 h in the presence of 2.5 mM SNAP. The cells were
detached with trypsin-EDTA and stained with annexin V-FITC and PI
according to the manufacturer’s instructions. A3
region, vital cells (annexin V-/PI-);
A4 region, early apoptotic cells (annexin
V+/PI-); A2 region, late
apoptotic and dead cells (annexin V+/PI+);
A1 region, damaged cells and cell debris (annexin
V-/PI+). FALS, forward angle
light scatter. SS, side scatter. Data from one of two
experiments are shown.
|
|
Metastasis Studies.
Mice injected with B16F1 plus L-NAME (regimen D) had
a 5-fold increase in the number of liver surface metastatic nodules in
comparison with the group receiving B16F1 cells alone.
Histomorphometric analysis on sections from three different lobes
showed a 20-fold larger histological area occupied by metastases in
B16 + L-NAME groups than in the B16 only controls. A total of
90 ± 6% of the nodules from both groups were found on
the liver surface. On histological examination, all small tumors that
might not be scored under the dissecting microscope appeared to
originate from TPVs, as described previously (7)
. The
results are shown in Table 3
.
View this table:
[in this window]
[in a new window]
|
Table 3 Effects of L-NAME on the formation of metastases by B16F1 cells
Unlabeled B16F1 (1 x 105 cells in 150 µl of
saline) were injected into the mesenteric vein of the mouse without
(controls) or with L-NAME (2.5 mg/kg L-NAME was coinjected with the
cells, and 5 mg/kg L-NAME was injected i.p. 20 h after injection).
The mice were killed between 3 and 7 days after injection. The
metastatic nodules (0.5 mm in diameter) on the surfaces of all lobes
of the liver were counted under a dissecting microscope.
Histomorphometric analysis was performed as described in "Materials
and Methods."
|
|
|
DISCUSSION
|
|---|
The goal of this study was to test the hypothesis that the hepatic
microvasculature regulates the arrest and fate of metastasizing cancer
cells through the release of NO.
NO is a small diatomic free radical, but it plays complex roles in
diverse biological processes in organisms ranging from bacteria to
mammals (17)
. The main difficulty in measuring NO
production is caused by its extremely labile nature and short
half-life, which is in the millisecond range. Among the indirect NO
detection methods widely accepted in the literature
(18, 19, 20)
, we chose to use spin trapping of a NO adduct
in vivo and then measure the trapped NO signal ex
vivo by EPR spectroscopy. The main advantages of spin trapping are
that NO has a high affinity for the iron complex in the spin trapping
agent and can instantaneously form a much more stable
NO-Fe2+-(DETC)2 complex
in vivo at the site of NO release and that the trapped
NO-Fe2+-(DETC)2 complex
generates a typical three-line structure on the EPR spectrum that can
be distinguished from other signals. This technique has been shown to
be a reproducible and sensitive NO detection method (13, 14, 15
, 21)
. We argue that it is the sensitivity and specificity of the
EPR spin trapping method that helped us to capture the transient burst
of NO release in the liver after the melanoma cell injection. We
further quantified the NO peak concentration to be in the range of
50–100 nmol/gram of wet tissue using a well-calibrated in
vitro standard. In addition to the three-line structure of
NO-Fe2+-(DETC)2 complex on
the EPR spectrum, a quartet hyperfine spectrum of
Cu2+-(DETC)2 was also
detected (14
, 15
, 21)
, which showed a partial overlapping
with the NO-Fe2+-(DETC)2
spectrum. Our data were expressed as relative EPR signal intensity
(equivalent to the amplitude of peak to trough of the three-line
structure) with the
Cu2+-(DETC)2 spectrum
subtracted, so that they represent the pure
NO-Fe2+-(DETC)2 signal
intensity. To our knowledge, this is a more specific approach because
most analyses measure the EPR NO concentration on the basis of the
spectrum only.
The induction of NO production immediately after B16F1 cell injection
suggested that the hepatic microvasculature is capable of eliciting
rapid vascular responses probably within seconds of cell entry into the
liver. The inhibition of the response by L-NAME confirmed
the generation of NO. Because we did not detect NO production by
populations of fluorescent bead-labeled B16F1 cells, we suggest that NO
was produced by the liver and to a greater proportion by the sinusoidal
endothelial cells because two-thirds of the injected cells were found
in the sinusoids at 0 h after injection. The injection of 15-µm
fluorescent microspheres also induced a rapid NO production, but at a
significantly lower level, suggesting that NO production may be both
microvascular and tumor cell dependent. The basal levels of NO
production detected in the normal livers and saline-injected livers
reflected the NO product in the hepatic microvasculature under
physiological conditions. It is possible that the later occurrence of
NO production detected after 8 h of injection originated from a
different cellular source [e.g., Kupffer cells or
hepatocytes, (22)
], which may contribute in part to the
observed sinusoidal cytotoxicity in arresting melanoma cells between 8
and 24 h after injection. The extremely transient nature of the
initial NO peak is not fully understood. It may be that the induction
of NO occurred only during the entry of cells into the liver (5–10
min), that NO was quickly released into the sinusoids, and that the
Fe2+-(DETC)2-trapped NO was
flushed out of the liver by the blood circulation [4.3 ± 1.9 pl/s in the livers of nude mice, (23)
]. These
possibilities would be consistent with an intravascular event.
Metabolism of
NO-Fe2+-(DETC)2 complex
could be another reason for the transient nature of the peak, although
little information is available on the in vivo metabolism of
NO-Fe2+-(DETC)2 complex
(24)
. The observation that arrest of inert microbeads
elicited a NO burst similar to that associated with cell arrest
suggests that mechanical forces may initiate transient NO production.
Because the quantity of NO released was greater after cell injection,
other tumor cell-regulated pathways may also be involved.
A significantly greater proportion of melanoma cells underwent
apoptosis in the sinusoids than in the TPVs. The observation that
L-NAME treatment inhibited the initial NO burst and also
inhibited the magnitude of DNA end labeling of the B16F1 cells in the
sinusoid suggests that apoptosis of the melanoma cells in this region
is NO dependent. Because L-NAME did not affect apoptosis in
the TPVs, a separate mechanism may be involved in this zone. It has
been well recognized and documented that there is a functional and
structural heterogeneity of the hepatic endothelial, parenchymal, and
sinusoidal lining cells and extracellular matrix in the space of Disse
across different zones of the liver acinus (25)
. The
endothelial layer has different cellular components in the TPVs
(endothelial cells) and sinusoids [endothelial cells and Kupffer cells
(26)
]. We have previously shown that the expression of
various adhesion molecules on the endothelium of the liver
microvasculature is inducible (25)
and that this can alter
the site of arrest of tumor cells that enter the microvasculature
(7)
. The present study suggests that the cytotoxic
properties of the liver may also have a zonal distribution. This is
supported by our previous evidence that cells that arrest in the TPVs
are more likely to form metastases than cells in the sinusoids
(7)
. The effects of NO do not appear to be related to an
altered distribution of the tumor cells, although the NO effect does
appear to be site specific. The cell arrest data (65 ± 1% in sinusoids and 35 ± 1% in TPV) fit well with our
previous work, which showed melanoma cells arrested primarily in
the presinusoidal portal veins in IL-1-stimulated animals and
arrested primarily in the sinusoids in unstimulated animals
(7)
.
The data suggest that NO is an effector of tumor cell destruction in
the sinusoids. The cytotoxicity-inducing effect of a brief exposure of
B16F1 cells to NO was confirmed in vitro. The data do not
infer an exclusive role for NO and do not exclude the possibility that
other sinusoidal factors are cytotoxic. Hydrogen peroxide
(H2O2), superoxide anion
(O), peroxynitrite (ONOO-),
Pit cells, and Kupffer cells are all potentially cytotoxic to tumor
cells in the sinusoids (27, 28, 29, 30)
. Apoptosis of melanoma
cells in the TPVs may be mediated by these other mechanisms because
L-NAME had no effect on this and therefore is not NO
dependent. Because some cells showed deformation and membrane
disruption immediately after arrest (31)
, mechanical
injury may account for some of the cell death we have observed. Because
it is not possible to sample the entire liver of each animal, the
present study best represents the changes observed at the sample site.
Our ongoing studies have suggested to us that even by using a double
labeling procedure, we are still underestimating the tumor cell killing
in general because of: (a) tumor cell clearance by the liver
(7)
; (b) tumor cell fragmentation after
mechanical arrest (31)
; and (c) clearance of
the end stage apoptotic cells (32
, 33)
. Therefore,
measurement of ApopTag-labeled cells should be considered as only one
index of cell death. Moreover, the increased liver metastases seen
after administration of L-NAME support the hypothesis that
NO-mediated tumor cell cytotoxicity is one determinant of metastatic
outcome and reinforces the significance of the apoptotic data.
Our results are supported by the findings of Edmiston et al.
(34)
, who showed in vitro that unstimulated
murine sinusoidal endothelial cells are selectively toxic to a
population of weakly metastatic cells by the production of reactive
oxygen species. These workers have recently demonstrated that normal
liver fragments in culture are nontoxic to weakly metastatic colorectal
cancer cell lines, and only ischemic and reoxygenated liver fragments
killed those cells (5)
. Our experiments provide evidence
for the existence of a mechanism by which cell arrest can trigger the
release of the cytotoxic free radicals, which may account for one of
the potential mechanisms of metastatic inefficiency (35)
.
Our study and the one by Luzzi et al. (36)
both
support the principles of metastatic inefficiency, although
different methodologies were used in each study. Our data are
consistent with the observations of Rocha et al.
(37)
, who have demonstrated a role for endothelial cell NO
in suppression of lymphoma metastasis in the liver. Other evidence also
indicates that NO may be able to trigger apoptosis in rat cardiac
myocytes (38
, 39)
.
In the present study, we have observed that the arrest of B16F1
melanoma cells or inert 15-µm polystyrene microbeads in the hepatic
blood vessels induces an immediate burst of hepatic NO, followed by
delayed expression of NO after 8 h. This effect was related to the
injection of a bolus of cells (5 x 105 cells in 150 µl of saline) that might be
predicted to cause a major hemodynamic disturbance within the injection
time (5–10 min). Others have shown that blood flow can regulate the
expression of endothelial NO synthase (40)
and that
endothelial NO synthase can be activated by mechanical forces including
shear stress, pulsatile and cyclic circumferential stretch, and
isometric contraction of the endothelium (41
, 42)
.
Observations on humans and animals have indicated that many cancer
cells gain access to the circulation during the natural history of even
small cancers (1)
. However, it is not known whether the
arrest of individual cells in the liver during spontaneous metastasis
is capable of inducing local NO production.
As an interesting molecule, NO has been shown to play opposite roles in
both inhibition and stimulation of tumor growth (43, 44, 45)
.
Our study implies that either inhibition of local NO production or
acquisition of NO resistance by the tumor cells would be conducive to
successful hepatic metastasis. The latter is consistent with the
findings of Ambs et al., who showed that tumor
cell-associated NO production promoted cancer progression in nude mice
and that this effect was tumor suppressor gene product p53-dependent
(46
, 47)
.
In summary, our data have identified a natural defense mechanism
against cancer metastasis in the hepatic microvasculature whereby the
arrest of cancer cells in the liver induces endogenous NO release,
leading to sinusoidal killing of cancer cells and reduced metastasis
formation.
|
ACKNOWLEDGMENTS
|
|---|
We thank Dr. B. MacNeil for help in statistical analysis and
discussion of the data. Dr. D. Goetz kindly provided 15-µm
fluorescent polystyrene microspheres. We also thank Mrs. A. Lawless for
his participation in the work of in vitro calibration of
NO-Fe2+-(DETC)2 signal
in vivo and M. Webber for his help in figure preparation.
|
FOOTNOTES
|
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 Supported by Medical Research Council of Canada
Grants MT-14356 (to D. M. N. and F. W. O.) and MT-14477 (to
B. B. H.) and a University of Manitoba Fellowship (to H. H. W.). 
2 To whom requests for reprints should be
addressed, at the Department of Pathology, Room D212, 770 Bannatyne
Avenue, Winnipeg, Manitoba, Canada R3E OW3. Phone: (204) 789-3338;
Fax: (204) 789-3931; E-mail: worr{at}cc.umanitoba.ca
3 The abbreviations used are: NO, nitric oxide;
EPR, electron paramagnetic resonance; DETC, diethyldithiocarbamate;
TdT, terminal deoxynucleotidyl transferase; TUNEL, TdT-mediated UTP end
labeling; L-NAME, NG-nitro-L-arginine methyl
ester; D-NAME, NG-nitro-D-arginine methyl
ester; SNAP,
S-nitroso-N-acetylpenicillamine; TPV,
terminal portal venule; PI, propidium iodide; IL, interleukin.
Received 1/14/00.
Accepted 8/17/00.
|
REFERENCES
|
|---|
-
Orr, F. W., Wang, H. H., Lafrenie, R. M., Scherbarth, S., and Nance, D. M. Interactions between cancer cells and the endothelium in metastasis. J. Pathol., 190: 310–329, 2000.
-
Jiang W. G. Cell adhesion molecules in the formation of liver metastasis. J. Hepatobiliary Pancreat. Surg., 5: 375-382, 1998.[Medline]
-
Khatib A. M., Kontogiannea M., Fallavollita L., Jamison B., Meterissian S., Brodt P. Rapid induction of cytokine and E-selectin expression in the liver in response to metastatic tumor cells. Cancer Res., 59: 1356-1361, 1999.[Abstract/Free Full Text]
-
Anasagasti M. J., Alvarez A., Martin J. J., Mendoza L., Vidal-Vanaclocha F. Sinusoidal endothelium release of hydrogen peroxide enhances very late antigen-4-mediated melanoma cell adherence and tumor cytotoxicity during interleukin-1 promotion of hepatic melanoma metastasis in mice. Hepatology, 25: 840-846, 1997.[Medline]
-
Jessup J. M., Battle P., Waller H., Edmiston K. H., Stolz D. B., Watkins S. C., Locker J., Skena K. Reactive nitrogen and oxygen radicals formed during hepatic ischemia-reperfusion kill weakly metastatic colorectal cancer cells. Cancer Res., 59: 1825-1829, 1999.[Abstract/Free Full Text]
-
Umansky V., Schirrmacher V., Rocha M. New insights into tumor-host interactions in lymphoma metastasis. J. Mol. Med., 74: 353-363, 1996.[Medline]
-
Scherbarth S., Orr F. W. Intravital videomicroscopic evidence for regulation of metastasis by the hepatic microvasculature: effects of interleukin-1 on metastasis and the location of B16F1 melanoma cell arrest. Cancer Res., 57: 4105-4110, 1997.[Abstract/Free Full Text]
-
Barbera-Guillem E., Alonso-Varona A., Vidal-Vanaclocha F. Selective implantation and growth in rats and mice of experimental liver metastasis in acinar zone one. Cancer Res., 49: 4003-4010, 1989.[Abstract/Free Full Text]
-
Li L. M., Nicolson G. L., Fidler I. J. Direct in vitro lysis of metastatic tumor cells by cytokine-activated murine vascular endothelial cells. Cancer Res., 51: 245-254, 1991.[Abstract/Free Full Text]
-
Li L. M., Kilbourn R. G., Adams J., Fidler I. J. Role of nitric oxide in lysis of tumor cells by cytokine-activated endothelial cells. Cancer Res., 51: 2531-2535, 1991.[Abstract/Free Full Text]
-
Geng Y. J., Hellstrand K., Wennmalm A., Hansson G. K. Apoptotic death of human leukemic cells induced by vascular cells expressing nitric oxide synthase in response to
-interferon and tumor necrosis factor-. Cancer Res., 56: 866-874, 1996.[Abstract/Free Full Text]
-
Kong L., Dunn G. D., Keefer L. K., Korthuis R. J. Nitric oxide reduces tumor cell adhesion to isolated rat postcapillary venules. Clin. Exp. Metastasis, 14: 335-343, 1996.[Medline]
-
Doi K., Akaike T., Horie H., Noguchi Y., Fujii S., Beppu T., Ogawa M., Maeda H. Excessive production of nitric oxide in rat solid tumor and its implication in rapid tumor growth. Cancer (Phila.), 77: 1598-1604, 1996.[Medline]
-
Kubrina L. N., Caldwell W. S., Mordvintcev P. I., Malenkova I. V., Vanin A. F. EPR evidence for nitric oxide production from guanidino nitrogens of
Top
ABSTRACT
INTRODUCTION
