(CANCER RESEARCH 46, 3105-3110, June 19861
Influence of Glucose Concentration on Growth and Formation of Necrosis in
Spheroids Derived from a Human Bladder Cancer Cell Line'
Ian F. Tannock2 and Irma Kopelyan
Departmentsof Medicine and Medical Biophysics,Ontario Cancer Institute and University of Toronto, Toronto, Ontario, Canada M4X 1K9
ABSTRACF
Spheroids derived from the human bladder cancer cell line MGH-U1
were initiated in spinner culture and then transferred to multiwell plates
which contained medium with varying concentrations of glucose and
pyruvate. Growth of individual spheroids was monitored, and at different
times after transfer spheroids were sectioned and the diameter of the
necrotic center and the thickness of the viable nm were determined. In
normal mediumcontaining glucose(1 g/liter, 5.5 mM), spheroiddiameter
increasedlinearly with time, growing from —400@m
to —@12®
gsmin 8
days, and most spheroids did not develop central necrosis. Increase in
glucoseconcentration up to 5 g/Iiter had no effect on spheroid growth.
Lower glucoseconcentrationdecreasedthe rate of spheroid growth, but
large effects wereobservedonly at glucoseconcentrationlower than 100
mg/liter. Spheroids developedcentral necrosisat 2—4
days after transfer
to glucose-deficient medium, and the diameter of the necrotic center
nutrient in a blood vessel and the process of cell death in
neighboring tumor cells. Spheroids are multicellular aggregates
oftumor cells (7, 8), which may be propagated in tissue culture.
Spheroids resemble tumor nodules since they may develop
central necrosis, have a decreasing gradient of cell proliferation
from the periphery, and form tight junctions between cells.
They provide a useful model for studying the penetration of
nutrients into solid tissue and their influence on central necrosis
(9—12),although they cannot model effects that are influenced
by blood flow in vivo. In the present study we have varied the
concentration of glucose and pyruvate in medium surrounding
spheroids derived from a human bladder cancer cell line to
determine the influence ofthese metabolites on spheroid growth
and formation of necrosis.
increased more rapidly than the diameter of the spheroid. There was an
approximately linear relationship betweenthickness of the viable rim in
5—6-dayspheroids and glucose concentration
in the range of zero (rim
thickness, -@450gim) to 500 mg/liter (rim thickness, @-400
aim). The
presenceor absenceof pyruvate (110 mgfliter) in the medium had no
effect on spheroid growth or formation of necrosis.Theseresults suggest
that limited penetration of glucose may be one of the factors that
contribute to cell death in solid tumors.
INTRODUCI1ON
Cell death and necrosis occur commonly in solid tumors, but
the mechanisms which lead to cell death remain largely un
known (1). In several types of human and rodent tumors the
edge of a region of necrosis has been observed to be parallel to
a neighboring blood vessel, leading to “cords―
of tumor cells
with a central blood vessel and surrounding necrosis; alterna
tively, tumor nodules may occur with surrounding stroma and
central necrosis (2—4).These structures imply that limited dif
fusion of essential nutrients from tumor blood vessels or the
accumulation of toxic catabolites may be involved in the caus
ation of cell death. The distance between a blood vessel and a
neighboring edge of necrotic tissue has been reported to be in
the range of 100—200tim. These distances are consistent with
the expected diffusion distances for oxygen (2—5)and suggest
that hypoxia may be involved in the causation ofnecrosis. Also,
tumors were observed to grow more slowly and there was a
decrease in the radius of tumor cords when animals were placed
in a hypoxic environment (6). However, cells in tissue culture
can survive hypoxia for long periods if other nutrients are
present at physiological concentration. Thus limited supply of
several other nutrients probably contributes to the process of
necrosis.
The complex anatomy of tumor blood vessels and the physi
ological regulation ofcomponents of the blood make it difficult
to study the relationship between the concentration ofany given
Received6/I 1/85; revised I 1/6/85, 2/1 3/86; accepted2/17/86.
The costs of publication of this article were defrayed in part by the payment
of pagecharges.This articlemustthereforebeherebymarkedadvertisement
in
accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.
â€S̃upported
by
research
grants
CA
29526
and
CA
36913
from
the
National
Cancer Institute, NIH, and by a grant from the National Cancer Institute of
Canada.
2To whom requestsfor reprints should be addressed.
MATERIALS
AND METhODS
Cell Culture. The MGH-U1 cell line was derived originally from a
patient with bladder cancer and is of the same origin as cell lines
designated U and T24 (13) which have been shown to express an
activated Ha-ras oncogene(14). The cell line was kindly provided to us
by Dr. G. Prout and colleagues, Massachusetts General Hospital,
Boston, MA. MGH-Ul cells will grow as a monolayer on plastic and
will also form colonies in agar and xenografts in immunodeprived mice
(15). We have confirmed the identity of the cells by the presence of
marker chromosomes in their karyotype and by isoenzyme analysis.
MGH-U1
cells are maintained
as a monolayer culture in 75-cm2
plastic flasks containing a-medium + 10% FCS.3 a-medium contains
glucose (I g/liter) and pyruvate (110 mg/liter)
but no other sugars; the
medium also contains amino acidswhich can be metabolized to provide
substratesfor glycolysis. Cells are subcultured at weekly intervals fol
lowing detachment using 0.05% trypsin/0.02% EDTA. Cultures are
reestablishedfrom frozen stock at 3-month intervals.
In some experiments westudied the growth ofmonolayers in glucose
deficient medium. Medium waspreparedwithout glucoseand pyruvate,
and appropriate amounts of glucose were added to this medium. Fetal
calf serum was dialyzed against 40:1 NaCl:KCI with phosphate buffer
to remove glucose
and other small molecules.
The same number of
MGH-Ul cells (-@10@)
were seeded into multiple flasks containing
glucose-deficient medium + 10% DFCS with varying concentration of
glucose.Flasks were selectedat random at 2-day intervals, and the cells
were detached and counted using a hemocytometer. In most experi
ments the medium was changed at 2-day intervals in the remaining
flasks. Cells were also plated in Petri dishes in a-medium + FCS, and
plating efficiency was assessedby counting stained colonies 10 days
later.
Samplesof medium were analyzedfor glucoseconcentration using a
commercial kit (Sigma Chemical Co., St. Louis, Mo). This method is
based on the conversion of glucose to glucose 6-phosphate in the
presenceof ATP and hexokinase, followed by reduction of NADP to
NADPH when glucose 6-phosphate
is oxidized to 6-phosphogluconic
acid in the presence of glucose-6-phosphate
dehydrogenase. NADPH
is then measured by absorption at 340 nm using a spectrophotometer.
Recently some of our samples have been analyzed with a Beckman
glucose analyzer (Beckman Instruments, Brea, CA) which measures
oxygen consumption in the presenceof glucoseand glucoseoxidase.
Spheroids.The derivation of MGH-U1 spheroids has beendescribed
elsewhere (16). All experiments described in the present paper were
3 The
abbreviations
used
are:
FCS,
fetal
calf
serum;
DFCS,
dialyzed
fetal
serum.
3105
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1986 American Association for Cancer Research.
calf
GLUCOSE AND CELL DEATH IN SPHEROIDS
performed on a sublimeof MGH-Ul cells that was established after
passageof the parental line through spheroids. This sublime,designated
MGH-Ul/OCI-l, will routinely form spheroids when placed in spinner
culture, whereasthis is more rarely observedfor the parental cells.
Spheroids usedin the presentexperiments were grown at 31T in air
in 250-mI
spinner
flasks containing
a-medium
+ 10% FCS + 442-
hydroxyethyl)-1-piperazineethanesulfonic acid buffer (20 mM/liter),
using a rotor speed of 130 rpm. After 4—6days, spheroids of about 400
@m
diameter were selected from the spinners and pipetted one per well
into 24-well multiwell plates. The wells contained a 0.5-mI underlayer
of I % agar diluted in a-medium without glucose and pyruvate, with 1.5
ml liquid medium above. Liquid medium consisted of a-medium with
or without pyruvate and with varying concentration of glucose, plus
10% DFCS. In most experiments the liquid medium wasaspirated and
.@
replaced at 2-day intervals. The aspirated medium was pooled for
estimation of glucose concentration as described above. The multiwell
plates were maintained at 37C in 95% air/5% CO2. This experimental
design allowed measurements to be made on individual and uniform
spheroids, without the extensive shedding of cells that we and others
have observed from the surface of spheroids in spinner culture (17).
The growth of spheroids was studied in multiple experiments. At 2—
3-day intervals each multiwell plate was removed briefly from the
incubator and individual spheroids were observed through a transparent
cover using an inverted microscope. The maximum and minimum
diameters of each spheroid were recorded using an eyepiecereticule,
and the geometric mean diameter wascalculated. (Most spheroidswere
close to spherical; examination
after removal from the agar underlayer
did not reveal any systematic deviation from spherical symmetry.) The
mean diameter of multiple spheroids was then plotted against time to
generatea growth curve.
In order to study the relationship betweenglucoseconcentration and
formation of necrosis, spheroids were pipetted from the multiwell
dishes at varying times after exposure to normal or glucose-deficient
medium. The spheroids were exposed to mercurochrome to facilitate
their recognition. In most experiments spheroids were sectioned using
a cryostat after rapid freezing. Serial 5-sm sections were cut, followed
by fixation of the slides in 95% ethanol and staining with hematoxylin
and eosin. The largest cross-sections of spheroids (i.e., those through
their center) were examined under the microscope and the maximum
and minimum diameters of both the spheroid and its necrotic center
were recorded. There was no net shrinkage or expansion of spheroid
sections using this method. More recently, spheroids have been fixed
overnight in Bouin's solution, then embeddedsequentially in 1.5%agar
and paraffin, followed by serial sectioning and staining. The latter
procedure is more time consuming and leads to approximately 20%
shrinkage in linear dimension,
but gives higher quality sections.
Growth of Cells in Monolayer. The growth of MGH-Ul
in varying concentration
different concentrations of glucose plus 10% DFCS. Each point was obtained
from hemocytometer counts after detachment of cells in a single flask. Medium
was replenished at 2-day intervals. Qualitatively
similar results were obtained in
several replicate experiments, except that glucose, 20 mg/liter, was sometimes
unable to support growth of the cells.
The rate of depletion of glucose from medium by cells in
monolayer is shown in Fig. 2A. Glucose consumption decreased
from about 0.7 mg/105 cells/day to about 0.1 mg/b5 cells/day
during exponential growth of cells under control conditions;
glucose consumption was lower (@‘.-0.04
mg/b5 cells/day) for
cells growing at a slower rate in medium
containing
only 100-
mg/liter amounts of glucose.
Growth of Spheroids. The growth of spheroids in varying
RESULTS
as a monolayer
Doys
Fig. I. Growth of MGH-UI cellsas a monolayerin a-mediumcontaining
cells
of glucose was studied
concentrations of glucose was studied in multiple experiments.
In medium with a glucose concentration of 1 g/liter, the sphe
roid diameter increased linearly with time at a rate of about
100 @sm/day(Fig. 3A). Growth ofspheroids in medium contain
ing increased
concentrations
of glucose in the range of 2—10g/
in several experiments. Under control conditions (i.e., in amedium with glucose, 1 g/liter), cells grew exponentially with
liter was similar to that under control conditions (data not
shown). Spheroids grew quite rapidly in media containing 100—
a doubling time of 16—24h after an initial lag period of 2—3 1000 mg/liter glucose, but there was progressive slowing of
days (Fig. 1). Increase of the glucose concentration in the range
growth at glucose concentration below 100 mg/liter (Fig. 3A).
of 2—lO-g/liter had minimal effects on cell growth. Cells were However, unlike cells in monolayer, spheroids grew slowly in
unable to grow in the absence of glucose, regardless of the the total absence of glucose. The mean plating efficiency of
presence ofpyruvate. Growth ofcells was slower than in control
cells dissociated from spheroids was about 60% and did not
flasks at glucose concentrations
in the range of 20—100 mg/
depend on the glucose concentration in the medium during
liter, although the magnitude ofthis effect varied among exper
spheroid growth.
iments. Growth of cells was slower than in control flasks at an
The growth of spheroids in medium containing glucose, 10,
initial glucose concentration of 100 mg/liter regardless of 20, or 100 mg/liter, with or without pyruvate (1 10 mg/liter) is
whether the medium was left unchanged or was replenished at shown in Fig. 3B. There was no difference in spheroid growth
2-day intervals.
The mean plating
efficiency
of trypan
blue
in the presence
excluding cells was close to I 00% and was independent of
glucose concentration about 20 mg/liter. However, plating ef
ficiency fell rapidly for cells placed in medium containing no
glucose.
or absence of pyruvate.
In most of the above experiments the liquid medium in
multiwells was replenished at 2-day intervals, although there
was no effect on spheroid growth rate in one experiment where
the medium was left unchanged (Fig. 3A). Representative values
3106
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1986 American Association for Cancer Research.
GLUCOSE
AND CELL DEATH IN SPHEROIDS
the Beckman glucose analyzer indicate that this method is
accurate, to ±2%(SD), and this accuracy leads to a confidence
range for glucose consumption of about 0—0.02mg glucose/
spheroid/day. Glucose consumption by MGH-U1 spheroids is
about 10-fold lower than values reported by Li (10) for 9L rat
brain tumor spheroids grown in suspension culture. Since the
amount of glucose removed from medium (Fig. 2B) was of the
same order as the uncertainty in the most accurate detection
procedure available to us, we were unable to@investigate the
relationship between glucose consumption by spheroids, their
volume, and the concentration of glucose in medium. Since
spheroids could grow slowly for a limited time in medium
without glucose, the lower limit of consumption of exogenous
glucose is zero. Spheroids may, however, consume glucose that
was trapped between cells on transfer from initial spinner
culture and/or glucose that is released from cells which die
within spheroids.
Other investigators have studied the influence of glucose
concentration in the medium when spheroids were maintained
in spinner culture (10—12).We therefore compared spheroid
growth and glucose consumption by single spheroids in multi
wells with those placed in new spinners after their initial estab
0
Days
Fig. 2. A, measured glucose concentration in flasks containing MGH-U 1 cell
monolayers. Cells (1—2
x 10') were seededinitially in 20 ml medium containing
lishment. This comparison was made in medium containing
a
range of glucose concentration. The number of spheroids was
eitherglucose,I g/liter (V,0, @,
0), or glucose,100mg/liter(, •).
Themedium adjusted to provide approximately one spheroid per 2 ml me
was not replenished. Different symbols represent different experiments. B, mea
suredglucoseconcentrationin multiwells(0, •)
or spinners(0, U) containing dium (i.e., 100 spheroids per spinner), the same ratio of sphe
MGH-Ul spheroids (approximately 1 spheroid/2 ml medium). The medium was roids to medium as in multiwells. Under these conditions
replenishedat 2-dayintervals(0, 0) or left unchanged
(N,•).
spheroids grew slightly more rapidly in spinner culture, al
though the reverse is observed at a higher concentration of
of glucose concentration in the medium are shown in Fig. 2B. spheroids in the spinners. Examination of the medium from
Despite the strong influence of glucose concentration in the spinners showed the presence of numerous single cells as well
medium on spheroid growth, the data of Fig. 2B indicate a very as aggregates and small spheroids which were formed by cells
shed from spheroids; this was not found in multiwells. The
low rate of consumption of exogenous glucose. From the mea
number of cells shed from larger spheroids (@1 mm diameter)
sured change in glucose concentration in media taken before
was estimated by allowing spheroids to settle under gravity,
and after spheroid incubation in multiwells, the best estimate
followed by counting of cells in medium that was aspirated and
of glucose consumption by spheroids growing in media contain
trypsinized. The number of such cells in three spinner flasks
ing glucose, 200—1000 mg/liter, is about 0.01 mg glucose!
spheroid/day (equivalent to @0.02mg/105 cells/day). Replicate
containing @-100spheroids ranged from 2.8—7.3x 106 cells
measurements of glucose concentration on a single sample with after 6 days in culture. The total extraction of glucose in
Fig.3. Increase in mean diameter of
MGH-U1 spheroids in multiwells containing
medium with different concentrations of glu
cose and pyruvate. In A, wells contained
110
mg/literpyruvateandtheindicatedconcentra
tion of glucose.Mediawerereplenishedat 2day intervals(0, L@,
0, V) or left intact (U
glucose,20 mg/liter). In B, wells contained
pyruvate(110 mg/liter) or no pyruvatewith
varying glucoseconcentrations (10, 20, or 100
mg/liter) asindicatedin the key.Meandiam
eter ±SE (bars) is indicated for at least 10
spheroids.
Days
3107
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1986 American Association for Cancer Research.
GLUCOSEAND CELL DEATH IN SPHEROIDS
spinners was greater than in multiwells at an initial glucose
concentration of ‘@@500
mg/liter or 1 g/liter, but not at 100 mg/
liter (Fig. 2B). This increase in glucose extraction could be
accounted for by consumption by single cells and clumps at a
rate of about 0.2 mg/105 cells/day, well within the range of
glucose consumption found for MGH-U1 cells in monolayer.
Thus we did not find evidence for major differences between
spheroid growth and glucose extraction by MGH-U1 spheroids
under stirred conditions in spinner culture and unstirred con
ditions in multiwells.
Formation of Necrosis. Spheroids which grew in medium
containing a concentration ofglucose in the range of0.5—10g/
was due to a balance between peripheral cell production and
accumulation of necrotic tissue centrally.
DISCUSSION
We have demonstrated that the glucose concentration in the
medium influences the growth rate and formation of necrosis
in spheroids derived from the MGH-U1 human bladder cancer
cell line. Our results are in qualitative agreement with the work
of Freyer, Mueller-Klieser and their colleagues (12, 18) who
found that glucose concentration was a major determinant of
necrosis formation in EMT6/Ro spheroids of murine origin.
liter grew to a diameter of 1.0—1.2mm without central necrosis,
An unexpected finding was the failure ofpyruvate to influence
or with minimal necrosis occurring in the largest spheroids
spheroid growth or formation of necrosis. Pyruvate is known
(Fig. 4A). Spheroids which grew in lower concentration of to cross cell membranes (19); pyruvate is also formed normally
glucose developed central necrosis (Fig. 4B) and the relation
from glucose by glycolysis and is a substrate for production of
ship between the thickness of the viable rim and glucose con
ATP by the Krebs cycle. Failure of pyruvate to influence sphe
centration in the medium is shown in Fig. 5. These data were roid growth or morphology may suggest limited involvement
obtained for spheroids that were sectioned at 5—6days after
of the Krebs cycle in the energy metabolism of spheroid cells.
transfer to glucose-deficient medium and demonstrate an ap
It was also surprising to find that cells in spheroids, but not in
proximately linear relationship between rim thickness and glu
monolayer, could proliferate in the total absence of exogenous
cose concentration up to 500 mg/liter. This relationship was glucose. Viable cells may have utilized some carbohydrates that
not influenced by the presence or absence of pyruvate at a were released from damaged cells; this might occur more readily
concentration of 110 mg/liter.
when cells are in close proximity in spheroids as compared to
The rate offormation ofnecrosis was estimated by sectioning
when they are dispersed in a monolayer. Small amounts of
ofspheroids at varying intervals after placing them into glucose
glucose may also have been transferred within spheroids when
deficient medium (Fig. 6). Necrosis was first observed at 3—4 they were placed initially into multiwells.
days, and the diameter of the necrotic center then increased at
Li (10, 11) studied the relationship between glucose con
a rate that was faster than the rate of increase in spheroid
sumption and concentration for 9L rat brain tumor cells in
diameter, so that the thickness of the viable rim decreased.
culture and then used this relationship to solve the diffusion
These data were then used to generate the relationship between
equation to obtain the expected distribution of glucose concen
the volume of morphologically viable tissue and time in various
tration in spheroids derived from the same cell line. He found
types of media (Fig. 7). The growth of viable tissue in control
that the relationship between the radius of the necrotic zone
spheroids deviated only slightly from exponential. Even in the and the radius of 9L spheroids could be fitted by assuming that
total absence of glucose there was initial expansion of the viable necrosis occurred when the glucose concentration fell to 6 x
compartment, and growth of spheroids was not due simply to
iO-@mg/mi. His calculations required the assumption, however,
the accumulation of dead cells. However, in glucose-deficient
that glucose consumption was independent of the concentration
of oxygen and other key metabolites, and this seems rather
media the volume of morphologically viable tissue increased
unlikely.
toward a maximal value, so that later growth in larger spheroids
;.;@:@@.::7.
@
..-.
@
..@
@
..
., .
@
‘:
:-.
:
:
@‘
. .
...-..
.
•
.•:
.
.:
;:
.@ .:.
.
• •
‘@
‘S......
@
@
,@.
•.,@4.r%•.J.••,@
I .
.
‘@
•
Fig. 4. Cross-sections
of MGH-Ul spheroids.Left, spheroidgrownfor 4 daysin multiwellsundercontrol conditions(glucose,1 g/liter) containsno central
necrosis. Right, spheroid of similar size grown for 7 days in medium containing glucose, 100 mg/liter, contains a large region of central necrosis.
3108
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1986 American Association for Cancer Research.
GLUCOSE AND CELL DEATH IN SPHEROIDS
I
500
,
Spheroids
E
@@400
@
300
a
Relationship between thickness
of viable rim and glucose
concentration in 5-6 day
MGH- UI spheroids
@200
00
0
200
400
600
000
800
Glucoseconcentrationtmg/iiter)
Fig.5. Relationshipbetweenthicknessof theviablerim andconcentrationof
glucosefor MGH-Ul spheroidsafter 5—6
daysin multiwellscontainingthe
specifiedconcentrationof glucose.Differentsymbolsrepresentdifferentexperi
ments.
@
b)
4
Doys
Fig. 7. Estimated volume ofviable tissue (logarithmic axis) at different times
Necrotic A
Centre /
I.
(linear axis) during growth of MGH-Ul spheroidsin multiwellscontaining
different concentrations of glucose. Volume was calculated from estimates of
diameterof spheroidsandtheir necroticcenters(Figs.3, 5, and6) by assuming
sphericalsymmetry.
control conditions glucose, (1 g/liter) in muitiwells, or at low
concentration in spinner culture, their volume increases almost
exponentially to a large size (1.2-mm diameter). Many of these
Fig.6. Increasein meandiameterofMGH-Ul spheroidsandoftheir necrotic spheroids had no central necrosis while others developed mm
centers during growth in multiwells containing (a) glucose, 20 mg/liter, and no
imal necrosis at a diameter >1 mm. Most other types of
pyruvate. or (b) glucose, 100 mg/liter, and pyruvate, 110 mg/liter. Each point
spheroids, including those of human origin (20), have been
representsa separatespheroid, except mean values ±SE (bars) are indicated for
>6 spheroidsat sometime intervals.
reported to develop central necrosis at a mean diameter of <700
zm. MGH-U1 spheroids develop central necrosis at -@500 sm
The MGH-Ul cells used in the present study differed in two diameter when grown under more crowded conditions in spin
important ways from the 9L cells used by Li (10, 11): (a) ner culture, despite replenishment of the medium at 2-day
glucose consumption by spheroids was about 10-fold lower (and intervals. Thus formation of necrosis in spheroids may depend
because of this could not be measured accurately); and (b) as much on the culture conditions (i.e., number of spheroids
glucose consumption by cells in culture varied with cellular
per unit of medium) as on properties of the cell line. When the
concentration and was greater than for the same cells in sphe
concentration of spheroids is high it may be necessary to renew
roids. Also, MGH-U1 spheroids grew slowly and had a viable the medium at least twice daily (12) to avoid depletion of
rim of cells even in the total absence of exogenous glucose.
glucose and other essential nutrients. With the lower concen
Thus our data on spheroid growth and formation of necrosis
tration of spheroids (one spheroid per 2 ml medium) used in
the present study, medium changes were probably unnecessary
cannot be fitted by a simple diffusion-consumption model which
considers glucose alone. Rather, our results demonstrate that
(multiwells)or were required at less frequent intervals in spinner
lack of glucose is an important factor which contributes to culture because of consumption of glucose and other metabo
formation of necrosis but that the process of cell death is lites by cells shed from the surface of spheroids.
complex and depends on other factors as well. Absence of
We elected to study individual spheroids grown in multiwells
glucose does not cause cell death in the presence of adequate
on an agar underlayer since this system has the following
concentration of other key metabolites (e.g., oxygen, amino
advantages as compared to spinner culture: (a) it allows deter
acids such as glutamine, etc.) but probably does so when these
mination of the rate of growth of individual spheroids rather
other metabolites are also depleted.
than that of representative spheroids obtained at random from
Spheroids described in the present paper appear to show the population; (b) the size of spheroids remains more uniform
some major differences as compared to those described by and does not depend critically on factors such as speed of the
others. When individual MGH-Ul spheroids are placed under
rotor, (c) shedding of cells from the surface is minimal or does
0
I
2
3
4
5
0
Days
I
2
3
4
3109
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1986 American Association for Cancer Research.
GLUCOSE
AND CELL DEATH IN SPHEROIDS
not occur, removing the potential artifact of consumption of
glucose and other metabolites by single cells and by smaller
spheroids that form from them in suspension.
Possible disadvantages of the multiwell system are that lack
of stirring might lead to depletion of nutrient metabolites
around the surface of the spheroids and that a lower concentra
tion of metabolites in the agar underlayer might lead to asym
metry in the vertical plane. We found no evidence for systematic
departures from spherical symmetry either on gross examina
tion after removing spheroids from the agar underlayer, or after
histological sectioning. The area of contact between spheroid
and agar is small, and glucose and other metabolites probably
equilibrate rapidly between agar and liquid medium. It is pos
sible that lack of stirring led to a lower consumption of glucose
by spheroids in multiwells, but the higher glucose extraction
measured in spinners (Fig. 2B) could be explained entirely by
consumption of glucose by cells (and their progeny) that had
been shed from the surface of spheroids.
REFERENCES
I. Wyllie, A. H. The biology ofcell death in tumors. Anticancer Res., 5: 131—
136,1985.
2. Thomlinson, R. H., and Gray, L. H. The histological structure of some
human lung cancers and the possible implications for radiotherapy. Br. J.
Cancer, 9: 539—549,
1955.
3. Tannock, I. F. The relation betweencell proliferation and the vascularsystem
in a transplanted mouse mammary tumour. Br. J. Cancer, 22: 258—273,
1968.
4.
5.
6.
7.
8.
9.
In the current study we found an effect of glucose concentra
of necrosis may occur in spheroids growing in
medium containing a higher concentration ofglucose if the PO2
and/or levels of other important metabolites such as glutamine
are reduced to a concentration similar to those in tumor blood
vessels (9, 12). Spheroids will provide an important model for
understanding the relationship between the distribution of glu
cose, oxygen, and other metabolites and the formation of ne
crosis in solid tumors.
ACKNOWLEDGMENtS
analysis
on MGH-Ul
V.,
Haselton,
P.
5.,
and
Buckley,
C.
H.
Tumourcords
in
52
human
10. Li, C. K. N. The glucosedistribution in 9L rat brain multicell tumor spheroids
and its effect on cell necrosis.Cancer (Phila.), 50: 2066—2073,
1982.
11. Li, C. K. N. The role ofglucose in the growth of9L multicell tumor spheroids.
Cancer(Phila.),50:2074—2078,
1982.
12. Mueller-Klieser, W., Freyer, J. P., and Sutherland, R. M. Evidence for a
major role of glucose in controlling development of necrosis in EMT6/Ro
multicell tumor spheroids. In: H. I. Bicher and D. F. Bruley (eds.), Oxygen
Transport to Tissue, pp. 487—495.New York: Plenum Publishing Corp.,
1983.
13. O'Toole, C. M., Povey, S., Hepburn, P., and Franks, L. M. Identity of some
human bladder cancercell lines. Nature (Lond.), 301: 429—430,1983.
14. Parada,L. F., Tabin, C. T., Shih, C., and Weinberg, R. A. Human EJ bladder
carcinoma oncogeneis homologue of Harvey sarcomavirus rca gene.Nature
(Lond.), 297: 474—478,
1982.
15. Kovnat, A., Armitage, M., and Tannock, I. F. Xenografts of human bladder
cancerin immune-deprived
mice.CancerRes.,42:3696—3703,
1982.
16. Erlichman, C., and Vidgen, D. Cytotoxicity ofAdriamycin in MGH-U1 cells
grown as monolayer cultures, spheroids, and xenografts in immune-deprived
mice.CancerRes.,44:5369—5375,
1984.
17. Landry, J., Freyer, J. P., and Sutherland, R. M. Shedding of mitotic cells
from the surfaceof multicell spheroids during growth. J. Cell. Physiol., 106:
23—32,
1981.
18. Freyer, J. P., and Sutherland, R. M. The role of glucose in regulating
quiescentcell populationsin EMT6/Ro spheroids.Radiat.Res.,91: 342,
1982.
19. Halestrap, A. P. Transport of pyruvate, NAD, lactate into human erythro
cytes.Evidencefor the involvementof the chloridecarrier,anda chloride
independent carrier. Biochem. J., 156: 193—207,
1976.
20. CaIsson, J., Nilsson, K., Westermark, B., Ponten, J., Sundstrom, C., Lars
son, E., Bergh, J., Pahlman, S., Busch, C., and Collins, V. P. Formation and
growthof multicellularspheroidsof humanorigin. Int. J. Cancer,31: 523—
We thank Dr. J. Trent, University of Arizona, for performing kary
otype
J.
bronchial and cervical squamouscell carcinomas: inferencesfor their cellular
kinetics and radiobiology. Br. J. Cancer, 51: 407—413,1985.
Tannock, I. F. Oxygen diffusion and the distribution of cellular radiosensi
tivity in tumours. Br. J. Radiol., 45: 515—524,
1972.
Tannock, I. F., and Steel,G. G. Tumor growth and cell kinetics in chronically
hypoxic animals. J. NatI. Cancer Inst., 45: 123—133,
1970.
Moskowitz, M., Amborski, G. F., and Wicker, C. H. Structure development
in aggregens.Nature (Lond.), 211: 1047—1049,
1966.
Sutherland, R. M., McCredie, J. A., and Inch, W. R. Growth of multicell
spheroids in tissueculture as a model of nodular carcinomas.J. Nail. Cancer
Inst., 46: 113—120,
1971.
Franko, A. J., and Sutherland, R. M. Oxygen diffusion distance and devel
opmentof necrosisin multicellspheroids.Radiat.Res.,79:439—453,
1979.
tion on formation of necrosis at concentration in the medium
below 500 mg/liter. Tumors have been found to utilize about
30% of the glucose supplied to them (21) and the above value
is probably close to the lower limit of glucose concentration
which maybe found in blood draining a tumor. Lower values of
glucose concentration might occur in tumor capillaries. Also,
we have reported the effects of glucose concentration only in
the presence of normal or high concentration of oxygen and
other metabolites. The PO2 in tumor blood vessels is likely to
be closer to 40 mm Hg than to 150 mm Hg (95% air), the value
to which the medium was exposed in the current experiments.
Radiobiological experiments do not suggest a significant hy
poxic fraction when our spheroids are irradiated in air, in
contrast to results for most experimental tumors. It is probable
that formation
Moore,
cells.
533,1983.
21. Gullino, P. M., Grantham, F. H., and Courtney, A. H. Glucose consumption
by transplanted tumors in vivo. Cancer Res., 27: 1051—1040,
1967.
3110
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1986 American Association for Cancer Research.
Influence of Glucose Concentration on Growth and Formation of
Necrosis in Spheroids Derived from a Human Bladder Cancer
Cell Line
Ian F. Tannock and Irina Kopelyan
Cancer Res 1986;46:3105-3110.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/46/6/3105
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1986 American Association for Cancer Research.