.. I " In ‘I-‘I;
3- - .‘.. II t
'- .. I -- {It}. ”1):? fly»; . 1.
1 "‘1"? ii; \‘IvI‘LJ. I.
"T I I

'. “12“. I.... v.
I_- "A J

h;

' I3,h
L'fi ,

3%;

”If, I,
“(I ' I“ . :0 . 1
.' I'm 1.,‘ I . . ' , ‘4 ,L‘A “
.II[ [I ‘.I ':\\ ”1'11”,“ .1" (‘1' .1, :"1 1:10;...” 31;?" 31"-..1 .“I l . .(HII ." ‘ I; I: '4'? i); L
1‘” - .Il {' ' I " ; '1 . :hr()“"l'j '6 1 I I £141“ '0' ;f\'llll'1\’| ‘1]. '1 h" '-'I- _‘ “1".” _‘. .1'.‘ . _ . I‘ '
113‘. {I I1“ ".21": _ . I "It :ullu ‘..‘
..") I m 10H“‘ ”mg“; *' 1'1. . 1. ‘1“, i H“ ."n‘ff “t, I m . IJ;‘1“1I.'.:-." . 1:?“ fl ‘1 . K
\II J I I? ,' “1:”: , _ ‘4 , . W: . ' ‘ '.
k! zjflidp ' f l ' ' 1’ U 3' ‘. l‘ I

‘ " WW1; 'I-1 r. ,,"1
I n ”"1311" JP 1 2| _ l" - .111? ‘l' WWII”); :1”:
~ - 11-11 ‘ ”-11"*.--'1"I'*) ' .I; 1'11 12L 4%:qu
,.- I1 1" 1“ a: 4 .',o' 1'1” ‘ ‘5‘! [Jill-I111 ”‘1! h 1‘11“” 1.}
”‘3' I

11‘1'11'Wi1 fIr'-!}:I'I1J‘YIII"AE.~.' “1 1111/1; Il'-""I(:3[J"I""1\ 1 111’_'-;'.3'5}:.'I.§{[":~"I"1'IIEY‘-11 -‘ . 1 11,1 :1. . I
I 1. - i. . '.“,I '-,':1,"‘I. I" «‘-I'~ ." II . . . .
'. .1431»; 9.1.1.. 1 . I ”*1- " W: 11¢ " .I , .. . . ~1 I»-
. i',111..\\:j‘1.j'1'$ 11'1"::".1‘1 . I‘ . ' 'l ' “ Rammfi Hannah? V n.1,? 1“") .II‘.‘ 1;; '51 ’ o 'I
- . - -, .. ._.
' '1'!!! “:15 I '1'

Jur;
'--,1- 11:19,}. .;.

413.11..
‘LligJ \‘0 (I‘ll; .

'l

. ‘1" ‘1"...411‘,‘ ‘:

“ . {WW/:6), 1 ‘ 1 - "1'dj. .

ugh “15):: £ “1;"; , | 1 “37%;.

1”- "IJ <11 £4331
WV II

OII

II 'It
I"- 4.
I.III

 

 

$5151 I
f My .1501“, 1.
I: 'u ’le “ ’IVII’W

‘ I} i'. V.

 

 

';'fC1-3 ,.
_ ‘1‘
1“ '1 ;.‘ '
‘ :1 r _
’ . fin *
r! t r '. 13‘
..- '2‘ )i? 3.3
T 1-
E1, :1} ‘t ' X ’
,1 E
‘wmat... M A
g.
M

This is to certify that the
thesis entitled
MOLECULAR ANALYSIS OF DENSITY-DEPENDENT INHIBITION
OF GROWTH IN 3T3 FIBROBLASTS

presented by
Peter Allan Steck

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Biochemistry

 

Major professor

% { (UL—q
A Y

Date May 21, 1981

0-7639

    
 
 
  
    
 

4L ovum: mas:
Ln. \ 25¢ per day per item
- l
, j (E "“3.“ f RETURNING LIBRARY MATERIALS:
,\ .‘7‘:."',' 4 Place in book return to remove
\ “W” 4 charge from circulation records

   
 
   
     

MOLECULAR ANALYSIS OF DENSITY-DEPENDENT INHIBITION
OF GROWTH IN 3T3 FIBROBLASTS

BY

Peter Allan Steck

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Biochemistry

1981

ABSTRACT

MOLECULAR ANALYSIS OF DENSITY-DEPENDENT INHIBITION
OF GROWTH IN 3T3 FIBROBLASTS

By

Peter Allan Steck

The mechanisms of density-dependent inhibition of growth in
cultured fibroblasts were studied using the 3T3 cell line. Treatment
of sparse, proliferating cultures of 3T3 cells (target cells) with
medium conditioned by exposure to density-inhibited 3T3 cultures
resulted in an inhibition of growth and DNA synthesis in the target
cells when compared to parallel cultures treated with unconditioned
medium (UCM). This differential effect of conditioned medium (CM) and
UCM on target cells was demonstrated using three assay systems: (a)
assessment of total cell number; (b) measurement of [3H]thymidine
incorporated into acid-precipitable DNA; and (c) determination of the

percentage of radioactively labeled nuclei in individual cells after

incorporation of [3H]thymidine. The difference in the total
incorporation of [3H]thymidine in CM-treated and UCM-treated cells
was reflected by a difference in the percent of labeled cells. There
was no difference in the average number of grains per labeled cell in
the two cultures.

The preparation and assay of the growth inhibitory activity in CM
were optimized by varying several parameters of incubation with source
cells during the period of conditioning and with target cells during
the testing of the activity. In addition, it was determined that the
inhibitory effects of CM on DNA synthesis and cell proliferation
effects were reversible upon replacement with fresh growth medium. The
data also indicated that the CM prepared from 3T3 source cells had
similar inhibitory effects on normal fibroblasts from other sources,
but its effects on virally-transformed fibroblasts and non-fibroblast
cell lines were much less pronounced.

The growth inhibitory activity was fractionated from CM by
ammonium sulfate precipitation and gel filtration, yielding one
fraction that was 35-fold enriched in specific bio-activity. Analysis
of the chemical and biological properties of this highly active
fraction indicated that: (a) it consists of two polypeptides (Mr =
10,000 and Mr = 13,000), (b) it is an endogenous cell product,
synthesized by the 3T3 cells and secreted or shed into the medium; and
(c) it is not cytotoxic and its effects on target cells are reversible.
These results suggest that we have succeeded in partially purifying
from CM an endogenous growth regulatory factor that nay play a role in

density-dependent inhibition of growth in cultured fibroblasts. Ne

propose the term Fibroblast Growth Regulator to describe this class of
molecules.

The binding of Fibroblast Growth Regulator to 3T3 cells has been
analyzed using [35SJmethionine-labeled inhibitor. The key features
of this interaction are: (a) there are approximately 3-4 x 105
binding sites per cell; (b) more ligand is bound at 37° than at 4°;
(c) on a per cell basis, approximately the same amount of ligand is
bound at high cell density as at low cell density; (d) there is no
evidence of cellular heterogeneity in the binding of the inhibitor; and
(e) the binding can be partially inhibited by calf serum, as well as
several purified growth stimulatory factors. These results suggest
that a system has been developed for analyzing the antagonistic actions
of growth stimulatory and inhibitory factors in the regulation of

density-dependent inhibition of growth on a molecular level.

TO MY PARENTS

ii

ACKNOWLEDGEMENTS

To Dr. John L. Wang, I would like to express my sincere
appreciation for his advice and encouragement during my entire graduate
training.

I would like to express my gratitude to Patricia Voss, John
Blenis, and Justina Calamia for their collaborative efforts and
thoughts that aided in the progress of this work.

I would also like to thank all the members of Dr. Wang's

laboratory for their friendship and advice in many matters.

m

TABLE OF CONTENTS

LIST OF TABLES. . . . . . . . . . . . . ....... . ..... v
LIST OF FIGURES . . . . . . . . .'. . . . . . . . . . . . . . . . vi
LIST OF ABBREVIATIONS . . . . . . . . . . . . . . . . . a . . . .viii
INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . 1
CHAPTER

I. LITERATURE REVIEW . . . . . . . . . . . . . . . . . . . . . 6

Growth Stimulatory Factors. . . . . . . . . . . . . . . . . 7
Endogenous Growth Inhibitors. . . . . . . . . . . . . . . . 10
Growth Control in Vitro . . . . . . . . . . . . . . . . . 13
Cell to Cell Contact Inhibition . . . . . . . . . . . . . . 15
Limitation of Growth Factors. . . . . . . . . . . . . . . . 19
Accumulation of Growth Inhibitors . . . . . . . . . . . . . 22
II. ASSAYS OF CELL PROLIFERATION AND DEMONSTRATION OF A
GROWTH INHIBITORY ACTIVITY. . . . . . . . . . . . . . . . . 27

Materi a1 5 and Methads I O O O O O O O O O O O O O O O O O O 28
Resu‘ts O O O O C O O O O O O O O O O O O O O O O O O O O O 38
DTSCUSSTO". C O O O O O O O O O O O O O O O O C O O O O O O 72

III. MOLECULAR PROPERTIES OF A FRACTION ENRICHED IN GROWTH
INHIBITORY ACTIVITY 0 O O O O C C O C C O C O O O O O O O O 77

Materials and Methods . . . . . . . . . . . . . . . . . . . 78
Resu1ts O O O O O O O O 0 O O O O O O O O 0 O O O O O O O O 84
DESCUSSion. 0 O 0 O O O O O O O O O O O O O O O O O O O O 0 105

IV. BINDING INTERACTIONS OF A GROWTH INHIBITORY ACTIVITY
WITH TARGET CELLS . . . .

O O O O O O O O O O O O O O O O O 110

Materials and Methods . . . . . . . . . . . . . . . . . . . 111
Resu‘ts O O O O O O O O O O O O O O O O O O O O O O O O '0 O 115
D1. scusslion. O O O O O 0 O O O O O O 0 O O O O O O O O O O O 135

BIBLIOGRAPHY C O O O O O O O O O O O O O O O O O O O O O O O O O 140

iv

LIST OF TABLES
Table Page

I Growth stimulatory factors. . . . . . . . . . . . . . . . . 8

II DNA synthesis in sparse cultures of 3T3 cells
co-cultured with density-inhibited cells. . . . . . . . . . 43

III The effects of Conditioned and Unconditioned Medium
on DNA synthesis and viability of target cells. . . . . . . 50

IV The cellular specificity of the effects of Conditioned
Medium on growth inhibition . . . . . . . . . . . . . . . . 59

V The effects of cell number, cell density, and the
volume of the medium on the accumulation of growth
inhibitory activity in CM . . . . . . . . . . . . . . . . . 65

VI The effects of ammonium sulfate fractionation and
dialysis on the growth inhibitory activity in

serum'free COHthTOflEd Medium 0 e e e e e e o e e e e e e o 67

VII Activity, specific activity, and recovery of fibroblast

growth regulator at various stages of the fractionation
protOCOIO O O O O O O O O O O O 0 O O O O O O O O O O O O 0 86
VIII Reversibility of the effects of CM(SF) and FGR-s. . . . . .104
IX Specificity of the binding of FGR-s to target cells . . . .118

X Quantitation of radioactivity of cell bound FGR-s . . . . .124

LIST OF FIGURES
Figure Page

1 Schematic representation of the cell cycle . . . . . . . . . 2

2 Schematic representation of three hypotheses of density-
dependent inhibition of growth . . . . . . . . . . . . . . . 16

3 Schematic diagram of the protocol used to prepare
Conditioned Medium and Unconditioned Medium. . . . . . . . . 31

4 Growth curve of 3T3 fibroblasts. . . . . . . ...... . . 4O

5 Schematic diagram of co-cultivation of sparse and
dense ce1150 O O O O O O O O I O O O O I O O O O O O O O O O 42

6 Kinetics of [3H]thymidine incorporation into 3T3 cells
treated with Conditioned Medium and Unconditioned Medium . . 45

7 Autoradiography grain count distribution . . . . . . . . . . 47

8 Kinetics of increase in cell density of 3T3 cells treated
with Conditioned Medium and Unconditioned Medium . . . . . . 49

9 Photograph of representative 3T3 cells . . . . . . . . . . . 52
IO Reversibility of the effects of Conditioned Medium . . . . . 54

11 The effect of varying the density of target cells on
the inhibitory activity of Conditioned Medium. . . . . . . . 56

12 The effect of varying the time of exposure of Conditioned
Medium to the target cells on the inhibitory activity. . . . 58

13 The effect of varying source cell parameters for
preparation of Conditioned Medium. . . . . . . . . . . . . . 62

14 The effect of varying the length of exposure of Growth
Medium to source cells . . . . . . . . . . . . . . . . . . . 64

15 Chromatography of the ammonium sulfate precipitate

fraction of serum-free Conditioned Medium on a column
Of sephadex 6-15 0 O O O O O O O O O I O O O I O O O O O O O 71

vi

Figure

16

17
18
19

20

21
22
23
24
25
26
27
28
29
30
31

32

Chromatography of the ammonium sulfate precipitate
fraction of serum-free Conditioned Medium on a column of
sephadex 6-50 0 O 0 O O O O 0 O O O O O O O O l O O O O O O

Polyacrylamide gel electrophoresis in $05 of serum-free
Conditioned Medium and derived components . . . . . . . . .

Polyacrylamide gel electrophoresis in $05 of 35S-labeled
component C O O O O O O O O O O O O O O O O O O , O O O O O O

Polyacrylamide gel electrophoresis in $05 of 125I-labeled
component C O O O O O O I O O O O O O O O O O O O O O O O O

Chromatography of the growth inhibitory activity derived from

several different preparations of serum-free Conditioned
Medium on columns of Sephadex G-SO. . . . . . . . . . . . .

The effect of pronase treatment on the growth inhibitory
activity of soluble Fibroblast Growth Regulator . . . . . .

The effect of various serum-free Conditioned Medium
derived components on the viability of 3T3 cells. . . . . .

Cell count distributions for colonies of 3T3 cultures
treated with serum-free Conditioned Medium derived
components. 0 O O O O O O I O O O O O 0 O O O O O O O 0 O 0

Kinetics of the binding of 353-1abe1ed FGR-s fraction .

Concentration dependence of the binding of 35S-labeled
FGR-s fraction. . O . ’ O . . . . . . . . . . . . C C O . . .

The binding of 35S-labeled FGR-s as a function of the
3T3 target cell density . . . . . . . . . . . . . . . . . .

Representative photograph showing the binding of FGR-s on
target cells as analyzed by autoradiography . . . . . . . .

Polyacrylamide gel electrophoresis analysis of FGR-s before
and after binding to 3T3 target cells . . . . . . . . . . .

Inhibition of the binding of FGR-s to target cells by calf
serum and by bovine serum albumin . . . . . . . . . . . . .

The effects of serum concentration on the growth inhibitory
activity of serum-free Conditioned Medium . . . . . . . . .

Chromatographic analysis of FGR-s on a column of Sephadex

Inhibition of the binding of 355-1abe1ed FGR-s to
target cells by growth stimulatory factors. . . . . . . . .

vii

Page

88

92

'94

95

97

99

100

102

116

121

123

126

128

129

130

131

134

FBS
DME-FBS
CM

UCM
CM(SF)
UCM(SF)
PBS
TCA
SDS
BSA
FGR-s
MSA
EGF

FGF
PDGF

LIST OF ABBREVIATIONS

Mitosis

Gap one

DNA Synthesis

Gap two

Dulbecco Modified Eagle's Medium

Fetal Bovine Serum

DME Supplemented with FBS to a Final Concentration of 10%
Conditioned Medium

Unconditioned Medium

Conditioned Medium Prepared in the Absence of Serum
Unconditioned Medium Prepared in the Absence of Serum
Phosphate Buffered Saline

Trichloroacetic Acid

Sodium Dodecyl Sulfate

Bovine Serum Albumin

Soluble-Fibroblast Growth Regulator

Multiplication Stimulating Activity

Epidermal Growth Factor

Fibroblast Growth Factor

Platelet-derived Growth Factor

viii

INTRODUCTION

The life of a cell is usually described in terms of what is
classically called the cell cycle, and reflects the orderly recurrence
of growth, DNA synthesis and cell division (1). This cycle has been
divided into four unequal phases which delineate nuclear events (Figure
1); (a) mitosis (M), (b) gap one (61), (c) DNA synthesis (S), and (d)
gap two (62) (1,2,3). DNA synthesis and mitosis insure the orderly
reproduction and inheritance of genetic material, respectively. These
two periods are experimentally defined. In contrast, 61 is between H
and S, and G2 which is between S and M, are periods of metabolic
activity, but are not characterized by any single biochemical event.

The time required to traverse the cell cycle is highly variable.

For most animal cells, this variance occurs in GI, whereas Gz,
M, and S are relatively constant (3,4). Therefore, the control of
passage through the cell cycle is postulated to occur in G1. This
control results in the modulation two alternative cellular states,
quiescent and proliferative.

Proliferative cells are cells that are actively progressing
through the cell cycle and more importantly the total number of cells

is increasing. In contrast, quiescent cells are defined as resting

 

Figure 1. Schematic representation of cell cycle.

cells whose population is constant and the vast majority of the cells
are_in 61 phase. These cells remain viable for long times and with
the proper stimulation are able to enter S phase (5). .

Three basic models have been proposed for the control of a normal
cell's passage through the cell cycle. Pardee (6) suggests that there
is a restriction point (R) in 61 at which the cell either progresses
through another round of the cell cycle or shifts into a resting state
GO (Figure 1). Alternatively, Prescott (7) postulates that a cell
progresses through a continuous but expandable GI phase. The third
approach first proposed by Smith and Martin (8), redefines the cell
cycle to consist of two states, A and 8. State A consists of most of
61 and is of indeterminate length, while state B contains S, G2, M
and parts of 61 and is of predetermined length. The shift between
these two states occurs by a transitional probability that is
determined by environmental conditions and cell type (9). Henceforth,
I will refer to the model proposed by Pardee (Figure 1) in discussions
of the cell cycle, even though any of the models could be used. In any
case, the biochemical control mechanism that mediates the transfer
between the proliferative and quiescent modes is unknown.

Growth control has been suggested to be regulated by the delicate
balance between the opposing forces of endogenous growth stimulators
and inhibitors (1,10). Growth stimulators are substances which induce
resting cells to enter into the proliferative cycle, Go to 61
transition of Figure 1. A large number of endogenous growth
stimulators have been isolated and chemically characterized (11). In

contrast, growth inhibitors cause the opposite transition, shifting

cells from a proliferative to a quiescent state (Gl to Go

transition of Figure 1). Much less progress has been made in studies
on growth inhibitory factors when compared to the status of growth
stimulators. A major problem has been a lack of a well-defined
experimental system to study their molecular nature and mechanism of
action. The use of cell culture has allowed certain aspects of growth
control to be more rigorously investigated, while still maintaining
relevance to in vivo homeostasis (10,12).

One good example of proliferative to quiescent transition in
tissue culture cells is the observation that the growth rate of normal
cells in culture decreases as the number of cells approaches a critical
cell density forming a monolayer. This phenomenon has been termed
density-dependent inhibition of growth (13). Cells can be released
from this growth restriction by increasing the available growth area
(14) or by decreasing the local cell density (15). Thus, this
capability of normal cultured cells to maintain a relatively constant
equilibrium density indicates an Operative growth control mechanism in
Vitro.

The system we have chosen for our studies of growth control
utilizes the tissue cultured cell line 3T3 fibroblasts, obtained
originally from mouse embryos (16). Besides the obvious advantage of
working with a tissue culture system, 3T3 cells possess the property of
being highly responsive to density-dependent inhibition of growth.
There is some experimental evidence to suggest that these cells produce
specific growth inhibitors (17-20). The work described in this thesis
has been directed at the isolation and characterization of the

molecules responsible for growth inhibitory activity found in medium

conditioned by exposure to 3T3 fibroblasts. To this end, the first
major objective was to define a reliable assay system for the detection
of growth modulators. Once the ability to determine the presence of
the inhibitory activity was demonstrated, the next step involved the
identification of the responsible factors. The successful attainment
of this goal permitted investigations directed at the biochemical
mechanisms of growth control. The results presented in this thesis
suggests that a well defined system has been developed using the 3T3
fibroblast cell line to study growth regulation in terms of (a) the
isolation and molecular characterization of growth inhibitory factors,
and (b) the mechanism by which these inhibitory factors exert their

effects.

LITERATURE REVIEW

Normal tissues and organs maintain a uniform mass throughout the
adult life of an animal. This implies that the rate of cell renewal in
tissues is in balance with the rate of cell death and that the number
of functional cells is directly proportional to the functional demand.
It has been proposed that endogenous growth regulators, some
stimulatory, others inhibitory, play a key role in the modulation
between the quiescent and proliferative modes of existance for a cell
(1,21). Induction of cellular proliferation may be stimulated under
the proper environmental conditions by a wide variety of mitogenic
factors (growth factors) which have received much attention. In
contrast, little work has been done on endogenous growth inhibitors.

It should be emphasized that the control of cellular proliferation
resides in the balance between the two opposing forces of mitogenic and
inhibitory factors. Therefore, in the analysis of growth control by
endogenous inhibitors one must be cognizant of the activities of growth

stimulators.

Growth StimulatoryiFactors

 

A growth factor is defined by its biological activity of
stimulating the multiplication of cells in a nutritionally complete
environment (11,22). In addition, growth factors should express their
biological effects at concentrations considered to be near their
physiological level, which is generally at very low concentrations.
Although a large number of mitogenic preparations have been reported,
only a few molecules enjoy the status that their chemical nature has
been well characterized and thus can rightfully be denoted as growth
factors. These are summarized in Table I.

Growth factors are often classified as to their principal source,
since little is known about the specificity of target cells especially
in vivo. The major sources of growth factors are conveniently
classified into three categories: (1) factors from the bloodstream, (2)
factors from tissues, and (3) factors from cultured cells (Table I).
The placement of a growth factor in one of the three categories does
not exclude it from being present in other locations. For example,
multiplication stimulating activity (MSA) has been isolated from both
serum (the bloodstream) and from cultured cells (35). Furthermore,
many growth factors appear to be related.

MSA, nonsuppressible insulin-like activity and somatomedin A,
three mitogenic polypeptides extracted from plasma or serum, all
produce insulin-like activity when they are bound to target cells
(25,11). Interestingly, these three growth factors have molecular
weights in the range of 5000 to 7000 (Table 1). Moreover, these growth
factors have been shown to bind to the MSA plasma membrane receptor of
chicken embryo fibroblasts (39,40). This receptor, however, was

demonstrated to be distinct from the insulin receptor.

.ES—vms mumpaeou a—_o:o_u.c»=: :.
m—puu amuse» ocau—au mama.» :. co—aucmuwmoca cu—:__uu ouu_::_um o» once o>—uuouuu a:u.=.z+
.auu =.uco_aoumoca . gun «Locum» ac—uopsepum ace—cu .mmu mm_—mu cu>__
use coon soc» uo>pcov cocoa» m=_uu—=s.um =o.uau—_qpu_=s .4mu-<mz ”couuou zuzocm awe—nocn.w
. .uuu "cognac guzocm :a—ca>o .uuo "Logan; guzocm o>cm= .uuz "League :uzoca _u5cmu_am
.mmu «Au_>_uuu mc—uo.:e_um copuuu_—a—u_=a .<mz 4coauum saxoca uo>—Louiuupmua_a .muom
«ou.uaoa o_a=—om au_>—uoo ox—_i=._=m:_ u—apnmucaaamco: .mi<4_mz "use can: mcopuow>ocaa<n

 

 

"mwuocuoom
“any mama—eeee.c om. a~dua
m_—ou coo—am
Aem.om. :eeeax beam coo.o~-coc.oe emu
m__au
“may mama—eeee.a oce.ofi-cco.m ugo-<wz emcee—=9 u:mm.~
Acme _e\ue Ha.o mume_eoee_a oco.n~ Lac
Anny. _e\ae o~ ee.ee>o cco.m_-coo.o~ Lac xeeu.=a.e
ANMV _e\ae _.o meeeaaz oeo.e~ . Luz
_.._eeb.eu
Adn.cmv _a\ae m.o meme—eoee.u occ.e . Lem wee—m xga__.xeea=m
AGNV meme—eeee_a cco.m <m= aseum
.m~.-v mume_beee.a ao¢.n_ Lena mob—oue_e
Mo~ mama—neeeuaeoeea coo.m~ e.uu_eaoeebsem
mN meme—eeee_a cao.s u e_eeeeaeeem
Ae~.m~v mama—eeee_a ace.“ m-<3_mz . sam._e
+umca m__au a=a_m:
moccacomoa saawcpx comes» cu—auu_az ago—unca—muo oucaom Leno:

 

msouuum auouapaawum garage ._ u—nah

Furthermore, platelet-derived growth factor (PDGF) isolated from
serum and fibroblast growth factor (FGF) isolated from tissues have
been demonstrated to have interchangeable, stimulating properties on
certain cells (41). Both of these growth factors have molecular
weights of 13,000 (Table I). These results have led to the suggestion
that there are classes of growth factors, with similar properties, that
exhibit different cellular specificity in vivo (22).

The attainment of highly purified growth factors has allowed
investigations focused at the biochemical mechanisms of growth
stimulation. Epidermal growth factor is the best characterized growth
factor in terms of its interaction with target cells and the cell's
subsequent biological reSponses. Specific, saturable receptors for EGF
have been demonstrated on a wide variety of cultured cells and tissues
(42). Human fibroblasts were estimated to possess 40,000-100,000
binding sites per cell with an apparent dissociation constant of
2-4 x 10'10 M (43,44). Furthermore, using a photo-reactive
derivative of EGF, Das et al. (45,46) showed that one cell surface
receptor for EGF may be a glycoprotein of molecular weight 190,000.
Recent studies have been concerned with the fate of the bound EGF.

Carpenter and Cohen (47) have presented data which suggests that
after the initial binding of EGF to its receptor, the EGF receptor
complex is internalized and subsequently degraded in lysosomes. It is
not known at which stage(s) the biochemical signal that eventually
results in cellular proliferation occurs. Recent experiments have
demonstrated that EGF bound to isolated membranes of human fibroblasts
resulted in increased protein phosphorylation (48). Of particular

interest was the finding of higher levels of phosphotyrosine (49).

10

Although it is too early to assess the biological role of increased
phosphotyrosine levels in mitogenic stimulation, it is an intriguing
result, considering the importance of phosphotyrosine in viral
transformation (50,51) and other metabolic pathways (52). The
elucidation of the biochemical mechanisms involved in the induction of
cellular proliferation will yield a better understanding of the
regulation of growth and indicate possible points of interaction with

growth inhibitors.

Endogenous Growth Inhibitors

 

Weiss and Kavanu (53) proposed the existence of negative feedback
inhibitors that would act in concert with growth factors to control
cellular mitosis in vivo. Bullough and Laurence (54) applied this
concept to their studies of growth control of various epidermal cells
and suggested that the negative feedback mitotic inhibitors be called
chalones. A chalone was defined as a growth inhibitor that is
synthesized by cells within a tissue to act only on cells of the same
tissue type. Furthermore, a chalone should lack species specificity in
exhibiting its effects. Finally, the action of a chalone must not be
toxic and consequently its effects are reversible.

The existence of chalones was first suggested on the basis of
wound healing experiments on ear epidermis in mice (54). The number of
cellular mitoses was determined in the intact epidermis on the opposite
side of an ear that was previously wounded. An increased mitotic rate
was found across from the center of wound. If growth stimulators were
released from cells adjacent to the wound, then more mitotic activity

would be expected in the intact epidermis opposite the edges of the

11

wound. However, if the increase in the number of mitotic cells was due
to the release of cells from control of an inhibitor normally present,
then the highest mitotic activity would be found opposite the center
of the wound and indeed this was the observed result. This result
supported the concept of endogenous mitotic inhibitors or chalones. A
number of studies have since focused on the isolation and
characterization of the putative epidermal chalone; no purified
molecule has yet been reported (55-62).

The extensive studies on the epidermal chalones has also
stimulated interest in a wide variety of other cell types.
Investigations have suggested that endogenous growth inhibitors, each
with their own cellular specificity, are integral components in the
growth control of all tissues. The systems studied have included
granulocytes (55), lymphocytes (56), fibroblasts (57), liver cells
(58), mammary cells (59), melanocytes (60) and a number of other
tissues (21,61). To date, however, no chalone has been purified to a
sufficient extent that the existence of these molecules can be
unequivocally documented.

Two molecules, however, have been isolated and chemically
characterized that contain growth inhibitory activity, interferon and a
factor. The growth inhibitory activity associated with these molecules
was determined after these proteins had been characterized in regard to
other biological activities: antiviral activity of interferon and
sexual conjugation in yeast by a factor. Nevertheless, the potent
affects of these two molecules on cell growth and proliferation
strongly suggest that they are the best presently available paradigm

systems for the study of growth inhibitors.

12

Interferon is defined as a proteinaceous factor which exerts
antiviral activity, at least in homologous cells, through cellular
metabolic processes involving synthesis of both RNA and protein (62).
There are three distinct sources that produce interferon (63);
leukocytes (IFN-a), fibroblasts (IFNee) and immune T lymphocytes
(IFN—v). Further studies on this family of glycoproteins
(Mr = 18,000-36,000) have shown a multiplicity of biological effects.
The activities attributed to purified interferon include modulation of
humoral and cell mediated immune responses (64), inhibition of cell
motility (65), alteration of cell surface (66,67) and cytoskeleton
organization (68,69) and inhibition of cellular proliferation (70-72).
The ability of interferOn to inhibit growth was originally observed in
the delayed appearance, decreased size and fewer nunber of tumors in
experimental animals which were treated with interferon followed by
oncogenic viruses or transplantable tumors (73) when compared to
animals not treated with interferon. Many studies have since focused
on the anti-tumor effects of interferon.

Gresser and coworkers have shown that normal cell division is also
inhibited by interferon preparations (71,72) or by electrophoretically
pure interferon (74). Recent investigations by Tamm and coworkers (70)
demonstrated that the anti-proliferation activity of interferon on
cultured human fibroblasts could be detected at a minimum dose of
approximately 30 ng/ml. This concentration of interferon is of the
same order of magnitude as that required of growth factors (Table I).
However, even at high doses the rate of cell division was reduced by
only 50-60%. The anti-proliferative activity was shown not to be

attributable to toxicity, even though its effects are not readily

13

reversible. Thus interferon is the first well characterized mammalian
endogenous growth inhibitor which can serve as a model system and may
aid in the elucidation of mechanisms of growth control.

The other well characterized endogenous growth inhibitor is a
factor. Haploid cells of yeast Saccharomyces cerevisiae can exist as
one of two alternative mating types which are referred to as 'a' and
'a'. Diploid zygotes can be formed by conjugation of yeast with the
opposite mating type. It has been shown that a type cells excrete a
diffusible substance which inhibits cellular division of 'a' type cells
(75). This substance facilitates the mating between cell types and has
been called a factor. a Factor blocks the passage of an 'a' cell
through the cell cycle at a particular location in 61 (76).
Furthermore, this substance has been found to be a polypeptide of
molecular weight 1,400 daltons whose amino acid sequence has been
determined (77). Therefore, this molecule also can be classified as a
definitive endogenous growth inhibitor and will undoubtedly aid in

studies of growth control.

Growth Control In Vitro

Abercombie and Heaysman (78) described the phenomenon of cessation
of cell movement by fibroblasts in culture when the cells come into
contact with each other. This inhibition of movement could be reversed
by the removal of neighboring cells or by the increaSe of growth area.
The ability of similar cell types to act as barriers to each other was
termed "contact inhibition". Studies on neoplastic transformation of
cultured cells showed a good correlation between the loss of contact

inhibition of cell movement and transformation in vivo (79).

14

Therefore, contact inhibition was extended to include inhibition of
cell division (80). However, the role of cell to cell contact in
growth inhibition was questioned by experiments with freshly isolated
fibroblasts which showed some continued growth and cell division even
after the initial cellular monolayer was formed (81). Also,
multilayers of cells could be produced by continuous addition of fresh
medium to the culture (82). For these reasons, Stoker and Ruben (13)
proposed to call inhibition of cell division in crowded cultures
“density-dependent inhibition“ of cell growth.

A large number of factors have been shown to effect the rate of
cellular proliferation in vitro. The range of these factors include:
(a) inorganic ions, particularly calcium and magnesium (83); (b)
nutrients, such as amino acids and carbohydrates (84); (c) vitamins and
cofactors (85); (d) growth stimulators, exemplified by epidermal growth
factor, and many other found in serum (11); and (e) growth inhibitory
factors (21). In addition, it has been demonstrated that there is a
tight correlation between cell shape and DNA synthesis in cultured
normal cells (86). A basic question in study of density-dependent
inhibition of growth is which factors or combinations of factors serve
to regulate growth control. An important advancement in the
elucidation of the molecular mechanisms of density-dependent inhibition
of growth was the development of a normal cell line which is highly
responsive to this form of growth control.

Todaro and Green (16) established the 3T3 cell line which is
characterized by extreme sensitivity to density—dependent regulation of
growth. The 3T3 cells originally isolated from mouse embryo

fibroblasts possess the distinctive property that they maintain a

15

constant density in a monolayer, even with frequent refeeding.
Moreover, disruption of the monolayer allows the quiescent cells to
resume a proliferative state until a new monolayer is achieved (15).

On the basis of analyses performed in a number of laboratories,
the mechanisms responsible for this phenomenon of density-dependent
inhibition of growth have been arbitrarily classified under three major
hypotheses (Figure 2). The first hypothesis proposes that cell to cell
contact is a necessary requirement for growth inhibition (Figure 2a).

A second hypothesis suggests that depletion of growth factors in the
medium surrounding the cells causes cessation of growth (Figure 2b).
The third hypothesis postulates that soluble growth inhibitors are
released into the medium and that these inhibitory factors accumulate
to a sufficiently high concentration such that they act on
proliferating cells (Figure 2c). These three theories are not mutually
exclusive, nor do they represent all the possibilities. However, for

this review I shall examine each hypothesis on an independent basis.

Cell to Cell Contact Inhibition

 

The observation that most normal cells in culture cease to
proliferate once a monolayer of cells is achieved prompted the
hypothesis that cell to cell contact is an important condition for
density-dependent growth control. Fisher and Yeh (87) reported that
DNA synthesis in 3T3 cells grown in colonies occurred predominantly at
the periphery of the colony, while the cells at the center did not
incorporate [3HJthymidine. In contrast, CHL-l cells, which do not
exhibit density-dependent regulation of growth, exhibited an even
distribution of DNA synthesis throughout the colony cultures.

16

Hypothesis Proliferatlng Cells Quiescent Cello

Cell to Cell Contact 02%

e e
3
ll Depletion of Growth Factors 0.0—0 0
8

III Soluble Inhibitors ‘ I

Figure 2. Schematic representation of three hypotheses to account for
density-dependent inhibition of growth in cultured
fibroblasts. The symbols 5 and 1 denote growth stimulatory
factors and growth inhibitory factors, respectively.

17

Moreover, Schutz and Mora (88) demonstrated that dense quiescent 3T3
cells on one side of a thin membrane filter did not affect the growth
of sparse 3T3 cells on the opposite side of the filter. These results
imply that cell to cell contact, and not just close proximity of
neighboring cells, is required for growth inhibition.

Density-dependent inhibition of growth is also dramatically
demonstrated in the phenomenon of "wound healing“. Nhen 3T3 cells are
mechanically removed from a confluent monolayer, the cells adjacent to
the newly devoided area (the wound) migrate into the area and
proliferate until the monolayer is restored (15). The proliferation of
the sparse cells in the wound occurs while crowded cells in the same
medium remain quiescent. This observation, that cells surrounded by
their neighbors are quiescent while cells not encircled by other cells
proliferate, supports the hypothesis of cell to cell contact inhibition
of growth.

Analogous to wound healing studies, Kolodny and Gross (89) showed
that increasing the available growth surface by the removal of physical
barriers, such as glass beads, resulted in cell division to fill in the
open area. This method has the advantage over wound healing
experiments that the newly available growth surface did not contain
cells, which upon their removal may release growth stimulating factors,
especially to the cells next to the open area.

A question which arises from the previous results is whether
neighboring cells present a physical barrier to continued growth or if
a specific chemical signal is transmitted between cells resulting in
inhibition of growth and cell division. Eagle and Levine (90) reported

that human diploid fibroblasts and ammion cell lines respond to

18

density-dependent growth regulation in culture. However, neither cell
type could inhibit the proliferation of the other when the cells were
cultured together. In addition, Martz and Steinberg (91) investigated
the effect of cell to cell contact by time lapse cinematography in
growing 3T3 cultures as the cells approached their saturation density.
The studies demonstrated that simple visual cell to cell contact did
not produce immediate inhibition of cell division, but rather a delay
of about one generation time was often required. All of these results
suggest that specific interactions at the cell surface between cells
may play an important role in growth regulation.

Recent studies by Nhittenberger and Glaser (92) showed that a
plasma membrane-enriched fraction of 3T3 cells inhibited DNA synthesis
in sparse 3T3 cultures but not in their SV 40 virally transformed
counterparts. The growth inhibitory activity of the surface membrane
preparation was extracted by the detergent octylglucoside, and the
inhibitory activity was shown to be reversible, nondialyzable and heat
labile (93). Furthermore, the inhibitory effects of the membranes were
demonstrated not to be due to a decreased diffusion around the sparse
3T3 cells or binding of growth factors to the isolated membranes,
thereby depleting the medium of growth factors (94,95). In analogy to
density inhibited 3T3 cells the effects of the membranes could be
overcome by the addition of mitogenic serum factors and resulted in
renewed DNA synthesis. These results support the hypothesis that cell
to cell contact is an important factor in growth regulation.

The role of the cell surface in mechanisms of growth control is
also suggested by experiments where the cell surface was extracted with

urea. The presence of urea in the culture medium significantly reduces

19

cellular response to contact inhibition of movement (96). Also, Natraj
and Datta (97) have shown that an inhibitor of DNA synthesis can be
extracted from 3T3 cells by treatment with 0.2 M urea in PBS. The
inhibitory activity of this factor was abolished by glycosylation with
N-acetyl-D-glucosamine; the inactive factor could in turn be converted
to the active form by treatment with N-acetyl-e-D-glucosaminidase.
Furthermore, the conversion from active to inactive forms was
correlated with increased uptake of low molecular weight nutrients
which occurs upon release of density-dependent inhibition (98,99).
Therefore, these investigations suggested that these cell surface
factors which inhibit normal cell proliferation may be the same
molecules that are responsible for density-dependent growth

regulation.

Limitation of Growth Factors

 

Two broad classes of substances are necessary to achieve cell
growth and division in culture, nutrients and growth stimulators (1).
Low molecular weight nutrients consist of metabolic substrates or
cofactors and can be supplied in synthetic medium. Under restrictive
concentrations of some nutrients, such as glutamine or isoleucine
starvation, cells will shift specifically into a Go resting state
(100). This form of growth control, however, is generally considered
not to be operative in vivo or in vitro, except under extreme
conditions (101). In contrast, growth stimulators (growth factors) are
thought to be essential components in the control of cellular division.
Almost all cultured cells require growth factors in order to

proliferate (11).

20

The question arises, what specific roles do growth factors play in
density-dependent inhibition of growth. Holley and Kierman (102)
demonstrated that the final saturation density of 3T3 cells is directly
proportional to the final serum concentration in the medium.

Therefore, they proposed that depletion of essential growth factors in
serum was responsible for density-dependent inhibition of growth. In
support of this notion, Dulbecco and Elkington (103) varied the
available growth area (dish size) while keeping the number of cells
plated and the medium volume constant, and then determined the final
density of 3T3 cells. The results suggested that the final cell number
per dish was most dependent on serum limitation. Furthermore, growth
could be restored in resting 3T3 cultures by addition of isolated
growth factors (104). These studies were extended by Smets (105) who
showed that used medium could be restored to its original growth
supporting capacity by dialysis against fresh mediun or by addition of
low molecular weight metabolites or serum. Finally, used medium
removed from density-inhibited 3T3 cultures, and placed onto
proliferating 3T3 cultures supported the growth of these cells poorly
unless fresh serum was added (102).

In contrast, other investigators have reported that used medium
supported the growth of sparse 3T3 cells, but at a lower rate than
fresh medium (15). This difference has been reconciled by the fact
that the two groups used different exposure times to make the used
medium (91). This conflict exemplifies a basic problem in the study of
growth control, in that with the magnitude of the number of factors
involved, a well defined system must be used to study density-dependent

inhibition of growth.

21

Dulbecco (106) investigating density-dependent inhibition of 3T3
cells suggested that the regulation of growth was controlled by at
least two components, serun and topoinhibition. The function of serun
was assigned into four general classes: (1) to initiate DNA synthesis;
(2) to supply factors necessary for mitosis; (3) to aid in survival of
cells in culture; and (4) to supply factors to overcome topoinhibition.
In contrast, topoinhibition was defined as the phenomenon of decreased
growth and proliferation as cells accumulate extensive cell to cell
contact. The neighboring cells were thought to act as a barrier to
continued growth of normal cells in a monolayer. Several possible
mechanisms have been suggested to account for the influence crowded
cells exert on each other to regulate cellular growth and division.

Stoker (107) proposed that growth limitation of cells in a
monolayer was due to failure of growth factors to penetrate a boundary
diffusion layer around the cells. Sparse cells were suggested to have
a greater exposed surface area to the bulk medium than dense cells in a
monolayer. Disturbing the microbarrier layer by pumping medium over
the monolayer (107) or by shaking a culture of 3T3 cells (108) resulted
in increased [3H]thymidine incorporation into the effected cells.

This hypothesis was also used to explain topoinhibition observed in
wound healing experiments.

Cells at the edge of the wounded area were proposed to possess a
greater surface area, in the form of a vertical component, that are
exposed to the bulk medium (107). Further experiments showed that
factors involved in the initiation of DNA synthesis acted more
effectively on sparse cells than on dense cells (109). Moreover,

recent experiments by Holley and coworkers (110) demonstrated that low

22

density monkey epithelial cells (BSC-I cells) bind more of the potent
mitogen epidermal growth factor than high density cells. These results
suggest that cells next to the wounded area may interact more
effectively with serum factors resulting in cellular proliferation,
while dense cells bind fewer growth factors and remain quiescent.

The diffusion barrier layer could also be used to explain the
results of Thrash and Cunningham (111). They reinvestigated
density-dependent inhibition of growth using 3T3 cells under varying
growth conditions. They demonstrated that the growth area (cell
density) was the limiting factor when the ratio of medium volume per
cell was high. However, under conditions where the medium volume per
cell was low, growth was limited by medium components. The dependence
on medium components could be overcome if the cells were refed daily.
Frequently refed cells, however, showed decreased proliferation when
the growth area became limited. These results suggest that neighboring
cells proximity is an important factor in density-dependent inhibition

of growth.

Accumulation of Growth Inhibitors

 

An alternative view to depletion of growth factors resulting in
inhibition of cell growth and division is accumulation of growth
inhibitors in the bulk medium. Green and coworkers (15) found that
medium exposed to density-inhibited 3T3 cells when removed and placed
on cultures of sparse 3T3 cells, decreased the cells growth rate. The
inhibitory effects could not be attributed to accumulation of toxic
substances. The inhibitory activity of medium exposed to 3T3 cells was

further documented by Rheinwald and Green (17,18). In attempting to

23

establish keratinizing epidermal cells in culture from human skin
explants or a mouse teratoma cell line, the investigators found that
skin or teratoma fibroblasts consistently overgrew the epidermal cells.
However, the growth of these fibroblasts was much suppressed by the
presence of a feeder layer of lethally irradiated 3T3 cells or by
medium previously exposed to 3T3 cells. The growth of the epidermal
cells was not affected by this treatment (17,18). These results are
consistent with an accumulation of a growth inhibitor; however, they
are also consistent with selective depletion of fibroblast growth
factors.

More recently, Harel et al. (19) have shown that phosphate
metabolism and cell growth in sparse cultures of 3T3 cells were
inhibited when they shared the same medium as dense cultures.

Phosphate metabolism, which is correlated to the growth status of 3T3
cells (98), was demonstrated to decrease within twenty minutes of the
beginning of the co-culture, if the cultures were shaken. Depletion of
growth factors is not expected to occur in this short time interval.

Furthermore, Canagaratna and Riley (112) investigated the patterns
of nuclear incorporation of radioactive thymidine in 3T3 cultures where
the local cell density varied from sparse cells to dense cells in the
same medium. They determined that DNA synthesis was critically
dependent on local cell density. More detailed analysis of the data
showed that there was an inverse relationship between local cell
density and the proportion of labeled cells. Moreover, this
density-dependent regulation of DNA synthesis was exhibited in
relatively sparse cultures well before the onset of cell to cell

contact (113). Although these results add more evidence to support the

24

role of putative soluble growth regulators, their existence still needs
to be documented.

Yeh and Fisher (20) proposed the existence of a growth inhibitory
factor that aids in maintenance of density-dependent inhibition of
growth in 3T3 cells. The basis for their study was the observation
that when fresh medium containing serum was added to cultures of
density-inhibited 3T3 cells a synchronous wave of RNA synthesis was
observed in most cells, followed later by a wave of DNA synthesis and
mitosis in a few cells (15). However, if medium previously exposed to
3T3 cells was added to the culture of density-inhibited cells, no wave
of RNA synthesis or DNA synthesis was observed. Now, if serum growth
factors were solely responsible for the stimulation of mitogenic
activity, then the addition of serun free medium should not elicit a
wave of RNA synthesis or DNA synthesis. However, a wave of RNA
synthesis was observed in cultures to which serum free medium had been
added. These results suggested that the used medium contained a
soluble growth inhibitor. Further, characterization of this inhibitory
activity found it to be of low molecular weight (Mr = 400-900) and
heat stable. These results suggest that density-dependent inhibition
of growth in cultured 3T3 cells may be mediated, at least in part, by
the accumulation of endogenous growth inhibitors in the medium.

In addition to the 3T3 cell system, growth inhibitory factors have
also been studied in a number of other cultured cells. Lipkin and
Knecht have reported that a glycoprotein of molecular weight 160,000
can be isolated from culture medium of contact-inhibited hamster
melanocytes. This factor can reversibly inhibit growth in vitro of a

broad spectrum of malignant and normal cell types of ectodermal,

25

mesodermal and endodermal organs (114-116). Moreover, an
electrophoretically identical protein has been found in culture of
contact-inhibited fibroblasts and epidermal cells.

Lower molecular weight growth inhibitors have been reported to
have been obtained from a number of other cell types. Analysis of
growth properties of the human fibroblast cell line, NIe38, suggests
that decreased growth at high cell density is due to the accumulation
of an inhibitor (117,118). This inhibitory activity appears to be
associated with a 30,000 molecular weight protein which is both trypsin
sensitive and heat labile (119). More recent experiments have
suggested, however, that the apparent molecular weight of 30,000 may be
due to the complexing of the inhibitor with RNA molecules and that the
actual value of the molecular weight may be considerably lower (120).

A growth inhibitor with a similar molecular weight range
(Mr = 10,000-30,000) has been reported to be released from normal
human fibroblasts by saline washes of confluent cultures (121). In
addition, Mizel et al. (122) have demonstrated that medium exposed to
murine sarcoma virus-transformed 3T3 fibroblasts contained potent
immunosuppressive factors that inhibit in vitro thymocyte
proliferation. The inhibitory activity was found to migrate on gel
filtration chromatography with apparent molecular weights 12,000 and
8,000.

Finally, Holley and coworkers have reported (123) that the growth
of the monkey epithelial cell line, BSC-1, is limited at high cell
densities partly by the action of three inhibitors. These inhibitory
molecules are released by the BSC cells themselves and appear to

consist of lactate, ammonia, and an independent substance. Recent

26

attempts at isolation of the unidentified inhibitory substance have
shown the presence of two high molecular weight growth inhibitory
factors in medium exposed to BSC-1 cells (124). The inhibitory
activity was determined to be effective at ng/ml concentrations, and
reversibly inhibit passage of BSC-1 cells in 61 phase of the cell
cycle. Also, the action of the inhibitors could be overcome by
addition of epidermal growth factor or fresh calf serum to the test
cultures (124).

The inhibitory activity of these factors were tested on a variety
of different cell types. The inhibitors were most active when tested
on BSC-1 cells, less active on isolated rat and human epithelial cells
and not active when assayed on fibroblasts (124). The demonstration of
potent growth inhibitory activity in medium exposed to cultured cells
greatly supports the notion that endogenous growth inhibitors play an
important role in in vitro growth control.

It must be emphasized that of all of the reported soluble
endogenous growth inhibitors not one has yet been purified to a
sufficient extent such that the chemical nature of these molecules can
be convincingly described. To this end, the aim of our studies is
directed at develOping a system to investigate the role of endogenous
inhibitory factors in growth control of the 3T3 fibroblast cell line.
The work contained in this thesis is divided into three parts: (a)
demonstration of inhibitory activity in medium exposed to 3T3 cells;
(b) identification of the responsible factors; and (c) some binding

properties of the inhibitory factors.

ASSAYS OF CELL PROLIFERATION AND DEMONSTRATION
OF A GROWTH INHIBITOR ACTIVITY

Despite the large number of phenomenological observations on
soluble endogenous growth inhibitors, analysis of the mechanisms of
“density-dependent inhibition of growth are obviously limited by the
lack of purified preparations of the growth regulatory molecule(s).

Two key requirements for the successful fractionation of a molecule are
(a) a well-defined source of the material and (b) reliable assays of
the molecule's activity. In the first chapter, we demonstrate that
treatment of proliferating 3T3 cells with medium conditioned by
exposure to density-inhibited 3T3 cultures resulted in an inhibition of
growth and division when compared to parallel treatment with
unconditioned medium. we have characterized some properties of the 3T3
system for the preparation and assay of the growth inhibitory activity.
The results suggest that a well-defined system for the purification and
molecular analysis of these factors may be developed from the 3T3 cell

line.

27

MATERIALS AND METHODS

Tissue Culture

The cell lines, 3T3 fibroblasts and LCC-RKI rabbit kidney cells
were obtained from American Type Culture Collection (Rockville, MD).
We have also obtained 3T3 cells and their SV-4O virus transformed
counterparts from Dr. Howard Green (Massachusetts Institute of
Technology, Cambridge, MA). The 3T3 cells from both sources gave
comparable results. Chicken embryo fibroblasts were obtained from ten
day old eggs following the procedure previously described (125). We
have used hamster fibroblasts cell lines, Nil-2 and polyoma virus
transformed Nil-2 cells (126), and also the neuroblastoma cell line
N4TG3 (127). The saturation density of the untransformed 3T3, Nil-2
and chicken embryo fibroblasts were 5 x 104 cells/cmz, 4 x 105
cells/cmz, and 9 x 104 cells/cmz, respectively.

All cells were cultured in Dulbecco Modified Eagle's Medium (DME)
(K.C. Biological, Lenexa, KA) supplemented with 100 U/ml penicillin,
100 ug/ml streptomyocin and fetal bovine serum (FBS) (GIBCO, Long
Island, NY) to a final concentration of 10%. The cells were incubated
in Corning plastic tissue culture flasks (Corning Glass Works, No.

25100, 25110, and 25120, Corning, NY) or multi-well tissue culture

28

29

dishes (Costar, No. 3524, Cambridge, MA) maintained at 37° in a humid
atmosphere of 10% C02 and balanced air. The 3T3 fibroblasts were
passaged and used for a maximum of three months, after which they were
discarded, and a fresh sample was grown from frozen cultures kept at
-80°.

Routinely, our stock culture of 3T3 cells was passaged every two
days, before the cells reached a confluent state. The growth medium
was first removed, and the cells were washed and then incubated in PBS
with 0.25% trypsin (Nutritional Biochemicals, Cleveland, OH) and 4 x
10'4 M Versene at 37° for ten minutes. The cells were dislodged
from the growth surface, transferred and centrifuged at 1320 x g for
three minutes. The trypsin solution was removed and the cells were

resuspended in DME-FBS and seeded at the desired density.

Co-Cultivation of Dense and Sparse Cultures

In these experiments, cells were seeded on glass coverslips (2.5 x
1 cm) cut from microscope slides (Corning, No. 2948). For dense
cultures (5 x 104 cells/cmz), the 3T3 cells were inoculated, on
coverslips, at 2 x 104 cells/cm2 and allowed to grow for 72 hours.
For sparse cultures (104 cells/cmz), the cells were plated at 2.5 x
103 cells/cmz, on coverslips, and incubated for 24 hours. Both
types of coverslips were then incubated with 0.75 "Ci/culture of
14C-labeled amino acids (58 Ci/matom of 14C, Amersham-Searle,
Arlington Heights, IL) for 24 hours. After washing three times with
DME-FBS, the coverslips were then moved into a new culture dish
(Costar, No. 3506) in various combinations and 1.5 ml of fresh DME-FBS

was added to each dish. The cultures were incubated at 37° in a

3O

humidified chamber equilibrated with 10% C02 atmosphere and shaken on
a turntable at 30 run.

After 24 hours of co-cultivation, the cells were pulsed with
[3H]thymidine (1 uCi/dish, 1.9 Ci/mmole, Schwartz Mann, Orangeberg,
NY) for three hours at 37°. The coverslips were then washed three
times with cold PBS and the cells on each individual coverslip were
scraped into two ml of PBS using a rubber policman. The cells were
then placed on Whatman GF/A filters and the amount of acid-precipitable
[3H]thymidine incorporated was determined as described in DNA

synthesis assays.

Preparation of Conditioned Medium

We have "conditioned“ growth medium by exposing it to a confluent
monolayer of cells and then tested its capacity to support growth and
cell division in sparse cultures. For this purpose the following terms
are defined explictly: (a) conditioned medium (CM) -- DME-FBS that has
been incubated in a flask containing 3T3 fibroblasts and maintained at
37° in a humid atmosphere of 10% C02; (b) Unconditioned Medium
(UCM) -- DME-FBS incubated under identical conditions but in a flask
containing no cells; (c) Source Cells -- cells (usually
density-inhibited 3T3 fibroblasts) that are used to condition the
medium; (d) Target Cells -- cells (usually sparse cultures of 3T3
cells) that are used to test the capacity of CM and UCM to support
growth and cell division.

CM was prepared according to the protocol diagrammed in Figure 3,
unless otherwise mentioned. The source cells were seeded at an initial

density of 2 x 104 cells/cm2 and allowed to grow for at least three

Figure 3.

31

 

 

 

 

[commuomuiuegg] fifiibnmummuiqug
Source Cell 8 I 2 x to‘cells/ cm’)
7211
Growth Media

@— (0.1ml/cm’)

Confluent (4-5 x io‘oelis/cm’)

 

 

 

 

Stimulation 24 h
Feeding
h Medals No cells
(0.1 rnl/ cm )
Conditioning
Media 2‘ h
auditioned Centrifuge U Unconditioned
sterile Harvest U 1470 X 9 Media Harvest
cu kimu' ucu
(0.5 nil/cm”) (0.5 nil/en?)
23333.
Target ______y- 0 0
° 2 2 2 2 o
l 24 h
s
Assay ( l-i ) Tdn incorporation
protein content determination

cell number determination

Schematic diagram of the protocol used in the preparation of
Conditioned Medium (CM) and Unconditioned Medium (UCM). The
cell densities are expressed as the number of cells per

cm2 of growth surface in the tissue culture dish.

Similarly, the volume of medium added at the various steps
are expressed in this diagram as the number of mi of medium
used per unit area of growth surface.

32

days. Upon reaching a confluent monolayer, the medium was removed and
“stimulation feeding" was performed by adding fresh DME-FBS (0.1 ml per
cm2 of growth surface). The purpose of this feeding was to induce

one final round of cell division and assure a confluent monolayer.
After 24 hours, the medium was again removed and replaced with the same
volume of fresh DME-FBS as medium to be conditioned. After 24 hours
the medium was then transferred and centrifuged at 1470 x g for ten
minutes to remove cellular debris and particulate material. The
supernatant was collected and used as 04. For transformed cell lines
that fail to respond to density-dependent inhibition of growth, CM was
prepared by adding DME-FBS to source cells when a confluent monolayer
was observed. In all experiments, UCM was prepared in parallel from
the same batch of DME-FBS and incubated under the same conditions but
in the absence of cells.

For experiments that required source cells at less than the
saturation density, the cells were plated at one fourth of the desired
density. Four hours later, DME-FBS (0.1 ml/cmz) was added for
stimulation feeding. After 24 hours the old medium was removed and
DME-FBS (0.1 ml/cmz) was added for production of CM, for experiments
of constant volume at varied source cell density. For experiments in
which the volume of the medium and the cell density varied
proportionally, medium was added such that the ratio of source cells
per unit volume was constant, assuming the desired density was
attained. In all cases, the number of source cells used to condition a
batch of medium was counted, after collection of CM.

Serum free conditioned medium (CM(SF)) was prepared by a similar

protocol as for CM, except DME did not contain FBS during conditioning.

33

Parallel incubations in the absence of cells yielded UCM(SF). FBS was
added to both CM(SF) and UCM(SF), to a final concentration of 10%,
before the medium was tested for target cells.

Target cells used to test the growth capacity of CM and UCM were
always obtained from our stock cultures. They were routinely seeded at
an initial density of 2.5 x 103 cells/cm2 in a 24 well culture dish
(Costar, No. 3524). For experiments in which the density of the target
cells were varied, the cells were seeded at one half the desired
density. After 24 hours, the medium was removed and 1 ml of CM or UCM
was added. The growth of the CM and UCM treated target cells was
usually determined after 24 hours of incubation, using the assays

described below.

Assays of DNA Synthesis

 

The target cells to be assayed for DNA synthesis were pulsed with
1 pCi/culture of [3H]thymidine for 3 hours at 37°. After the pulse,
the radioactive medium was removed and the cells were washed three
times with cold PBS. The cells were then removed from the growth
surface by trypsin treatment and deposited on Whatman GF/A filters.
They were washed with PBS, 5% trichloroacetic acid, and methanol. The
filters were counted in a scintillation counter in five ml of
scintillation cocktail (0.2 g dimethyl-POPOP, 8 g PPO, 1000 ml Triton
X-100 and 2000 ml of toluene).

For autoradiography, the cells were plated at an initial cell
density of 4 x 103 cells/cmz, and incubated for one day in Leighton
tubes (Costar, No. 3393). The cultures were then treated with UCM or

CM for 24 hours and then pulsed with 0.1 uCi/culture [3H] thymidine

34

for 3 hours at 37°. The cells on the Leighton tube slides were then
washed five times in cold PBS, 5% tirchloroacetic acid, fixed in
absolute ethanol, dipped in NTB-2 Nuclear Track emulsion (Eastman Kodak
Co., Rochester, NY) and exposed for 18-24 hours (128). Exposing slides
for shorter or longer times yielded proportionally smaller or larger,
respectively, average number of grains per cell. After final fixation,
slides were stained in Giemsa (GIBCO), and dipped in xylene and a
coverslip was mounted with Permount (Fisher Scientific Co., Pittsburg,
PA).

At least 70 labeled cells were systematically counted in every
case. The error in grain counting was estimated by multiple counts of
the same cell. This analysis indicates that the standard error of
visual grain counting method is 5-10% in each individual case. For
reversibility experiments, the amount of 14C-labeled thymidine in
each culture was increased to 2 uCi/culture and the labeled cells were
exposed for 48 hours. These procedural modifications facilitated the

counting of percent labeled cells.

Assays of Cell Number and Cell Viability

Two different methods were used to determine the total number of
cells in a given culture. In the first method, the cells were washed
three times with PBS and were removed from the growth surface by
trypsin treatment. The cells were then centrifuged, resuspended in
PBS, diluted with trypan blue, and counted in a corpusle counting
chamber (Hausser Scientific, Blue Bell, PA). Alternatively, the cells
were washed with PBS, and then dissolved in 0.1% sodium dodecyl

sulfate. An aliquot of this sample was then used for total protein

35

determination in the culture by method of Lowry et al. (129). Over a
concentration range of 104-105 cells, the amount of protein in a
given culture was directly proportional to the number of cells, as
counted by the first method.

The viability of cells treated with UCM and CM were determined
while the cells remained attached to the growth surface. After removal
of growth medium, the cells were incubated with trypan blue (0.08% in
PBS) for ten minutes at room temperature. The staining solution was
then removed and the viable cells were counted using an Olympus
inverted microscope.

Target cells were also labeled with 14C-amino acids (0.75
uCi/culture) for 48 h. After washing, these cells were treated with CM
or UCM. At various times thereafter, the cells were washed with PBS
and were dissolved in 100 pl of 0.1% sodium dodecyl sulfate. The
amount of labeled target cells remaining after CM or UCM treatment was

determined by counting the 14C radioactivity in each culture.

Reversibility and Renewed Applications of Conditioned Medium

In two series of experiments, it was necessary to replace the
medium on target cells after initial exposure to CM or UCM. For
refeeding experiments, parallel cultures were treated with CM. After
24, 48 and 72 hours, the CM in one set of target cells was removed and
replaced with an equal volume of freshly prepared CM. The other set of
target cells retained the CM added initially. DNA synthesis in refed
or not refed cultures was assayed every 24 hours. Concurrent

experiments were carried out using UCM.

36

For the reversibility experiments, parallel cultures were treated
with CM for 12, 24, or 48 hours. At the specified time, the medium was
removed from the target cells and replaced with an equal volume of
fresh DME-FBS. At times corresponding to O, 6, 12, and 24 hours after
the replacement of medium, DNA synthesis was assayed by the
incorporation of [3H]thymidine and the number of cells was estimated
by protein determination (129). Parallel experiments were carried out

with UCM treated cultures.

Fractionation and Molecular Properties of Conditioned Medium Components

For the fractionation studies, solid ammonium sulfate was added to
CM(SF) to a saturation of 80% at room temperature. All subsequent
operations were performed at 4°. The supernatant was decanted after
centrifugation at 12,400 x g for 15 minutes. The precipitate was
resuspended in 3 ml of DME for dialysis experiments. Cellulose
dialysis tubing (A.H. Thomas, Philadelphia, PA) with molecular weight
cutoff of 12,000 was prepared as previously described (130). Both the
supernatant and precipitant fractions were dialyzed against PBS (4
liters, 3 hours), PBS (2 liters, 1.5 hours) and DME (1.5 liters, 1.5
hours). Similar procedures were carried out on UCM(SF); however, no
precipitate was observed in this case. To all fractions, FBS was added
to a final concentration of 10%, before the fractions were tested on
target cells.

For column fractionation, the ammonium sulfate precipitate was
redissolved in 1 ml of DME and applied to a column (1 x 60 cm) of
Sephadex G-15 equilibrated with DME. Fractions (1 ml) were assayed for

protein content (50 ul aliquots from each fraction) using method of

37

Lowry et al. (129), or for radioactivity, where appropriate. To the
remainder of the fraction, 0.1 ml of FBS was added. Then these
fractions were tested on target cells» The same procedure was followed

for UCM.

RESULTS

Growth Characteristics of 3T3 Cells

Because the present series of experiments depends upon a detailed
analysis of the growth characteristics and assays of proliferation of
3T3 cells, we have made a particular effort to define the growth curve
of the fibroblasts under our culture conditions. When identical
cultures were seeded at a density of 500 cells/cmz, a lag period was
observed for two days before significant growth, as assayed by counting
the number of cells per cultUre (Figure 4a). The increase in cell
number was exponential (Figure 4a inset) for six days before it reached
a plateau, at a cell density of about 4-5 x 104 cells/cmz. The
generation time and saturation density of our 3T3 cell line was taken
to be 22 hours and 5 x 104 cells/cmz, respectively.

A similar growth curve was obtained by assaying for total amount
of protein in each culture (Figure 4b). Furthermore, the number of
cells, in a cell number range of 104 to 105 cells, was directly
proportional to the amount of protein in a given culture (Figure 4a
inset).

The measurement of DNA synthesis by incorporation of [3H]thymi-

dine is the third parameter used to characterize the growth curve of

38

39

Figure 4. The growth curve of 3T3 fibroblasts cultured in DME-FBS as
assayed by (a) assessment of the cell density of the
culture, (b) determination of the total protein content of
the culture, and (c) assay of DNA synthesis as measured by
the incorporation of [3H]thymidine. Cultures were
inoculated at a density of 500 cells/cmz; at the indicated
times, parallel cultures were taken for the various assays.
Data points are the averages of measurements on duplicate
cultures. The inset in (a) is a semi-logarithmic plot of
the same data to show exponential gorwth of the 3T3 cells
over a period of 144 hours. The inset in (b) shows the
linear relationship between the total amount of protein in a
given culture and the number of cells in that culture. At
the point indicated by the arrow, the cultures were judged
to be confluent by microscopic observation.

40

W.

m
mm
mini

.895: :8. lair...

 

 

 

  

 

 

 

IIIA) 2p
.9 x «suite

 

 

20 Cell Nunber lx10'4l

aerial

.2: 532a

WWW

tin)

 

Tic

66

 

 

NT

n. i
A»? a 58. c319”—

41

3T3 fibroblasts (Figure 4c). It was found that DNA synthesis increased
proportionally with cell number until the eighth day and decreased
abruptly thereafter. The peak value of DNA synthesis corresponded
roughly to the point when the cultures were judged to be confluent by
microscopic observation, even though the actual cell number continued
to increase for 24 hours or more after this point. These results
characterize then the growth curve of 3T3 cells and form the basis for
studying the properties of a growth inhibitory activity associated with
the 3T3 cell line.

Evidence for a Growth InhibitoryiActivity in Medium Exposed to

 

Density-Inhibited Cells

The hypothesis that density-inhibition of growth may be mediated
by soluble products was tested by using the experimental procedure of
Harel et al. (19) in whiCh two coverslips inoculated at two different
cell densities were co-cultured in fresh DME-FBS (Figure 5). DNA
synthesis was measured by [3HJthymidine incorporation into acid
insoluble DNA, and then normalized to total protein. The amount of
protein was determined by the amount of [140] amino acids
incorporated into acid insoluble material. DNA synthesis in cells on a
coverslip containing approximately 104 cells/cm2 (sparse cultures)
was significantly decreased when co-cultured with a coverslip
containing about 5 x 104 cells/cm2 (dense cultures), then when
cultured alone (Table II). This inhibition was not observed when two
sparse coverslips were cultured together. These results on DNA
synthesis are in agreement with the results of Harel et al. (19) who

measured cell growth and phosphate metabolism by the same experimental

42

 

 

Figure 5.

 

 

  

’///////.

IDE) ems

      

 

 

 

 

 

 

Schematic diagram illustrating the experimental design for
culturing two coverslips next to each other in a tissue
culture dish. The letter S represents a coverslip
containing a sparse culture (5 10 celis/cmz) of 3T3
fibroblasts and the letter 0 represents a coverslip
containing a dense (2_4 x 104 cells/cmz) of 3T3 cells.

The dimensions of the coverslip were 2.5 cm by 1 cm. The
coverslips were cultured in a tissue culture dish (3.5 cm
diameter) containing 1.5 ml of fresh DME-FBS.

43

.mmczapsu muoo_—a_cu soc; mu_=mmc do moamco>m on»
«commence moapm> concede; och .ucoe_coaxo cowum>Pupzoiou use o» co_ca uaumconcou:_ mu_um o:_so
umpone_iue o» one «a: peg» xuw>_uumo_umc mo pesoEe on» an um:_scouou mm: :o_umcoacoo:_
o:_upsxgumzmw do uo_cog mmpza oz» mzvczu mappmco>ou pm=u_>_u:P «so do acoucou =_ouoca we»
.moczupzo ouuuwpq_cu
Eocw mapamoe do mommco>m on» «command; moapm> uopcoqoc on» m—mPcmumE opnmo_a_oochu_um
ou:_ cmuucoacooc_ ocpupexgpn: u we “cacao as» »a umeammme no: mwmogucxw <zo+
.Amsu m_—mu «OH x qu mocaupau emcee .o ”Am58\m__oo

 

 

eomwv moc=u_=o emceam .m "mam moc=»_=u do macho acocmewpu on» com vow: mcopuew> cane been
"moaocuood
m.~ «can NNHN a o .m
~.mH mom mmmm m m .e
~.~ mmem mmsu o
m.m mwum sums o co .m
m.m~ use ommu m
m.~H mom mmem m mm .N
m.~ omafi Kmmm o
o.~ mew awed m cm .H
acmucou :wmuocm A+saov A+squv qp_mco>ou do ucms_coqu

\m_.aebeam (z: seabeou e_eboee m_maeoe»m (za seameaa ..au

 

mp_au
eee.e.ee_-su.meeo gee: eboe>_e_=o-ou m__eo mam co mae=e_=o esteem e_ m_maeoe»w <zo .HH b.3ee

44

method. Moreover, these results lend strong support to the hypothesis
that growth control in 3T3 fibroblast may be mediated, at least in

part, by soluble factors endogenous to the cells.

Effects of Conditioned Medium onL3HlThymidine Incorporation in 3T3
Target Cells

 

A key problem associated with any growth inhibitor purification
project is the development of a reliable assay system. Three major
technical problems involved with growth assays are: (a) cytotoxic
effects associated with growth inhibitory fractions; (b) altered
metabolism of radioactively labeled thymidine used in growth assays,
unrelated to DNA synthesis; (c) nucleoside precursors diluting out the
specific activity of labeled thymidine, and thus mimic the action of
growth inhibitors. We have used three independent assays to
demonstrate the CM from 3T3 cells can no longer efficiently support
growth and cell division of 3T3 cells.

To sparse cultures of 3T3 cells plated at a density of 2.5 x 103
cells/cmz, CM and UCM prepared as illustrated in Figure 3, was added
to the cultures. DNA synthesis, as measured by [3H]thymidine
incorporation into acid insoluble DNA, of target cells treated with CM
showed a significant decrease when compared to parallel cultures
treated with UCM (Figure 6). The levels of [3H]thymidine
incorporation in UCM treated target cells was similar to that observed
in the normal growth curve (Figure 3c).

Since the incorporation of [3H]thymidine is the assay most often
employed in our studies, the following terms are defined to express the

inhibitory activity associated with CM as compared to that of UCM.

45

 

l

A e—eUCM

gr o—-—oCM

2

x2-

E

Cl.
.0 0
m1-

;3 .
,_,

O

 

t(h)

Figure 6. The kinetics of [3H]thymidine ([3H]TdR) incorporation in
3T3 cells cultured in the presence of Conditioned Medium
(CM, o————o) and Unconditioned Medium (UCM, e————e). At
various times, parallel cultures were pulsed with 1 uCi of
[ HJthymidine for 3 hours. Data points are the averages
of measurements on duplicate cultures and are plotted at
times corresponding to 1.5 hours after the start of each
pu se.

46
"Percent growth“ is defined as the ratio of amount of [3H]thymidine
incorporated by CM target cells as compared to UCM cultures, multiplied
by 100 (% growth = CM/UCM x 100). The inhibitory effect of CM is
defined as “percent inhibition" (% inhibition = 100% - % growth).

Effects of Conditioned Medium on DNA Synthesis in 3T3 Cells Assayed by

Autoradiography

 

To ascertain that the inhibition of [3H]thymidine incorporation
in target cells by CM was due to a suppression of cellular DNA
synthesis rather than to alterations of the transport or pool sizes of
the 3H-labeled thymidine, we have assayed for the incorporation of
[3H]thymidine at an individual cell level by autoradiography. Target
cells were treated with CM or UCM for 24 hours and assayed for
[3H]thymidine incorporation by scintillation counting of acid
precipitable radioactivity. The percent inhibition was 49%. When
parallel cultures treated with the same preparation of CM and UCM were
assayed by autoradiography, we found that the percent of labeled cells
was greater in UCM-treated cultures (56%) than in CM-treated cultures
(35%). The fraction of labeled cells decreased to 29% when cultures
were treated with CM for 48 hours. Therefore, the total incorporated
radioactive thymidine in CM-treated cells decreased in parallel with
the percent of cells undergoing DNA synthesis.

More importantly, we have found that the average number of grains
per labeled cell was invariant within the error of estimation for CM-
and UCM-treated cultures, for a given experiment. Moreover, the
distributions of grain counts for labeled cells of both cultures appear

to be similar (Figure 7). Chi-square analysis of the data in Figure 7

47

/“/4
+
p

 

 

 

 

 

//

 

 

 

Number of cells

 

 

 

 

 

 

 

 

 

 

 

 

 

sh ii) 150 200
Grains per cell

Figure 7. Grain count distributions for 3T3 cells cultured in the
presence of (a) Conditioned Medium and (b) Unconditioned
Medium. The following numerical data were obtained for
Conditioned Medium cultures in: (a) 200 total cells counted,
70 labeled cells (35% labeled cells), average number of
grains per labeled cell, 115. The corresponding data for
Unconditioned Medium cultures in (b) were 200 total cells
counted, 111 labeled cells (56% labeled cells), average
number of grains per labeled cell, 108. The average number
of grains per labeled cell for each of the cultures is
indicated by the arrow in the grain count distribution
graphs in (a) and (b).

48

showed that these distributions are the same at the 90% confidence
level. All of these results lend further support to the notion that
the inhibition of [3H]thymidine incorporation in target cells by CM

reflects a true inhibition of DNA synthesis in 3T3 cells.

Effects of Conditioned Medium on Cell Number and Cell Viability

 

We have also characterized the effect of CM on growth and cell
division by two other assay systems, cell counts and total protein. It
was determined that proliferation of the target cells exposed to CM was
significantly inhibited, as measured by cell counts (Figure 8),
compared to UCM cultures that showed normal growth characteristics
(Figure 4). Moreover, the generation time calculated for target cells
exposed to UCM and CM was 24 and greater than 100 hours, respectively.
To confirm that cell growth suppression was attributable to effects of
CM, the total protein of the cultures were determined. The difference
observed between UCM and CM was similar to that previously seen. Also,
because we have established that the amount of protein was directly
proportional to cell concentration over a large range, this result
supports the contention that inhibitory activity is due to CM, and not
assay artifacts.

In addition, we have found that there was no significant differ-
ence in viability (Table III) of cells treated with UCM or CM. First,
the viabilities of cells were identical for target cell cultures
treated with UCM and CM for up to 48 h, when assayed by trypan blue
exclusion tests (Table III). Because nonviable cells could have been
lost from the dishes, we prelabeled the target cells by culturing them

in the presence of 14C-amino acids. After treatment with UCM or CM,

49

40)

H UCM
C>--f-() (:IVI

CeIIS/ cm2 ( x 10'4)

 

 

0 4'0 so 120 160

Figure 8. The kinetics of the increase in cell density in 3T3 cells
cultured in the presence of Conditioned Mediun (CM, o————o)
and Unconditioned Medium (UCM, e————e). At various times,
parallel cultures were trypsinized and the cells collected
for determination of the total cell number in the cultures.
Data points are the averages of measurements on duplicate
cultures.

50

 

 

 

Table III. The Effect of Conditioned and Unconditioned Medium on DNA
Synthesis and Viability of Target Cells
Time Retention of prelabeled Cell Viability+
h) 14C proteins Number of stained
cells/total cells counted)
_CM 991 _CM M
0 4,201 1 114 4,238 i 226 0/212 0/100
3 3,929 2 143 4,274 1 187 0/202 1/233
6 3,325 1 171 3,321 e 130 0/229 0/282
24 2,601 1 271 2,686 i 91 1/228 ,1/285
48 1,984 2 75 2,028 i 19 0/202 0/216
Footnotes:

+Cell viability was assayed by the trypan blue dye exclusion

technique described in Materials and Methods.

51

the 14C-radioactivity retained were found to be similar in both
sets of cultures (Table III).

Finally, the morphology of the cells was not effected by treatment
with UCM or CM (Figure 9). It should be noted that as the target cells
treated with UCM became confluent, the cells assumed an orderly
cobblestone appearance, characteristic of crowded 3T3 cells (15).
However, since the target cells treated with CM did not approach
confluency, their morphology remained characteristic of sparse 3T3

cells (Figure 9).

Reversibility_of the Effects of Conditioned Medium

In order to demonstrate that the inhibitory effects of CM on DNA
synthesis was reversible, and not due to cytotoxic effects, parallel
target cell cultures were treated with CM for 24 h. The medium was
then removed and replaced with an equal volume of fresh DME-FBS. At
various times thereafter, DNA synthesis was assayed by the incorpora-
tion of [3H]thymidine. Concurrent experiments were also carried out
with UCM treatment. In order to adjust for the fact that after 24
hours, the total number of cells in UCM-treated cultures was greater
than that in CM-treated cultures, the incorporation of [3H]thymidine
in each culture was normalized by the amount of protein present in that
culture. The data showed that the inhibition was reversible within ten
hours after removal of CM (Figure 10a). Similar results were obtained -
when target cells were exposed to CM for 12, 48, and 72 hours before
replacement with fresh DME-FBS.

Moreover, we have demonstrated the reversibility of the effects of

CM by autoradiography. The percentage of radioactively-labeled cells

Figure 9.

52

 

Photographs of representative 3T3 cells cultured in
Conditioned Medium (a) and Unconditioned Mediun (b) for 24
hours and in Conditioned Medium (c) and Unconditioned Medium
(d) for 60 hours. The cells were inoculated at an initial
density of 5 x 103 cells/cmz. The photographs were

taken on a Leitz MPV 2 microscope using phase contrast
optics and Kodak Tri-X pan film. The magnification factor
for the photographs was x 100. The bar represents 10 uM.

Figure 10.

53

The reversibility of the effect of Conditioned Medium.
Target cells (5 x 103 cells/cmz) were treated for 24

hours with Conditioned and Unconditioned Medium. The medium
was then removed from each culture and replaced with an
equal volume of fresh DME-FBS. (a) At various times
thereafter, DNA synthesis was assayed by the incorporation
of [3H]thymidine, normalized by the protein content of the
culture, and expressed as percent of growth as defined in
“Materials and Methods". Data points represent the averages
of measurements on duplicate cultures. (b) Percent of
radioactively labeled cells as determined by autoradiography
in cultures treated with Conditioned Mediun ( ) and
Unconditioned Medium (H ) at various times af er
reversal with fresh DME-FBS. Data points represent the
averages of triplicate determinations 3 standard error of
the mean.

54

 

 

 

 

 

m=oO 30.30.. R

 

Time after reversal (h)

55

increased from 32% in cells treated with CM for 24 hours to
approximately 55% in cells replenished with fresh DME-FBS (Figure 10b).
Control cultures treated with UCM showed labeling in about 45% of the
cells throughout this experiment. Taken together with the previous
demonstration that the viability of the target cells was not effected
by CM, these results strongly suggest that the inhibitory activity of

CM cannot be ascribed to cytotoxicity.

Target Cell Parameters

The effect of variation of the target cell density on the growth
inhibitory activity is shown in Figure 11. At target cell density of
5 x 103 cells/cm2 or greater, the percent inhibition was
consistently 60-70%. However, a sharp decrease in the ability of CM to
inhibit growth of 3T3 cells was observed at a target cell density below
2.5 x 103 cells/cmz. The mechanism of this density-dependent
response to CM is not understood, but a possible explanation may be
that at high cell densities, the target cells contribute their own
inhibitory factors during exposure to CM. In any case, the data
indicates that the optimal density to test the effects of CM on target
cells with minimal interference was 5 x 103 cells/cmz.

The kinetics of DNA synthesis was also studied as a function of
time of exposure to CM and UCM on target cells. The 3T3 cultures were
originally initiated at 2.5 x 103 cells/cmz, to allow for longer
exposure before reaching a confluent state. DNA synthesis, assayed by
[3H]thymidine incorporation, was markedly decreased in cultures
exposed to CM as compared to UCM cultures (Figure 6). Because the

level of [3H]thymidine in cells treated with CM for 24 hours was

Figure 11.

% inhibition

56

1C“)

END.

END.

_;

4!).

2C),

 

 

(”A

o 1' 2 :3. 21
Cells/ em2 I x10'4 I

The effect of varying the density of target cells on the
inhibitory activity of Conditioned Medium. Conditioned
Medium was prepared using 5 x 105 source cels per ml of
DME-FBS and 24 hours of exposure time; the inhibitory
activity was tested by treating target cells at different
densities with the Conditioned Medium for 24 hours followed
by a 3 hour pulse with [3H]thymidine (1 uCi/culture).

Data points represent the averages of measurements on
duplicate cultures and are expressed as inhibition of growth
as defined in "Materials and Methods."

57

consistently about 40% of that observed in cells treated with UCM, we
have chosen this time of exposure for our routine measurements of
inhibitory activity in the CM.

Consistently, it was observed that the level of [3H]thymidine
incorporation in CM treated cultures began to increase after 50 hours.
Although the details of this pehnomenon are not well understood, it may
reflect that the inhibitor(s) in CM are losing their effectiveness due
to denaturation or degradation. To test this hypothesis, the CM
treated cultures had their medium removed, and were refed with freshly
prepared CM every 24 hours. The inhibition in the refed cultures did
not exhibit a decrease in activity after 50 hours, but showed an
increase in the inhibitory activity (Figure 12). These results support
the hypothesis of inactivation of inhibitory factor(s). Also the
decrease in inhibitory activity at long term exposure to CM suggests
that CM cannot act through depletion of growth factors, because it
would be hard to envision a mechanism that would restore growth

factors.

Cellular Specificity of the Effects of Conditioned Medium

The inhibitory activity of CM prepared from 3T3 source cells was
also tested on a variety of other target cells. On the basis of the
data obtained so far (Table IV), the following generalizations can be
made. First, the normal fibroblasts, 3T3, Nil-2, and chicken embryo
fibroblast showed the greatest sensitivity to inhibition by CM.
Second, the SV-40 and polyoma virus-transformed fibroblasts showed much
less inhibition, when tested with the same CM, than the normal

coudterparts. Third, of the two non-fibroblast cell lines tested, the

Figure 12.

% Inhibition

58 '

1001 H CM continuous

0.--.0 Fresh CM added
80.
I ’0‘ ‘
,0’ ’ \ ‘°
60. , ’
4K).
20.

 

 

o 20 4'0 60 so 100
t (h)

The effect of varying the time of exposure of Conditioned
Medium to the target cells on the inhibitory activity.
Conditioned Medium was prepared using 5 x 105 source cells
per ml of DME- FBS and 24 hours of exposure time; the
inhibitory agtivity was tested by treating target cells (5 x
103 cells/cm for various times followed by a 3 hour
pulse with [3H]thymidine (1 uCi/culture). In one series
of cultures, the target cells were exposed continuously to
the Conditioned Medium added initially (I—l. ); in
another series of cultures, the Conditioned Medium was
replaced every 24 hours with fresh Conditioned Medium
(o————o). Data points represent the averages of
measurements on duplicate cultures and are expressed as
percegt inhibition of growth as defined in "Materials and
Metho s".

59

Table IV. The Cellular Specificity of the Effect of Conditioned
Medium on Growth Inhibition

 

Inhibition (%)

 

 

2 x 104 5 x 103

Target Cells* target cells/cm2 target cells/cm2

3T3 54.2 50.7

Nil-2 50.5 59.9

CEF 47.8 40.4

sv 313 14.6 44.9

py Nil-2 7.7 27.4

LCC-RKl 32.0 2.3

N4TG3 13.8 ‘ 29.1
Footnotes:

*The target cells used are: 3T3, a cell line derived from mouse
embryo fibroblasts; Nil-2, a cell line derived from hamster embryo
fibroblasts; CEF, primary cultures of chicken embryo fibroblasts; SV
3T3, SV 40 virus-transformed 3T3 cells; py Nil-2, polyoma
virus-transformed Nil cells; LCC-RK1, a cell line derived from
rabbit kidney cells; N4TG3, a neuroblastoma cell line.

The level of growth inhibition was measured by comparing the
incorporation of [3H]thymidine in Conditioned Medium (CM)- treated
cultures with the corresponding incorporation in Unconditioned Medium
(UCM)- treated cultures.

60

epithelial line LCC-RKl derived from rabbit kidney cells and the
neuroblastoma line N4TG3, both showed weaker inhibition by CM than the
normal fibroblasts. As was observed with transformed fibroblasts, the
neuroblastoma line also showed, quite strikingly, much less inhibition
at the high cell density. Although these experiments are obviously
limited by the few cell lines tested, the results nevertheless suggest
that the effect of CM prepared from 3T3 source cells exhibit some

cellular specificity with respect to the targets.

Source Cells Parameters

Several parameters involved in preparation of CM were varied and
then the CM was tested for inhibitory activity by [3H]thymidine
incorporation. The parameters studied were (a) variation of the
density of source cells; (D) variation of the volume of medium being
conditioned; and (c) length of the exposure of the medium to source
cells.

The effect of variation of the density of source cells in the
preparation of CM is shown in Figure 13a. The level of inhibitory
activity was highest in CM prepared from cultures that were arrested at
their saturation densities ( 5 x 104 cells/cmz). This activity
decreased proportionally with decreasing cell density. It should be
noted, however, that the inhibitory activity could be observed even
when CM was prepared from source cells at 4 x 103 cells/cmz, one
tenth of the saturation density of the 3T3 cells, that we have
observed. This suggests, at least tentatively, that cell to cell
contact is not absolutely required to produce the growth inhibitory

activity.

Figure 13.

61

(a) The effect of varying the density of source cells on the
inhibitory activity of the Unconditioned Medium. Parallel
cultures of 3T3 cells in plastic tissue culture flasks (25
cm2 growth surface) were used to condition a constant

volume (2.5 ml) of DME-FBS.

(b) The effect of varying the volume of the medium
conditioned by the same number of source cells on the
inhibitory activity of the Conditioned Medium. Parallel
cultures of 3T3 cells at the same density (4 x 104
cells/cmz) in plastic tissue culture dishes (10 cm2

growth surface) were used to condition different volumes of
DME-FBS; the parameter expressed on the abscissa of this
graph represents the volume of medium added per unit area of
growth surface.

The inhibitory activity of the Conditioned Medium was tested
on target cells using the following conditions: target cells
at 5 x 103 cells/cmz, 24 hours of exposure to

Conditioned Medium, 3 hour pulse with [3H]thymidine (1

' uCi/culture). Data points represent the averages of

measurements on duplicate cultures and are expressed as
percent inhibition of growth as defined in "Materials and
Methods".

62

% Inhibition .s
§ 0’
.0 .0 .8 18
I”

10
.0

 

0
0|

 

2 6 4

 

 

 

O 1 -
Ceiis/ cm2 (X104)
1001 b.
80.
5
E 60.
a
E 40.
31
20.
0 . - ' '
0 04 0.8 12 1.6 20

ml/ cm2

63

When parallel cultures of source cells at the same density (4 x
104 cells/cm?) were used to condition different amounts of medium,
there was a decrease in the observed inhibitory activity with
increasing volume of the medium (Figure 13b). In this figure, the
parameter used in the variation is expressed as the volume of CM per
unit area of growth surface. When large volumes of medium were
conditioned (21 ml per cm2 of growth area) by the same number of
cells, no significant inhibitory activity was observed. Because the
cell density was the same in these experiments, the results suggest
that a minimal number of cells was required to condition a given amount
of medimn.

In order to substantiate this hypothesis, we varied the volume of
the medium and the density of source cells proportionally (Table V).
The culture at the highest cell density did not yield the highest level
of inhibition. Rather, cultures at lower cell density but exposed to
smaller volumes of medium yielded the greatest inhibition. The key
parameter was, therefore, the number of cells used to condition a unit
volume of medium. These results indicate that cell to cell contact was
not required for the generation of the inhibitory activity and that the
levels of the activity possibly reflected the concentration of the
molecular species in the medium.

The effects of variations in the length of exposure of the medium
to source cells was studied using confluent source cell cultures (5 x
104 cells/cmz). A rapid increase of inhibitory activity (from 6%
to 52%) was observed during the first 12 hours of exposure (Figure 14).

The increase in the activity was minimal over the next 20 hours.

Figure 14.

% Inhibition

64

 

 

0‘2'0'40‘6'0'60
t(h)

The effect of varying the length of exposure of DME- FBS to
source cells on the accumulation of inhibitory activity in
the Conditioned Medium. 3T3 Cells (5 x 104 cellg/cm2

cultured in plastic tissue culture flasks (25 cm growth
surface) were used to condition a constant volume (2. 5 ml)
of DME-FBS for different lengths of time and the resulting
Conditioned Medium was tested on target cells using the
following conditions: target cells at 5 x 103 cells/cm,

24 hours of exposure to Conditioned Medium, and 3 hour pulse
with [3 H]thymidine (1 uCi/culture). Data points represent
the averages of the measurements on duplicate cultures and
are expressed as percent inhibition of growth as defined in
Materials and Methods.

65

.mocaupao woueocuiAzozv Es_uoz uoco_uwu:ooc= :_ copumcogcou:_ m:_c=oamogcou

mgu sup: mmcsupzu umpamcpiAzuv e:_uwz uo=o_u_u=ou cw m=_uwexgumz u do
:o_oecoacou:_ use m=_cmasou an uoczmeos we: >u_>_oum xuouwn_;=_ guzocm mo —m>mm oz»

.s:_ums Co —E g cowpwueou ou wow: m__mo ea conga: on» muoopemc o_u~c m_;h«

 

 

"monocaoom
m.e~ mod x m.e no“ x m.m m.H om” bad x m.H
¢.- moH x m.e «OH x e.H ¢.H me baa x ~.H
a.oa mofi x e.m ace x a.~ a.m me ecH x o.~
m.ee mo“ x m.m do“ x m.m m.~ mN mod x N.m
Amy zu co ue=_o> A~56\.__uev A_ev =8 ANEUV

 

=o_b_e_ee_ .edee=z __du xuamedo __.o co ue=_o> aee< gazeeu Lease: __uu

 

=8 e_ »u_>_ou< xeoe_e_eea eezeea co eo.sepse=ee< be»
:6 s=_euz use do de=_e> deb eee .»b_meda P_.u .Luee=z .Fuu to Saddam 6;» .> ._ae»

66

Properties of the Growth Inhibitory Activity

 

In order to eliminate the possibility that the failure of CM to
support growth and division in target cells was entirely due to
depletion of serum factors that are required for fibroblast
proliferation, we have prepared CM(SF) by exposing serum free DME to
source cells in the same fashion as we had described for the
preparation of CM. This CM(SF) was supplemented with fresh FBS to a
final concentration of 10% before it was tested on target cells. When
3T3 target cells were treated with CM(SF), DNA synthesis was
significantly inhibited as compared to parallel cultures treated with
UCM(SF) (Table VI). Using autoradiographic assays, we have confirmed
that this difference is in the percent of radioactively labeled cells.
Although the level of inhibitory activity in CM(SF) was consistently
lower than that observed in CM prepared in the presence of serum, these
results nevertheless indicate that the growth inhibitory factor(s) can
be collected from source cells without the complications of arguments
concerning serum depletion.

The availability of the growth inhibitory activity in CM(SF) has
also allowed us to characterize its molecular properties. We have
found that the inhibitory activity can be detected in the precipitated
fraction of CM(SF) after treatment with ammonium sulfate at 80%
saturation (Table VI). Little or no activity was found in the
supernatant fraction. Moreover, similar treatment of UCM(SF) did not
show any partitioning of growth supporting functions.

It should be noted that in order to remove ammonium sulfate, both
the supernatant and precipitate fractions were first dialyzed against

large volumes of PBS and DME before they were combined with FBS and

.emuoguoz new m~e_cmumz= :_ uo=_emu ma gpzocm do =o_o_nwg=_

acougoa mu oommmcqxo m_ »u_>_uoe xgou_n_;:w och .cmoe any do cocco ucmuccum

on» « mcopuecpscmuou oueo_pa_cu do momecm>m me vommmcaxm men came on» .mcao; em
so» sawed: um=o_u_ucou use e=_uoz um:o_uwu=ou=: o» vomoaxo A su\m__ou oH x my
m—_oo ummcou mpm =_ copumcoqcoo=_ o:_u_5»;um: H mo Fm>o— as» comocamc mace ugh
.em oguoz use m—u_eoooz= :_ umn_commu

mew mFm»Fm_u one covue=o_uuece mace—3m aswcosse =_ new: magnumuocg —mu:oe_emaxo
map .cowamcsuem so» cu woven me: ouuepam Eapcoas< .uwuue Eacwm o: =o_z

E=_uoz m.o_mmm newepuoz cocoa—no .wzo “Eaton ocw>on ~euom uoH suF: coucoswpnazm
s=_umz m.o—mmm um_e_voz cocoa—an .mmmimzo "ecu vow: m:o_uew>mcnno as».

 

67

 

”mwuozpoom
-- oo¢~ a oom.- ooHH.u coH.H~ euua.e_e .mza

eu~»_e_e

.mueu_q_ooca
~.em com 4 oom.mH seem a ooH.e~ «ommflezzv mzo

eu~ape_e

.ouoccoazm
o.m comm u coH.e~ comm u oom.e~ eomwAezzv uzo
~.H~ oooH u oom.m~ coma u oom.m~ use
m.~e can u ooo.mH ooom u oom.e~ mmu-mzo

Asauv Asquv
Auv s=_euz e=_umz

eeae_epeea eueo,b_eeoo eueo_aaeeoue= .mee_b_eeoo

 

e=_euz euee_e_eeeu eueu-e=eum e_ »u_>_od< aeoo.e_eea
gazeeu use :6 m_m»_e_a eee ee_beeo_eeeed doec_=m e=_eeee< co budcem use .H> apnee

68

tested on target cells. We have found, however, that dialysis of the
CM(SF) resulted in the loss of inhibitory activity (Table VI).

Although the possibility that the inhibitory factor(s) may be adsorbed
on the dialysis membrane or otherwise denatured by this treatment has
not been eliminated, the present results suggest that the inhibitor nay
be dialyzable and therefore, may be a substance of low molecular
weight. In any case, this loss of activity on dialysis may account for
the fact that the observed enrichment of inhibitory activity in the
precipitate fraction was far less than expected from a calculated value
based on the protein contents of the original and precipitant
fractions.

In order to circumvent the problem of activity loss during
dialysis, we have fractionated the ammonium sulfate precipitate on a
column of Sephadex G-15 equilibrated with DME. We have found that the
inhibitory activity was associated with fractions corresponding to the
void volume of the column (Figure 15). There was a direct correlation
between the inhibitory activity and the protein content of the
fractions. More importantly, the corresponding fractions from a
parallel experiment using UCM(SF) did not show any inhibitory activity.
We have ascertained that the inhibition by the fractionated CM(SF)
material was reversible, indicating that we have not concentrated a
toxic factor. In order to establish the concentration range within
which a linear response of the inhibition of thymidine incorporation
assay can be obtained, we have diluted the fractionated CM(SF) material
with UCM(SF). We have found that the assay was linearly dependent on
the concentration of the inhibitory fraction up to the highest

concentration available (~60% inhibition).

69

All of these results strongly suggest that a system has been
developed for the purification and molecular characterization of growth
inhibitory factors that may mediate density-dependent regulation of

cell division in tissue cultured fibroblasts.

Figure 15.

7O

Chromatography of the ammonium sulfate precipitate fraction
of serum free Conditioned Medium on a column (1 x 60 cm) of
Sephadex G-15 equilibrated with DME. The ammonium sulfate
(80% saturation) precipitate was redissolved in 1 ml of DME
and applied to the column. Fractions of 1 ml were collected
and assayed for (a) inhibition of [3H]thymidine (3H TdR)
incorporation in target cells and (0) protein content.

The fraction corresponding to the void volume of the column
is indicated by the V0 symbol. The horizontal bar

( »———4) represents the fractions which lysed the target
cells resulting in no incorporation of [3H]thymidine;

these fractions were found to contain ammonium sulfate. The
vgrtical bar labeled UCM(SF) represents the average level of
[ H]thymidine incorporation (1 standard error of the mean)
in target cells treated with fractions obtained from
chromatography of serum free Unconditioned Medium; the
fractions corresponding to those indicated by the horizontal
bar also showed no incorporation and were excluded in the
calculation of this average.

71

II.“ Tmrusoa

 

VPV

 

b

---A.‘A--

AAAA‘A‘A‘A‘A‘AAA‘
' '

 

N

I! 1‘

3. 2 1

.18 x .58 :3 Hz...”—

‘1

m

we

3.: £30...

 

Fraction Number

DISCUSSION

The mouse fibroblast cell line, 3T3, exhibits a form of growth
control in vitro in that it reaches only a very low saturation density
and can remain for long periods of time in a viable but nondividing
state. From analyses performed in a number of laboratories, three
major hypotheses have been proposed to account for density-dependent
inhibition of growth. The first hypothesis suggests that cell-to-cell
contact is a necessary requirement for growth inhibition (16,131). A
second hypothesis suggests that depletion of growth factors, in the.
bulk medium (102) or in a micro diffusion barrier surrounding the cells
(107), causes the cessation of growth. The third hypothesis postulates
that soluble growth inhibitors are released into the bulk medium and
that these inhibitory factors accumulate to a sufficiently high
concentration to act on the target cells.

Although these three mechanisms are not necessarily mutually
exclusive, the demonstration of a growth inhibitory activity in CM
reported in our present studies strongly suggests that at least part of
the observed growth regulation at high cell densities may be mediated
by soluble inhibitory factors (Hypothesis III, Figure 2). Our results

suggest that the growth inhibitory factor in CM prepared and tested in

72

73

the 3T3 system has the following key features: (a) the inhibitory
activity can be demonstrated using three different assays of cellular
DNA synthesis and proliferation; (b) it is not cytotoxic and its
effects on cell growth are reversible; (c) the inhibitory activity can
be accumulated in the medium before the onset of extensive cell to cell
contact; (d) the inhibitor can be collected in the absence of serum;
(e) the activity is lost upon prolonged exposure to target cells; and
(f) the inhibitory factor can be concentrated and fractionated by
precipitation and gel filtration, respectively. We shall consider
below the interpretation of these facts in relation to the data
accumulated in support of the three major hypotheses on growth control
in cultured fibroblasts.

There have been a number of lines of evidence suggesting that
medium conditioned by exposure to density-inhibited 3T3 fibroblasts may
contain soluble factors that inhibit cell growth and division
(Hypothesis III, Figure 2). Experiments by Harel et al. (19) as well
as our own, have shown that co-cultivation of coverslips containing
dense and sparse cultures of 3T3 cells resulted in inhibition of cell
metabolism and DNA synthesis in Sparse cultures. These results have
implicated a soluble inhibitor released by the density-inhibited cells
to act on the sparse target cells. It has been reported that
lethally-irradiated 3T3 cells or medium conditioned by exposure to 3T3
cells can specifically suppress the growth and division of human and
mouse fibroblasts but not epidermal cells (17,18). Finally, the
experiments of Yeh and Fisher (20) on comparing the effects of fresh
medium and conditioned medium on contact-inhibited target cells have

implicated a heat stable, low molecular weight factor in their

74

conditioned medium that mediates growth inhibition. The molecular
nature of this factor has not been further characterized.

An alternative view postulates that growth is controlled by the
presence or absence of growth factors. Holley and Kierman (102) have
proposed that depletion of growth nutrients, particularly the serum
factors, may be responsible for the inhibition of growth and division
of fibroblasts at high cell densities (Hypothesis II, Figure 2). It
has also been suggested that the entry of essential growth factors into
3T3 fibroblasts at high cell densities may be limited by a diffusion
boundary layer surrounding the cells (107,108). Although the lack of
growth factors will undoubtedly result in a decrease growth rate, the
role of the depletion of growth factors in density-dependent inhibition
of growth is in question.

Thrash and Cunningham (111) reported that 3T3 cells at high
density ceased to grow due to limited growth surface. Furthermore, the
idea of diffusion boundaries limiting the entry of growth factors has
been challenged by recent experiments which demonstrated that changing
the viscosity of the medium had little effect on the growth rate and
final saturation density of 3T3 cells (94). These results support the
notion that decreased cellular proliferation in dense cultures is due
to the presence of growth inhibitors rather than the absence of
stimulating factors.

In accord with the idea of the presence of growth inhibitory
substance, we have observed inhibitory activity in CM(SF) even though
it is tested in the presence of freshly added serum. More importantly,
the fact that we can concentrate the inhibitory activity by ammonium

sulfate precipitation and fractionate it by gel filtration strongly

75

suggests that the failure of CM(SF) to support cell division of 3T3
target cells cannot be ascribed to a depletion of growth nutrients from
the medium by source cells. This conclusion is further supported by
the observation that the effects of CM were diminished upon prolonged
exposure to target cells, suggesting that the inhibitor may be
denatured or degraded. If the depletion of nutrients from CM were
responsible for its activity, it would be hard to envision mechanisms
by which these nutrients can be restored in the presence of target
cells. '

Finally, the observation that confluent 3T3 cultures showed little
or no growth while subconfluent cultures proliferated rapidly has
suggested that cell to cell contact (Hypothesis I, Figure 2) was an
important requirement in the mechanism of growth regulation (15).
Recent experiments have demonstrated that cell surface membranes or
extracts of the cell surface of 3T3 cells inhibited DNA synthesis in
the same cells (92,93,97). It was suggested that these surface
membrane molecules which inhibit normal cell proliferation may be the
same molecules that are responsible for contact-dependent growth
regulation.

The relationship between the growth inhibitory activity in the CM
of our present studies and the similar activity observed in the
membrane fractions has not been determined. It is possible that the
same molecule can exert its effects both anchored on the cell surface
or released into the medium (Hypotheses I and III, Figure 12). The
effects of both inhibitory activities appeared to be much more
pronounced on normal fibroblasts than on their virally-transformed

counterparts (92 and Table IV). These results suggest that the

76

inhibitory molecules have similar cellular specificities with respect
to their target cells. Although we can accumulate our inhibitory
activity from source cells grown at densities well below saturation
density, the effect of CM was the greatest on dense target cells. The
data also indicates that a critical concentration may be achieved at
saturation density by a combination of surface bound and soluble forms

of the molecule.

MOLECULAR PROPERTIES OF A FRACTION ENRICHED
IN GROWTH INHIBITORY ACTIVITY

In the previous Chapter, I documented the observation that
treatment of proliferating cultures of 3T3 cells with CM resulted in an
inhibition of growth and DNA synthesis in the target cells when
compared to parallel cultures treated with UCM. In addition, I
described the optimization of source cell parameters for assay. These
data provide the background necessary for a major effort in attempting
to purify the putative inhibitory factor(s).

In this Chapter, I describe the isolation and characterization of
a growth inhibitory fraction obtained from conditioned medium of 3T3

cells.

77

MATERIALS AND METHODS

Cell Culture and Preparation of Conditidned Medium

Swiss 3T3 cells (American Type Culture Collection, CCL 92) were
grown at 37° in Dulbecco modified Eagle's medium (K.C. Biological)
containing 10% calf serum (Microbiological Associates, Walkersville,
MD). The detailed protocol for the preparation of serum free Condition
Medium (CM(SF)) has been previously described. The following
modification of the previous protocol were made in order to minimize
the concentration of serum proteins in our CM(SF). Growth medium from
confluent 3T3 source cells was removed and replaced with fresh medium.
After 24 h, the growth medium was removed, the cells washed twice with
phosphate buffered saline, and then DME (l0 ml/l50 cm2 of growth
area) was added to the cultures. CM(SF) was collected from the
cultures 24 h later and processed as previously described.

For experiments that required the source cells to be incubated
with radioactive compounds, the labeled precursor was added to the
medium during the production of CM(SF). The 14C-labeled precursors
were added to a final concentration of 3.3 uCi/ml for
[2-14CJthymidine (51.4 mCi/mmol, New England Nuclear),

[2-14C]uridine (58 mCi/mmol, Schwarz-Mann), and to 5 uCi/ml for

78

79

L-[u-l4c1amino acid mixture (1.87 mCi/mg, ICN Pharmaceuticals,
Cleveland, OH). D-[2-3HJmannose (20-30 mCi/mmol, New England
Nuclear) was used at a final concentration of 1 mCi/ml. To prepare
CM(SF) labeled with [35SJmethionine, source cells were labeled with
DME containing 3 ug/ml unlabeled methionine (one tenth of the
concentration normally found in DME) and l0 uCi/ml of

[35S]methionine (l0l4 Ci/mmol, New England Nuclear).

Assays of Growth Inhibitory Activity

Target cells were seeded at a density of 5 x 103 cells/cm2 in
a 24-well culture dish (2 cmZ/well, Costar) in DME containing 3 ug/ml
unlabeled methionine and 10% calf serum. The cells were then incubated
in [35$]methionine (1 uCi/well) for 24 h. The radioactive medium
was then removed and the fraction (0.45 ml) to be assayed for its
growth inhibitory activity was placed on the well along with 50 ul of
calf serum. After 24 h, the cells were pulsed with l uCi/well of
[3H]thymidine (1.9 Ci/mmol, Schwarz- Mann) for 3 h at 37°. The cells
were then rinsed twice with cold PBS, once with l0% trichloroacetic
acid (TCA) and then 0.55 ml of l% sodium dodecyl sulfate ($05) in 0.05
M NaOH was added. After incubation at 37° for 10 min, an aliquot (0.5
ml) was removed and added to 5 ml of scintillation cocktail containing
20 ul of 50% TCA. Autoradiographic analysis of cells labeled with

[3H]thymidine was carried out and analyzed as previously described.

Assays of Cell Viability
The viability of cells treated with UCM and CM were determined

while the cells remained attached to the plastic growth surface. After

80

removal of growth medium, the cells were incubated with trypan blue
(0.08% in PBS) for 10 minutes at room temperature. The staining
solution was then removed and the viable cells were counted using an
inverted microscope.

Target cells were also labeled with [14C]amino acids
(2 uCi/culture, 58 Ci/matoms of 14c, Amersham) for 24 h. After
washing, these cells were treated with UCM(SF), CM(SF), or fractions
derived from CM(SF) that contained growth inhibitory activity. At
various times thereafter, the cells were washed with PBS and were
dissolved in 100 pl of 1% $05. The amount of labeled target cells
remaining after the treatment with growth inhibitors was determined by

counting the 14C radioactivity in each culture.

Colony Formation Assay

3T3 cells were seeded in a very sparse configuration ( 150 cells
on 10 cm2 of growth surface) in 35 mm dishes (Falcon). The cultures
were incubated overnight to allow for cell attachment. The medium was
then removed and parallel cultures were treated with 0.81 ml of
UCM(SF), CM(SF), or FGR-s fraction plus 0.09 ml of calf serum. After
48 hours, the media were again removed and the cultures were refed with
a fresh batch of corresponding media and incubated for another 48
hours. The cultures were then washed and stained with 1% crystal
violet (Sigma) in 50% aqueous ethanol for one hour. The cultures were
washed three times with PBS. The number of colonies and the number of

cells per colony were then counted under an inverted microscope.

81

Fractionation and Gel Electrophoretic Analysis

CM(SF) (60 ml) was precipitated by the addition of ammonium
sulfate to a final concentration of 80% saturation. The precipitate,
obtained after centrifugation at 12,400 g for 15 min, was redissolved
in 1.5 ml of DME. Gel filtrations of the precipitated material were
carried out at 4° on a Sephadex G-50 column ( 90 x 1 cm) and a Sephadex
G-15 column (27 x 1 cm) equilibrated with DME. Several different
assays were carried out on the individual fractions from the columns.
First, the inhibitory activity of each fraction was assayed on growing
3T3 target cells by measurements of [3H]thymidine incorporation.
Second, 50 ul aliquots of each fraction were taken for the
determination of total radioactivity. Third, the TCA precipitable,
representing labeled macromolecular components, was determined by the
addition of 100 pl of each fraction onto GF/A (Whatman) filters,
washing twice with 10% TCA and once with absolute methanol. Finally,
the protein content of the fractions was also determined by the method
described by Lowry (129). Bovine serum albumin was used to establish a
standard calibration curve.

Protein fractions were labeled with 1251 using the method of
Greenwood et al. (132). After the labeling reaction, non-covalently
bound 1251 was removed by gel filtration on a column (38 x 2 cm)
of Biogel P-2 equilibrated with PBS.

Polyacrylamide gel electrophoresis in SDS was performed according
to the procedure of Laemmli (133), using 10% or l5% acrylamide running
gel and 5% acrylamide stacking gel. The gels were fixed and then
stained with Coomassie brilliant blue. After destaining, the gel was

subjected to fluorographic treatment as described by Bonner and Laskey

82

(134) using Kodak X-Omat R (XR-5) film. Alternatively, the gels were
sliced immediately after electrophoresis, solubilized by digestion with
30% hydrogen peroxide (60° for 5h), and then subjected to scintillation

counting.

Sensitivity of Growth InhibitoryiActivity to Hydrolytic Enzymes

The sensitivity of the growth inhibitory activity to various
enzymes was tested using proteases and nucleases coupled to solid
supports. Pronase coupled to carboxymethyl-cellulose beads and
ribonuclease A bound to polyacrylamide beads were purchased from Sigma.
Bovine deoxyribonuclease I (Worthington, Freehold, NJ) coupled to
Sepharose beads was prepared according to the method of Lazarides and
Lindberg (135).

All enzyme digestion experiments were carried out in pyrex,
siliconized test tubes precoated with bovine serum albumin to minimize
nonspecific adsorption of the proteins to the glass surface. The tubes
were incubated with bovine serum albumin (2 mg/ml) for one hour at 37°.
They were then washed two times with PBS and three times with DME. For
the pronase-bead experiments, 0.40 ml of UCM(SF), CM(SF), or FGR-s were
added to tubes containing 0.06 units of enzyme and the mixtures were
incubated at 37° for up to 18 hours. The highest concentrations tested
for ribonuclease-beads and deoxyribonuclease-beads were .01 units/ml
and .014 units/ml, respectively.

After the incubation, the reaction mixtures were centrifuged at
1470 x g for 5 min and the supernatant was removed from the pelleted
beads for tests of growth inhibitory activity and for gel

electrophoretic analysis. These procedures allowing us to effectively

83

remove the hydrolytic enzymes after the reaction period have obviated
the many problems associated with the addition of proteases and/or
protease inhibitors to the target cells, which in turn often changed
the proliferative properties of the cells. Parallel experiments were

carried out with UCM(SF) and CM(SF).

RESULTS

Fractionation of the Growth Inhibitory Activity in Conditioned Medium
We have reported previously the optimized parameters for the
collection of CM(SF) from source cells and for the assay of the growth

inhibitory activity on target cells. We have now carried out
fractionation of the growth inhibitory activity, resulting in a
preparation that yields two major protein components on polyacrylamide
gel electr0phoretic analysis (see below). The fractionation scheme
included ammonium sulfate precipitation and gel filtration on a column
of Sephadex G-50 (Table VII).

In these experiments CM(SF) was collected from source cells that
had been cultured in the presence of [35S]methionine. Thus the
protein components that were secreted or shed from the source cells and
accumulated in the CM(SF) were labeled with radioactivity. After each
step of the fractionation protocol, aliquots of material were taken for
the following assays: (a) determination of the protein content by
quantitating the total TCA-precipitable radioactivity due to 355
(Column 0, Table VII); (0) gel electrophoretic analysis in the presence

of SDS; and (c) assays of growth inhibitory activity.

84

85

The growth inhibitory activity of a CM-derived fraction was
determined by comparing the amount of [3H]thymidine incorporation on
a per cell basis in cultures treated with the fraction as compared to
UCM-treated cultures (percent inhibition = 100% - (CM/UCM) x 100). The
number of cells in the target cultures treated with CM-derived
fractions or with UCM was determined by the amount of
[35SJmethionine incorporated into the target cells prior to the
treatments with the fractions to be tested for activity. It should be
noted that since the CM(SF) was collected from 35S-labeled cells,
the CM-derived fractions that are added to the target cultures for
tests of activity also contained radioactivity due to 35S whereas
the UCM control cultures did not. We have found, however, that the
additional 355 carried by the CM-derived fractions did not affect
our determination of the number of target cells. This is mostly
because the 35S label carried by the CM-derived fractions was less
than 5% of the level of 35S radioactivity in the target cells and
much of this label could be removed by the PBS washing steps
immediately after the [3H]thymidine incorporation. The data obtained
from these assays are expressed as percent inhibition of DNA synthesis
(Column 8, Table VII). These data were used in the calculations of
total activity and specific activity of the fractionated samples
(Columns C and E, Table VII).

0n the basis of these assays, we found that the largest increase
in specific activity («v35 fold) was obtained by fractionating the
ammonium sulfate precipitate on a column of Sephadex G-50 (Figure 16).
The inhibitory activity was partitioned into two major components

(Components A and C, Figure 16a). Component A (Figure 16a), which was

86

TABLE VII. Activity, Specific Activity, and Recovery of Fibroblast Growth

Regulator at Various Stages of the Fractionation Protocol.

 

 

 

A B C D E F
Purification Total Total Specific
Step, Inhibitiona Activity Proteinb Activ t «over

. (%) - (cmelO' ) (XIOA)R (z )
CM(SF)c 53.3 6396 301 2.1 ---
(NH4) 2504
precipitated
Sephadex G-15 20.0 192 15.9 1.2 3.0
Sephadex G-50
Component A
(Fig. 1) 33.1 476 9.6 5.0 7.4
Sephadex G-50
Component C
(Fig. 1) 37.2 714 .94 75.9 11.1

 

b.

The growth inhibitory activity of CM-derived fractions is expressed as percent

inhibition of DNA synthesis due to .5 m1 3of the fractionated material. Percent

inhibition is defined as the amount of [3 H]thymidine incorporation on a per

cell basis in cultures treated with CM-derived fractions as compared to UCM-

treated cultures. The number of cells in the target cultures trgated with CM-

derived fractions and with UCM is determined by the amount of [3

methionine incorporated into the target cells prior to the treatments with the

fractions to be tested for activity.

The protein content of a fraction is quantitated in term? of the total

trichloroacetic acid precipitable radioactivity due to S in the

CM-derived fractions. In these experiments CM was collected from source cells

that had been culture 5in the presence of [35 SJmethionine.

When unfractionated, 55- -;abeled CM(SF) was assayed for inhibitory

activity, the amount of3 S radioactivity due to CM(SF) was too high to

agiow an accurate determination of the number of target cells prelabeled with
S] methionine. In this case, thg inhibitory activity was determined

simply by comparing the levels of E H]thymidine incorporation in CM(SF)-

treated and UCM(SF)-treated target cells, without normalization t0 the total

number of target cells. Tar et cultures prelabeled with [ SJmethionine

ggd then treated with CM(SF) and UCM(SF) for 24 h retained the same amount of
S-label (see Figure 22).

The growth inhibitory activity of the ammonium sulfate precipitate fraction

could not be measured directly because the high salt concentration lysed the

target cells. Therefore, this fraction was first desalted on a Sephadex G-15

column before the activity was determined.

87

Figure 16. Chromatography of the growth inhibitory activity derived
from CM(SF) on a column (90 x 1 cm) of Sephadex G-50
equilibrated with DME. The CM(SF) was collected from source
cells that had been cultured in the presence of
[3 5SJmethionine for 24 h. Fractions of 1.6 ml were
collected._

(a) Profile of the growth inhibitory activity assayed by the
inhibition of [ H]thymidine ([3 H]- Tdr) incorporation in
targgt cells. The data on the ordinate axis are expressed
as [ HJ-Tdr incorporation per cell in the target cultures.
The horizontal bars containing the letters A, B, C and 0
denote fractions which were pooled for analysis by
polyacrylamide gel electrophoresis. The fractions in pool 0
lysed the target cells resulting in no incorporation of

[ HJ-Tdr. The data represent the averages of triplicate
determinations + the standard error of the mean. The
vertical hatched bar at the right represents the average
level of [3 H]thymidine incorporation (+ standard error of
the mean) in target cells treated with fractions from the
column prior to the void volume (fractions 2-14).

(b) Profile of the protein content assayed by the Lowry
method (e——. ) or assayed by counting the
TCA-precipitable radioactivity due to [ 5SJmethionine
([3 5S]Met; o——-—o ). The vertical arrows indicate
the positions of elution of molecular weight markers:
bovine serum albumin (68,000); myoglobin (17,000);
ribonuclease T (11,000); and bacitracin (1,400).

88

e135 £22m

m m
d

 

 

 

 

zooxdcooz. 5... 7E ole APO. x can: 8.2 ”man.

 

Fraction Number

89

eluted at the void volume of the column, contained the major portion of
the protein of the sample. This conclusion was obtained both by
determining the protein content using the Lowry method or by
quantitating the amount of TCA-precipitable radioactivity due to
[35$]methionine (Figure 16b).

In contrast, a second component of growth inhibitory activity
(Component C, Figure 16a) was associated with a minute amount of
protein material. When TCA-precipitable 35$ radioactivity was
assayed, a small but reproducible peak was observed in the fractions
containing the growth inhibitory activity (Figure 16b). However, we
could not detect any protein material in these fractions using the
Lowry assay. Therefore, this Component C (Figure 16a) represents a
fraction enriched in terms of specific activity (Table VII). Moreover,
these data also suggest that the inhibitory activity associated with
Component C was a product synthesized by the 3T3 source cells and
not some serum-derived protein. The position of elution of Component C
on Sephadex G-50 suggested that the material contained polypeptide
chains with molecular weights of approximately 12,000.

We have also assayed for the growth inhibitory activity at the
level of individual cells by autoradiography after incorporation of
[3H]thymidine. The percent inhibition in this assay was determined
by comparing the fraction of labeled nuclei in CM-treated cultures with
the corresponding value in UCM-treated cultures. In UCM(SF)-treated
cultures, approximately 50% of the nuclei were labeled, consistent with
the value calculated on the basis of the cell cycle time (22 h (see
reference 136)), length of S-phase (8 h) and the length of the

[3H]thymidine pulse period (3 h). In contrast, the fraction of

90

labeled nuclei in cultures treated with CM-derived fractions were
significantly lower. The percent inhibition observed for CM(SF),
components A and C (Figure 16) were 33%, 19% and 26%, respectively.
Therefore, the inhibition of [3H]thymidine incorporation in target
cells treated with CM-derived fractions reflects a true reduction of
the percent of cells undergoing DNA synthesis rather than alterations

of the transport or pool sizes of the label.

Gel ElectrOphoretic Analysis and Identification of the Polypeptides
Associated with the Growth Inhibitory Activity

The materials containing growth inhibitory activity at various
stages of fractionation were desalted on columns of Sephadex G-15
equilibrated with water. The void volume fractions from these columns
were pooled, lyophilized, and then subjected to polyacrylamide gel
electrophoresis in $03. After electrophoresis, the gel was stained
with Coomassie blue; in addition, the gel was subjected to fluorography
to reveal 35S-labeled protein components.

The gel revealed a large number of proteins, as detected by
Coomassie blue staining and by fluorography for CM(SF), ammonium
sulfate precipitate and Component A (Figure 16a) from the Sephadex G-50
column (lanes 1-3, Figure 17a and b). In contrast, Component C (Figure
16a) yielded no Coomassie blue positive material (lane 5, Figure 17b)
and only two major bands on the fluorogram (lane 5, Figure 176). The
molecular weights estimated for the two bands (lane 5, Figure 17b) in
Component C (Figure 16a) were 10,000 and 13,000. These values are

consistent with the chromatographic behavior of the polypeptides

91

Figure 17. Polyacrylamide gel electrophoresis in $05 of CM(SF) and
various components derived from CM(SF) after fractionation.
The acrylamide concentration of the running gel was l5%.
The CM(SF) was collected from source cells that had been
cultured in the presence of [35$]methionine for 24 h.

(a) Photographs of Coomassie blue-stained gels.

(b) Fluorographs of 35S-labeled components in the same
gels.

In both (a) and (b), the panels are: (1) CM(SF); (2)
precipitate fraction of CM(SF) after ammonium sulfate
fractionation; (3) Component A; (4) Component 8; (5)
Component C; and (6) Component 0, respectiyely, of Figure
16. In panels 4-6, the maximal amount of 5S

radioactivity available ( 6,000 cpm/gel) was used. In
panels l-3, approximately 70,000 cpm/gel of 35$
radioactivity was used; this amount was chosen such that any
minor components, particularly those in the molecular weight
range of l0,000- l5,000, could be revealed. The arrows at
the left indicate the positions of migration of molecular
weight markers: bovine serum albumin (68,000); aldolase
(40,000); chymotrypsinogen A (25,000); cytochrome c
(l2,500); bovine pancreas trypsin inhibitor (6,400); and
insulin A chain (2,200).

MWx 103

68-

40>
25>

12.5»
6.4’
2.2-

68>

40'
25’

92

 

I e
12.5» -.

6.4'

2.2 »'

93

on Sephadex G-50 columns (Figure 16). Densitometric analysis of the
fluorogram of Component C (lane 5, Figure 17b) indicated that the two
bands accounted for 98% of the radioactive material in the gel. In
addition, a parallel experiment was carried out in which the gel sample
was not fixed and stained but was sliced into fractions immediately
after electrophoresis. These fractions were then solubilized, and
subjected to scintillation counting (Figure 18). More than 97% of the
radioactivity was recovered in the gel slices corresponding to the two
bands observed in the fluorogram of Component C (lane 5, Figure 170).
This result argues against the possibility that low molecular weight
components, representing major contaminants, could have diffused from
the gel during the fixation and staining procedures and therefore,
would have escaped detection in the fluorogram.

Finally, the material in Component C was labeled with 125I
and subjected to gel electrophoretic analysis. It was found that 98%
of the radioactivity was associated with two bands (Figure 19),
corresponding to those observed in the gels of the 35S-labeled
sample (lane 5, Figure 17b). This result suggests that there was no
other protein species in Component C that was previously undetectable
by Coomassie blue staining or [35SJmethionine labeling. All of
these results, coupled with the specific activity data (Table VII),
strongly suggest that we have successfully obtained a preparation
enriched for growth inhibitory activity. We shall hereafter refer to
this material (Component C, Figure 16a) as FGR-s (Fibroblast Growth

Regulator - soluble form).

94

 

1.0 ' .

 

 

353 Met (cpm x 10°)
' <2)
(I!

 

 

O 40 80
‘ Fraction Number

Figure 18. Polyacrylamide gel electrophoresis in $05 of 35S-labeled

Component C (Figure 1). Thg acrylamide concentration of the
running gel was 15%. The 3 S radioactivity profile was
obtained by slicing the gel immediately after
electrophoresis and then subjecting the solubilized
individual slices to scintillation counting. The arrows
indicate the positions of migration of molecular weight

markers:

bovine serum albumin (68,000); chymotrypsinogen A (25,000);
iytochgome C (12,500); and bovine pancreas trypsin inhibitor
6,400 .

95

 

O.|O i

0.05 r '

A580

 

 

 

. W

l

O 5 IO
Distance (cm)

 

 

 

Figure 19. Polyacrylamide gel electrophoresis in $05 of
1 51-1abe1ed Component c (Figure 16). The acrylamide
concentration of the running gel was l0%. The 1251

radioactivity profile was obtained as a densitometric
tracing of the fluorogram at 580 nm. The arrows indicate
the positions of migration of molecular weight markers:
bovine serum albumin (68,000); aldolase (40,000);
concanavalin A (26,000); and ribonuclease A (l3,000).

96

Chemical Characterization of FGR-s

Although we have identified the polypeptide chains associated with
FGR-s, the question of whether other chemical components may be
associated with the inhibitory activity remained to be addressed. To
determine the composition of the purified inhibitory fraction, parallel
cultures of source cells were grown in the presence of 14C-labeled
amino acids, thymidine, uridine and D-[2-3HJmannose. When CM(SF) was
prepared from source cells prelabeled with [140] amino acids and
then fractionated, we found 14C radioactivity in the fractions
corresponding to Components A and C (Figure 16a) in the elution profile
of the Sephadex G-50 colunn (Figure 20a). These results corroborate
the data obtained using the [35$]methionine label.

In contrast to these findings, collection of CM(SF) from source
cells prelabeled with [14C]thymidine or [14C]uridine did not
yield any radioactive components corresponding to the FGR-s fractions
(Figure 20c and d). These results suggest that there was no DNA or RNA
molecules associated with the inhibitory activity. This conclusion was
further corroborated by experiments in which the sensitivity of the
growth inhibitory activity to various enzymes was tested.

To carry out these experiments, the enzymes were first immobilized
on beads. The enzyme-bead complexes were then incubated with the
fractions to be tested for enzyme sensitivity. After the incubation,
the reaction mixtures were centrifuged and the supernatant was removed
from the pelleted beads for tests of growth inhibitory activity on
target cells and for gel electrophoresis. Incubation of FGR-s fraction
with pronase-beads followed by gel electrophoretic analysis showed a

decrease in the intensity of the radioactive bands (Mr = 10,000 and

Mr = 13,000). This was paralleled by loss of growth inhibitory

97

 

 

 

 

 

 

 

 

 

2 2
a ["0] 'Amino Acids c (“oi-Thymidin.
.3... 9. .9. .._..°
"9,.
9 | *- l- -l
X
E \
Q.
3
.>.~
Ii 4 p
g b ["CI-Mannose d ["C]- Ur idine
%
0
CK
E3 2~ - J1
l
O 30 60 O 30 60

Fraction Number

Figure 20. Chromatography of the growth inhibitory activity derived
from several different preparations of CM(SF) on columns (60
x l cm) of Sephadex G- 50 equilibrated with DME. The
CM(SF)‘ s were collected from source cells that. had been
cultured for 24 h in the presence of: (a) [14CJamino
acids (5 uCi/ml); (b) [2- H]mannose (l 4mCi/ml); (c)
[14C]thymidine (3. 3 uCi/ml) and (d) [14C]uridine
(3. 3 uCi/ml). Fractions of 1 ml were collected and 100 pl
aliquots of each fraction were assayed for radioactivity.
The horizontal bars containing the letters A, B, C and 0
denote the fractions which correspond to those pooled in
Figure 16. '

98

activity. After 18 h of incubation, no activity was detectable in our
assay system (Figure 21). In contrast, the activity was not affected
by treatment with deoxyribonuclease-beads and ribonuclease-beads.
Thus, the results of these enzyme sensitivity experiments provide
confirmatory data to those obtained in experiments using
radioactively-labeled precursors.

When CM(SF) was collected from source cells cultured for 24 h in
the presence of D-[2-3HJmannose, fractionation on the Sephadex G-50
column yielded radioactivity only in fractions corresponding to
Component A (Figure 20b). No radioactivity was detected in fractions
corresponding to Component C, even when the incubation medium contained
as much [2-3HJmannose as I mCi/ml. We have not tested as yet other
precursors for labeling glycoproteins such as L-[14C]fucose or

D-[14C]-glucosamine.

Viability and Reversibility of the Effects of FGR-s

Three series of experiments were performed to ascertain that the
inhibition of [3HJthymidine incorporation in target cells by FGR-s
was due to a true suppression of cell growth rather than to any
cytotoxic effects of the inhibitory fractions. First, the viabilities
of the cells, assayed by the trypan blue exclusion tests, were
identical for target cultures treated with FGR-s and with UCM(SF) for
up to 48 h. Because nonviable cells could have been lost from the
dishes, we prelabeled the target cells by culturing them in the
presence of [14C]amino acids. After treatment with FGR-s for 24
and 48 h, the levels of 14C label retained were found to be similar

to those observed for control cultures (Figure 22). The same results

99

 

 

 

 

 

 

acre ':3:1 11.0
1 /
'7
" jjéj an "
in 7 18
<
5 s 3/ .0, 5
g; 25‘ 77: ' E 5
// ~ -
‘ r/zf ’ .8
ii 7;: 36
757 Q
o/ 18 '

 

Time of incubation (h)

Figure 21. The effect of pronase treatment on the growth inhibitory
activity of FGR-s (open bars) and on the total intensity of
the FGR-s bands (Mr = 10,000 and Mr = %3,000) on
polyacrylamide gels (hatched bars). 3 S-labeled FGR-s
(10.45 ml) was treated with pronase coupled to
carboxymethylcellulose beads (0.06 units). One unit of
enzyme activity hydrolyzes 1 umole of N-benzoyl-L-arginine
ethyl ester per minute at pH 7 and 30°. Samples were tested
for growth inhibitory activity and were subjected to
polyacrylamide gel electrophoresis and fluorography as
described in Materials and Methods. The intensities of the
protein bands (Mr = 10,000 and Mr = 13,000) were

obtained by densitometric tracing of the fluorogram at 580
nm.

100

 

 

 

 

A 4
'1’
.9
x
E
Q.
8 -
>5
5: :2-
.2
8 .
U
.9.
U
U
8..
so I
' O 25 50

1 (hours)

Figure 22. The effect of various CM(SF)-derived components on the
viability of cells in target cultures as assayed by the
retention of [14C] amino acids previously incorporated
inzo cellular proteins. Target cells were cultured in
[1 CJamino acids (2 uCi/culture, 24 h), washed, and then
treated with CM(SF) ( .———. ), UCM(SF) ( o——_o ), FGR-s
(0—4 ), and fractions 10 ( H) and 16 ( I———e ~
from the Sephadex G-50 column shown in Figure 16. At the
times indicated, the amount of labeled target cells was
determined by counting the 14C radioactivity in each
culture. The data represent the averages of duplicate
determinations.

101

were also obtained when target cells were treated with CM(SF) and
UCM(SF).

In another series of experiments, 3T3 cells were seeded onto
culture dishes in a very sparse configuration such that individual
cells would give rise to colonies. After overnight incubation, the
medium was removed and parallel cultures were treated with UCM(SF),
CM(SF), and FGR-s for 96 hours. When the number of cells in each
colony and the number of colonies were counted, the control culture
(UCM(SF)) showed a wide variation in population distribution, ranging
from 10 cells per colony to 150 cells per colony (Figure 23a). In
contrast, when the sparsely distributed cells were treated with CM(SF)
or isolated FGR-s fraction, the population distributions were narrow
and the number of colonies with fewer than 40 cells per colony
increased significantly (Figures 23b and c).

Studentis-I tests were used to compare the means of the
distributions shown in the histograms of Figure 23. The mean values
for the distributions.of UCM(SF) and CM(SF) were significantly
different (t_= 6.17, d.f. = 227, p < .001). Similarly, the mean values
for UCM(SF) and FGR-s were significantly different (t_= 5.02, d.f. =
246, p < .0001). In contrast, the mean values for CM(SF) and FGR-S
showed significant similarity (t_= 1.29, d.f. = 268, p < 0.2).

Most importantly, we have found that there was no difference in
the total number of colonies in dishes treated with UCM(SF), CM(SF), or
FGR-s. All of these results suggest that CM(SF) and FGR-s inhibited
the growth of cells, thereby decreasing the size of the colonies
without affecting the overall number of the originally attached cells.

Moreover, the data provide a confirmation of the growth inhibitory

Figure 23.

102

 

 

 

a
I

 

L

Numburofcohnke
LL_§L,.

 

IS-

  

 

 

 

 

‘*T60 .
Nurber of cells/colony

Cell count distributions for colonies of 3T3 cultures
treated with (a) UCM(SF); (b) CM(SF); and (c) FGR-s. The
following numerical data were obtained: (a) UCM(SF), 146
total colonies counted, average number of cells per colony,
72; (b) CM(SF), 143 total colonies counted, average number
of cell per colony, 44; and (c) FGR-s, 129 total colonies
counted, average number of cells per colony, 48.

103

activity of FGR-s using an assay that is independent of [3H]thymidine
incorporation.

To demonstrate that the inhibitory effect of FGR-s on DNA
synthesis was reversible, parallel target cultures were treated with
FGR-s for 24 h. The medium was then removed and replaced with an equal
volume of fresh growth medium. At various times thereafter, DNA
synthesis was assayed by the incorporation of [3HJthymidine. The
data showed that the inhibition was reversible within 20 h after
removal of FGR-s (Table VIII). Taken together with the demonstration
that the viability of target cells was not affected by FGR-s, these
results suggest that the inhibitory activity of FGR-s cannot be

ascribed to cytotoxicity.

104

TABLE VIII. Reversibility of the Effects of CM(SF) and FGR-s

 

 

 

 

Conditions 24 h l2 h after reversal 24 h after reversal
(% of ContFEl)
CM(SF) 56 l05 l25
FRG-s 54 T30 127

 

a. Target cells (5 x l03 cells/cmz) were treated for 24h with
CM-derived and UCM fracti ns. DNA synthesis was then measured
by the incorporation of [ H]thymidine. The results are
expressed as percent of control ([3HJthymidine incorporated in
cgltures treated with CM-derived fractions divided by
[ H]thymidine incorporation in UCM cultures).

0. After the cultures were treated for 24 h with CM-derived or UCM
fractions, the medium was removed and replaced by an equal
volume of fresh growth medium. DNA synthesis in these cultures
was measured l2 h and 24 h later.

DISCUSSION

A number of lines of evidence have been accumulated in various
laboratories to indicate that the phenomenon of density-dependent
inhibition of growth in cultured 3T3 fibroblasts may be mediated, at
least in part, by the accumulation of endogenous inhibitory factors in
the medium as the cell number increases. Studies of Canagaratna and
Riley on the patterns of nuclear incorporation of radioactive thymidine
in cultures with local cell densities between 0.2 x l04 and 6.2 x
104 cells/cm2 indicated that DNA synthesis in these cells was
critically dependent on the local cell density (112). More detailed
analyses of the data showed that there was an inverse relationship
between the local cell density and the proportion of labeled cells and
that this density dependent regulation of DNA synthesis was exhibited
in relatively sparse cultures, well before the onset of cell-to-cell
contact (113,137).

In addition, Harel et al. (19) have shown that ph05phate
metabolism and cell growth in sparse cultures of 3T3 cells were
inhibited when they shared the same medium with dense cultures. We
have recently confirmed these observations using assays of cellular DNA

synthesis. Moreover, we have analyzed the inhibition at the level of

105

106

individual cells by autoradiography. It was shown that the inhibition
of DNA synthesis reflected a decrease in the percent of the cells
traversing the S-phase of the cell cycle.

Finally, the experiments of Yeh and Fisher (20) comparing the
effect of fresh medium and conditioned medium on contact inhibited
target cells have implicated a heat stable, low molecular weight factor
that mediates growth inhibition. We have also demonstrated that
treatment of sparse proliferating cultures of 3T3 cells with medium
conditioned by exposure to density-inhibited 3T3 cultures resulted in
an inhibition of growth and division in the target cells when
compared to similar treatment with unconditioned medium. All of these
results are consistent with the notion that a soluble inhibitor may
mediate the cessation of growth in cultures of high cell densities.

Despite these well-documented phenomenological observations,
however, analyses of the mechanisms of density-dependent inhibition of
growth are obviously limited by the lack of purified preparations of
the putative growth regulatory molecule(s). The results obtained in
the present study indicate that we have succeeded in isolating from the
3T3 fibroblast system a growth inhibitory preparation enriched in
specific activity. This preparation has the following key properties:
(a) it is an endogenous cell product, synthesized by the 3T3 cells and
shed into the medium; (b) it is a protein and its activity is sensitive
to treatment with pronase; (c) the constituent polypeptide chains have
molecular weights of l0,000 and 13,000; (d) it binds to the target 3T3
cells (see Chapter IV) ; and (e) it is not cytotoxic to the target cell
and its effects on cell growth are reversible. Although it is still

possible that the growth inhibitory activity is due to molecules

107

present in minute amounts well below our detection levels, the best
candidates for this activity appear, at the present, to be these two
protein components. We propose to designate this factor FGR-s, which
stands for Fibroblast Growth Regulator that is secreted or shed into
the medium in a soluble form.

Our most highly purified preparation of FGR-s yielded two major
protein components on polyacrylamide gel electrophoresis in sodium
dodecyl sulfate (Mr = 10,000 and Mr = l3,000, panel 5, Figure 17b).
The growth inhibitory activity of this material may be due to (a) one
of the polypeptides acting alone; (b) both of the polypeptides acting
independently; or (c) a complex of the two polypeptides associated
together, either one of which is not sufficient for activity. The
resolution of these possibilities must await the successful separation
of these two components under non-denaturing conditions so that the
activity of each polypeptide and the reconstituted system can be
compared. In the next chapter, we provide evidence that at least one
of these two FGR-s polypeptides (Mr = 10,000) is stably bound by the
target cells.

Moreover, we have yet to determine the relationship of the two
polypeptides to each other. For example, one (Mr = l0,000) may be a
fragment derived from the other (Mr = l3,000). In this connection,
it is also possible that one or both of the polypeptides in Component C
(Figure 16) represent proteolytic products of larger precursors
observed in Component A (Figure 16). This hypothesis is supported by
preliminary experiments which indicate that prolonged incubation of
CM(SF) at 37° followed by chromatography on Sephadex G-50 generated

more material eluting in the molecular weight range of Component C

108

(Figure 16). If our isolated FGR-s indeed represents fragments derived
by proteolytic cleavage of a larger polypeptide, then it would be
another example in which proteins isolated on the basis of biological
activity turned out to be protease-digested products of the native
molecule. This situation recalls the molecular weight differences
reported for gap junction preparations (138). In addition , it call
for caution when comparing the molecular weight of FGR-s with other
reported growth inhibitors: (a) hamster melanocytes, 160,000 (114-116);
(b) WI-38 fibroblasts, 30,000 (119,120); and (c) monkey epithelial
cells, (10,000 (123,124).

Recently, Whittenberger and Glaser have shown that the growth of
3T3 cells can be reversibly inhibited by a surface membrane fraction
from the same cells (92) and that the inhibitory components can be
solubilized by the nonionic detergent octylglucoside (93). Natraj and
Datta have also shown that an inhibitor of DNA synthesis can be
extracted from 3T3 cells by treatment with 0.2 M urea in PBS (97). It
was suggested that these surface membrane molecules which inhibit cell
proliferation may be the same molecules that are responsible for
contact-dependent growth regulation.

In both of these studies on the membrane bound growth regulatory
factor(s), no purified preparation of the inhibitory molecule has, to
date, been reported. It would be of obvious interest to establish the
relationship between the FGR-s from CM and the similar activity
observed in the membrane fractions. It is possible, for example, that
the same molecule can exert its effects both anchored on the cell
surface or released into the medium. The isolation of FGR-s and the

characterization of its polypeptide chains represent the first step in

109

the parallel lines of experiments on membrane bound and soluble forms
of growth regulatory factors that mediate density dependent inhibition

of growth.

BINDING INTERACTIONS OF A GROWTH INHIBITORY
ACTIVITY WITH TARGET CELLS

In the preceding chapter, I described the isolation of a growth
inhibitory fraction that is enriched in specific activity and that
yields two major components on polyacrylamide gel electrophoresis
analysis. The availability of a highly purified preparation of this
growth inhibitory activity presents the opportunity to study one
mechanism of density-dependent growth control at the level of direct
interaction of an endogenous growth regulatory factor with its target

cells. These data are documented in the present chapter.

110

MATERIALS AND METHODS

Cell Culture and Preparation of FGR-s

 

Swiss 3T3 cells were grown at 37° in Dulbecco modified Eagle's

, medium containing 10% calf serum. The protocols for the preparation of
serum free Conditioned Medium and of FGR-s have been described. To
prepare FGR-s labeled with [35S]methionine for the binding

experiments, source cells were labeled with DME containing 3 ug/ml
unlabeled methionine (one tenth of the concentration normally found in
DME) and 20-100 uCi/ml of [35S]methionine (1014 Ci/mmol, New

England Nuclear). Serum free Conditioned Medium and FGR-s were
obtained from these source cells and assayed for radioactivity as well
as for growth inhibitory activity using the procedures reported
previously. The protein content of the inhibitor fractions was
determined by the method of Schaffner and Weissman (I39). Ribonuclease

A (Sigma) was used to establish a standard calibration curve.

Assays of FGR-s Binding to Target Cells

 

Confluent monolayer cultures of 3T3 fibroblasts (1.2 x 106 cells

on 25 cm2 of growth surface) were washed twice with serum free DME at

111

112

or without 0.1% bovine serun albumin (BSA). Although the absolute
level of binding in the presence of BSA was in general 15% lower than
in the absence of the protein, we have obtained essentially the same
overall results under both sets of conditions. The target cells were
incubated with 1 ml of [35$]methionine labeled FGR-s fraction
(specific radioactivity, 1.4-4.9 x 105 cpm/ug) for 60 min. at 37°.
This incubation time was chosen for most of the binding experiments
because it represents a compromise between the attainment of saturation
in the kinetics of the binding of FGR-s and the possible endocytotic
uptake and degradation of both FGR-s as well as other growth factors
(see below and references 140-142).

After incubation, unbound radioactivity was removed by washing the
cell monolayer five times with 2 ml of DME. The cells were then
solubilized by addition of 1 ml of 1% sodium dodecyl sulfate (SDS) and
incubation at 37° for 10 min. The contents of the dishes were
transferred to counting vials containing 10 ml of scintillation
cocktail and assayed for 35S radioactivity.

The effects of calf serum, Epidermal Growth Factor (EGF),
Fibroblast Growth Factor (FGF), and Platelet-derived Growth Factor
(PDGF) on the binding of [35SJFGR-s to target 3T3 cells were
determined. For these experiments, the inhibitors were added to the
target cells immediately prior to the addition of the radiolabeled
FGR-s. The purified growth factors EGF, FGF, and PDGF were obtained
from Collaborative Research (Waltham, MA). I

To quantitate the nonspecific adsorption of [35S]FGR-s to the
plastic tissue culture dish, the following control experiment was

performed. Dishes containing no cells were incubated with growth

113

mediun (DME containing 10% calf serun) for one h'at 37°. The dishes
were then washed with serum free DME and binding studies were carried
out in DME containing 0.1% BSA. The dishes were washed and assayed for
radioactivity in the soluble fraction of the 1% SDS extraction as
before.

We have also determined the binding of cytochrome C (Mr=12,500)
to the target cells as a control ligand. Highly purified cytochrome C
was a gift of Dr. S. Ferguson-Miller. It was labeled with 14C
using [14C]succinic anhydride and reaction conditions described
for the derivatization of concanavalin A (143). The specific activity
of the 14C-labeled cytochrome C was 2.3 x 105 cpm/ug.

Finally, binding assays were also carried out on SV 40 virus
transformed 3T3 fibroblasts and on sheep erythrocytes. For the virally
transformed fibroblasts, the protocol described for 3T3 cells was used.
The procedures for assaying ligand binding of FGR-s to erythrocytes

have been described elsewhere (143).

Autoradiographic Analysis of FGR-s Binding

Cells were cultured in Leighton tubes (5 cm2/slide). Confluent
monolayers were treated with 35S-labeled FGR-s for one h; after
this incubation, the slides were washed five times with phosphate
buffered saline (pH 7.2). They were then treated for 30 min with 2%
glutaraldehyde in phosphate-buffered saline at room temperature. After
washing in water, the slides were dipped in NTB-2 Nuclear Track
emulsion (Eastman Kodak Co.) and exposed for six weeks. After final
fixation in methanol, slides were stained in Giemsa (GIBCO) and
observed under a microscope. Photographs were taken using Kodak Tri-X

pan 135 film.

114

Analysis of Cell Bound FGR-s Polypeptides by Gel Electrophoresis

In order to identify the 35S-labeled material that was bound
to the target cells, the following experiment was carried out. Cells
with bound [35SJFGR-s were washed five times with DME and then were
extracted with 1.5 ml of 0.1 N HCl containing 0.1% BSA (140). After
incubation at room temperature for 90 min, the soluble material was
collected and lyophilized. The radioactivity of both the solubilized
fraction and residue was determined.

The solubilized material was subjected to polyacrylamide gel
electrophoresis in $05 (133). The acrylamide concentration of the gel
was 15%. After electrophoresis, the gels were fixed and stained with
Coomassie brilliant blue and then were subjected to fluorographic
treatment as described by Banner and Laskey (134) using Kodak X-Omat R
(XR-5) film.

RESULTS

Binding of FGR-s to Target Cells

 

In the previous chapter, we described the isolation from medium
conditioned by 3T3 cells, of a fraction that inhibited the growth and
proliferation of target fibroblasts. This preparation, designated
FGR-s, consisted of two polypeptide chains (Mr=10:000 and
Mr=13,000). When confluent monolayers of 3T3 cells were exposed to
35S-labeled FGR-s for varying amounts of time, washed, and the
radioactivity assayed, there was a monotonic increase in the
cell-associated 35S label with increasing time. Maximal binding
was achieved after incubation at 37° for three hours (Figure 24).

There was no apparent increase in cell bound radioactivity with further
incubation. A similar binding kinetics was observed at 4°. The level
of binding at 4° was, however, considerably lower than that observed at
37°.

Because the radioactive species used in these experiments were
[35$]methionine-labeled FGR-s polypeptides, it was possible that
the radioactivity was being degraded into component amino acids and
reincorporated into cellular proteins. Several experiments were

performed to test this possibility. First, the 35S-labeled FGR-s

115

116

 

31
JS
1

 

 

[35% FGR-s bound (cpm x 10"

 

 

 

Time (h)

Figure 24. Kinefiics of the binding of 35S-labeled FGR-s fraction (6
x 10 cpm/culture; specific radioactivity, 2.6 x 105
cpm/ug) to confluent monolayers of 3T3 target cells.
a-----., 37°; o—o, 4°. The experiments were performed
in the presence of 0.1% bovine serum albumin. Data points
are the averages of measurements on duplicate cultures.

117

material that was not bound by the cells was collected, concentrated,
and refractionated on a column of Sephadex G-50. There was no
radioactive label eluted at a position corresponding to free
[35$]methionine. Second, binding studies carried out in the
presence of cycloheximide (10 ug/ml) yielded results similar to those
shown in Figure 24. Finally, we analyzed the cell bound radioactive
components by polyacrylamide gel electrophoresis. The bulk of the
radioactivity was associated with one polypeptide chain (Mr=10,000)
rather than with a variety of cellular proteins (see below).
Therefore, the bound 35S radioactivity most probably does not
represent reincorporation of the label into newly synthesized
proteins.

Control experiments were also carried out to ascertain that the
observed binding of [35$]FGR-s represented interaction of the
ligand with target cells rather than nonspecific adsorption on the
cells or on the tissue culture dish. When 35S-labeled FGR-s was
incubated with a dish containing no cells, there was very little
radioactivity bound (Table IX). In addition, 14C-succinylated
cytochrome C (Mr=12,500) showed little or no binding to the target
3T3 fibroblasts (Table IX). In this experiment, approximately 3% of
FGR-s ligand added was bound to the target cells while the
corresponding number for cytochrome C was 0.1-0.2%. We have obtained a
similar conclusion using an irrelevant fraction derived from the
purification of FGR-s. This material, which contained 355
radioactivity but did not exhibit any growth inhibitory activity,
showed much less binding to the cells (0.1-0.2%). We have also tested

the binding of FGR-s on several different cell types including SV 40

118

TABLE IX. Specificity of the Binding of FGR-s to Target Cells

 

1 11 111 IV V VI
Experiment Target Cells Ligand Ligand Added Ligand Bound Ligand Bound
(cpm) (CPM) (%)
1 None [35$]FcR-s* 93,600 356 0.4
2 3T3 [35$]FGR-s 93,600 2,657 2.8
3 3T3 [14c3succiny1- 123,000 171 0.1
cytochrome C ,
4 sv 40 373 [35$]FGR-s 93,600 1,084 1.2
s sheepf , [35$]FGR-s 93,600 0 0
erythrocytes .
6 3T3 [35$]FGR-s 93,600 108 0.1
plus FGR-s . .

 

*The specific radioactivity of the [35SJFGR-s preparation was 2 x
105 cpm/ug- .

The specific radioactivity of the [14cjsuccinyl-cytochrome C was 2.3 x
105 com/us. - -

Unlabeled FGR-s was present at-a concentration 15 times that of
[35$]FGR-s.

119

virus transformed 3T3 cells, and sheep erythrocytes (Table IX). Only
the SV 40 virus transformed 3T3 fibroblasts showed appreciable binding
of FGR-s although the level was less than half that of 3T3 cells.
Finally, when the binding of 35S-labeled FGR-s was carried out
in the presence of 15-fold excess of unlabeled FGR-s, the amount of
radioactive ligand bound was reduced by about 95% (Table IX). This was
expected simply on the basis of dilution of specific radioactivity of
the labeled FGR-s. A similar competitive inhibition was observed when
target cells were precoated with unlabeled ligand prior to the addition
of 35S-labeled FGR-s. All of these results strongly suggest that
our measurements of the amount of 35$ radioactivity on the 3T3

cells reflected the binding of FGR-s on its target cells.

Effect of Ligand Concentration and Target Cell Density

The binding of FGR-s was dependent on the concentration of ligand
added (Figure 25a). With increasing concentrations of FGR-s, there was
an increase in the amount of ligand bound to the cells. Moreover, the
growth inhibitory effect of increasing concentrations of FGR-s
paralleled the increase in the amount of binding observed. There was
more cell-bound ligand at 37° than at 4° at all ligand concentrations
tested. Saturation at 4° was obtained at a FGR-s concentration of
approximately 600 ng/ml; the corresponding value for 379 was a little
higher. At saturation, the supernatant was removed and added to a
monolayer of fresh cells. This "used" ligand solution was capable of
binding to fresh target cells at a level comparable to the original
FGR-s solution. These results indicate that the FGR—s remaining in the

medium was fully active.

120

Figure 25. Concentration dependence of the binding of 35S-labeled
FGR-s fraction to confluent monolayers of 3T3 target cells.
(a) Dose-response curves for the binding at 37° (o————o) and
at 4° (o————o). The specific rgdioactivities of the ligand
prgparations used were 4.9 x 10 cpm/ug at 37° and 1.6 x
10 cpm/ug at 4°. The experiments were carried out in the
presence of 0.1% bovine serum albumin. The binding data
were corrected for nonspecific binding of FGR-s to empty
culture dishes. Data points are the averages of measure-
ments on duplicate cultures. (b) Scatchard plot of the
binding data at 37° shown in Figure 25a. The data were
calculated on the basis of a polypeptide ligand with a
molecular weight of 10,000. The slope and x-intercept
yiglded, respectively, Ka = 5.9 x 10 M"1 and 4 x
10 receptors per cell. (c) Scatchard plot of the binding
data at 4° shown in Figure 25a. The data were calculated on
the basis of a polypeptide ligand with a molecular weight of
10,000. The slope and x—intercept yielded, respectively,

. Ka = 2.7 x 107 m-1 and 3 x 105 receptors per

cell.

121

 

 

A

N

 

 

FGR-s bound per IO‘s cells (no)

7% lo
Eon-é] (nglml x no“)

 

 

 

 

 

 

 

2 4 6 2 4
[Bouncfl (moles/cell no")

 

C.
,7, 1...
o
.92.:
\ Q -08
i" as
O I 44
L91 0 6

122

The binding data at both temperatures were plotted according to
the method of Scatchard (144) (Figures 25b and c). Least squares
analysis yielded the following parameters characterizing the
interaction between FGR-s and target cells: (a) Ka = 5.9 x 107
M'1 and 4 x 105 receptors/cell at 37°; (b) Ka = 2.7 x 107
M‘1 and 3 x 105 receptors/cell at 4°. These results suggest that
the extrapolated value for the total number of FGR-s receptors on 3T3
cells is about the same at both temperatures. However, the affinity of
the binding interaction was higher at 37° than at 4°. 'This difference
in affinity may account for the higher level of binding observed at 37°
(Figures 24 and 25a). Additional factors may contribute to the higher
level of cell-bound ligand at 37°, such as internalization of the
receptor bound ligand at this temperature.

The binding of [35SJFGR-s was also measured as a function of
cell density. The cell densities covered in this experiment ranged
from the saturation density of 3T3 cells at the high end (5 x 104
cells/cmz) to 2 x 103 cells/cm2 at the low end, a density well
below that used in our assays of growth inhibitory activity. There was
more ligand bound with increasing cell density (Figure 26a). On a per
cell basis, however, we could not detect any difference in the amount

of FGR-s bound (Figure 26b).

Autoradiographic Analysis of FGR-s Bound to Target Cells

In order to study the overall distribution of FGR-s bound to
target cells, cultures incubated with 35S-labeled FGR-s were
washed, fixed, and subjected to autoradiography. Under light

microscopy, the distribution of grains was found to be uniform over the

Figure 26.

123

 

 

 

 

 

(I
10
13 an.
Co:
§'e
X
«n
e g5“
83
2 §4-b
8,00
3 9
I ‘\ 29.
CE E A
(9
u_ ‘gi ~—4|——~ ‘_4.
l 1 LLLfii l L_

 

cells/cmzk I04)

The binding of 355-labeled FGR-s as a function of the

cell density of the 3T3 target cultures (9.8 x 104
cpm/culture; specific radioactivity, 1.5 x 105 cpm/ug).

(a) The total amount of radioactive ligand bound in cultures
of indicated cell density. (b) The amount of radioactive
ligand bound normalized by the total number of cells in
cultures of indicated cell density. The experiments were
carried out in the presence of 0.1% bovine serum albumin.
The binding data were corrected for nonspecific binding of
FGR-s to empty culture dishes. Data points are the averages
of measurements on duplicate cultures.

124

TABLE X. Quantitation of Radioactivity of Cell Bound FGR-s

 

I II III IV
Ligand Added Ligand Bound Ligand Extracted Ligand Remaining
(cpm) (cpm) (cpm) on residue
(cm)
93,840 3,410 2,400 + 991 .
(3.6%)* (70.4%) (29.1%) I

 

*This represents the percent of ligand added (column I) that was
bound on the target cells.

+This represents the percent of ligand bound (column II) that was
extracted by HCl.

1’This represents the percent of ligand bound (column II) that was not
extracted by HCl.

125

entire cell (Figure 27). We could not distinguish, at this level of
resolution, whether the ligand was exclusively bound on the outside of
the plasma membrane or whether some have been internalized. In any
case, there was no evidence that the label was preferentially localized
over any particular organelle such as the nucleus or at the junctions
between cells.

We have also found that the number of grains per cell was
approximately the same in all cases examined. In the experiment shown
in Figure 27, for example, the average nunber of grains per cells was
48 i 5. Therefore, with respect to FGR-s binding, there did not appear
to be any striking population heterogeneity, with one population
binding large amounts of the inhibitory fraction and another not
binding at all.

Finally, these studies also showed that there was no significant
difference in the morphology of the target cells treated with FGR-s or
with controls. This conclusion is supported by microscopic
observations made both-before and after the fixation and staining

procedures used for autoradiography.

Gel Electrophoretic Analysis of FGR-s Bound on Target Cells

Because the FGR-s fraction used in these binding experiments
consisted of two polypeptide components (Mr=10,000 and Mr=13,000),
it was of interest to determine which one or both of the polypeptides
was bound to the target cells. Therefore, cells with bound
[35SJFGR-s were washed and then extracted with HCl; approximately
70% of the bound radioactivity was solubilized (Table X). The soluble

material was lyophilized, subjected to gel electrophoresis and the gel

126

 

Figure 27. Representative photograph showing the binding of
35S-labeled FGR-s fraction on target cells as analyzed
by autoradiography. [35SJFGR-s (1.2 x 105
cpm/culture; specific radioactivity, 1.4 x 105 cpm/ug) was
incubated with target cells for 1 h at 37°. The cells were
washed, fixed and subjected to autoradiography for 6 weeks.
The bar represents 20 um. Magnification, X500.

127

was then analyzed by fluorography (134). The results showed that the
FGR-s material that-was bound to the cells and resolubilized was
predominantly the 10,000 molecular weight component (Figure 28).

A more quantitative analysis was carried out on the gel profiles
shown in Figure 28. The FGR-s fraction used for the binding
experiments showed two bands with the following distribution of
radioactivity: Mr=10,000 (60%) and Mr=13,000 (40%). After cell
binding and re-extraction, only the lower molecular weight component
was observed on the gels even after purposely exposing the fluorogram
for a long time (Figure 28c). These results suggest that the FGR-s
component binding stably to the target cells was the Mr‘102000

polypeptide.

Effect of Serum on the Binding and Functional Interactions of FGR-s

When the binding of 35S-labeled FGR-s to target cells was
determined at various concentrations of calf serum, a marked inhibition
of binding was seen at a serum concentration as low as 0.1% (’bD.05
mg/ml total protein) (Figure 29). At a serum concentration of 10% (m 5
mg/ml total protein), the serun level used in assays of the growth
inhibitory activity of FGR-s, the binding of the inhibitory fraction
was reduced to about 30% of that observed in the absence of serum. We
have also determined the growth inhibitory activity of unfractionated
Conditioned Medium and of FGR-s fraction as a function of serun
concentration. The results showed higher levels of activity at lower
serum concentrations, as would be expected on the basis of higher

levels of binding of the inhibitor (Figure 30).

128

 

a be

Figure 28. Polyacrylamide gel electrophoresis analysis of the

polypeptide composition of FGR-s fraction before and after
binding to 3T3 target cells. (a) 35S-labeled FGR-s

fraction before incubation with 3T3 cells (8400 cpm
electrophoresed); (b) radioactive material that was bound to
target cells, washed, and reextracted with 0.1 N HCl, 0.1%
bovine serum albumin (2000 cpm electrophoresed). In both
(a) and (b), the gels were subjected to fluorographic
exposure for 3 days. (c) same as (b) except the gel was
subjected to fluorographic exposure for 35 days.

129

 

 

 

 

Percent of Control Binding
an
o

 

 

 

[protein] (mg/ml)

Figure 29. Inhibition of the binding of FGR-s to target cells by calf
serum (o—o) and by bovine serum albumin (o—p).
Confluent monolayers of 3T3 cells were treated with calf
serum or bovine serum albumin at the concentrations
indicated. Immediately thereafter, 355-1abeied FGR-s
was added to the incubation mixture (1.2 x 105
cpm/culture). After 1 h, the cells were washed and the
amount of radioactivity bound to the cells was determined.
Control binding represents the level of binding observed in
the absence of any serum or bovine serum albumin. The
protein concentration of serum was determined to be 50
mg/ml.

130

 

END

4(a-

26L

inhibition ‘ (%)

 

 

 

o lo 20
[serum] (%)

Figure 30. The effect of serum concentration on the growth inhibitory
activity of serum frge Conditioned Medium assayed on 3T3
target cells (5 x 10 cells/cmz). The target cells were
treated with Conditioned Medium and Unconditioned Medium for
20 h in the presence of different concentrations of calf
serum. The cells were then pulsed with 3H-labeled
thymidine (0.1 uCi/culture) for 3 h to monitor DNA synthesis
by autoradiography. The per cent of growth inhibition was
determined by comparing the fraction of labeled nuclei in
cultures treated with Conditioned Medium with corresponding
fractions in cultures treated with Unconditioned Medium.
Data points represent the averages of triplicate
determinations i the standard error of the mean.

131

 

 

 

[3’53] Met (cpm 1: IO”)
0
U'

A
I
:2

 

 

 

 

 

0 25 , 50
Fraction Number

Figure 31. Chromatogr ahy of 35S-labeled FGR-s fraction on a column
(90 x 1 cm? of Sephadex G-50 equilibrated with Dulbecco's
modified Eagle medium in the absence (a) and presence (b) of
1% calf serum. Fractions of 1.6 ml were collected. The
vertical arrows indicate the positions of elution of
molecular weight markers: bovine serum albumin (68,000);
myoglobin (17,000); cytochrome C (12,500); and bacitracin

1,400 .

132

To test whether components in serum could interact directly with
FGR-s, resulting in either degradation or complexation, the inhibitory
fraction was chromatographed on columns of Sephadex G-50 in the
presence and absence of 1% serum (Figure 31). The positions of elution
for FGR-s under both sets of conditions were essentially identical.
Although the possibility that some low molecular weight serum component
can bind and inactivate FGR-s without drastically changing its
chromatographic properties remains to be explored, these experiments
suggest that FGR-s does not interact with a serum protein to form a
high molecular weight complex. In addition, the results also suggest
that FGR-s was not degraded by any serum proteases. This conclusion
was corroborated by experiments which showed that FGR-s incubated with
serun (1%) for 90 minutes and then subjected to polyacrylamide gel
electrophoresis yielded the same polypeptide bands (Mr=10,000 and
Mr=13,000). Therefore, the inhibition by serum of the binding of
FGR-s to target 3T3 cells is most likely explained by competition
between FGR-s and some serum components(s) for the target cell surface
receptors.

The specificity of the inhibitory effect of serum on the binding
of FGR-s to 3T3 cells was tested by searching for purified proteins
that may yield the same effect. Bovine serum albumin (BSA), the major
component in calf serum, showed little inhibitory effect at protein
concentrations equivalent to that found in serum (Fig. 29). The
concentration of BSA that was required to obtain the same level of
inhibition observed for 0.1% serum (0.05 mg/ml total protein) was 200
times higher. Thus, it appears that the inhibitory effect of serum on
the binding of FGR-s was a relatively specific competitive event rather

than a nonspecific protein adsorption effect.

133

Effect of Purified Growth Factors on the Binding of FGR-s

We have also tested the effect of several purified growth
stimulatory factors on the binding of FGR-s to target 3T3 cells. The
data showed that EGF, FGF, as well as PDGF, all inhibited the binding
of 35S-labeled FGR-s to the target cells. The maximal extent of
the inhibition and the concentration of the growth factor at which this
maximal inhibition was achieved are presented in Figure 32. The
inhibitory effect of the simultaneous addition of the various growth
factors was only slightly higher than that obtained with either EGF or
FGF alone (Figure 32).

134

 

 

 

 
  

 

   

     

 

 

 

,50- :5, E 2?.
a g .9 3
E40- 53 91 A ‘0
E ‘6 6 3
“iso— “‘1 “ E»

l

6 I? 8 %
UL ...

“52°' % 2 u. /
%
iglO’ a V

" A A .

 

 

 

Figure 32. Inhibition of the binding of 35S-labeled FGR-s to target
cells by purified growth stimulatory factors, Epidermal
Growth Factor (EGF), Fibroblast Growth Factor (EGF), and
Platelet-derived Growth Factor (PDGF). Confluent monolayers
of 3T3 cells were treated with the growth factors at the

oncentrations indicated. Immediately thereafter,

5S-labeled FGR-s (8.9 x 104 cpm/culture) was added to
the incubation mixture. After 1 h, the cells were washed and
the amount of radioactivity bound to the cells was
determined. The experiment labeled "cocktail" indicates that
all three growth factors were added at their respective
concentrations.

DISCUSSION

The experiments presented here document the specific interaction
of a fibroblast growth regulatory factor, FGR-s, with its target cells.
The key features of this interaction are: (a) there are approximately
3-4 x 105 receptors per cell; (b) the binding is greater at 37° than
at 4°; (c) on a per cell basis, approximately the same amount of ligand
is bound on dense cells as on sparse cells; (d) there is no evidence of
population heterogeneity in the binding of the ligand by the target
cells; and (e) the binding can be inhibited by calf serum, as well as
several purified growth factors such as EGF, FGF, and PDGF.

The specificity of the interaction between FGR-s and its 3T3
target cells was analyzed both in terms of the binding of irrelevant
ligands to the 3T3 cells as well as in terms of the binding of FGR-s to
other cell types and plastic surfaces. It was found that the level of
the binding of FGR-s to 3T3 target cells was well above that which can
be accounted for by these non-specific interactions. Moreover, the
level of binding was correlated with the extent of growth inhibition in
three different experiments. First, both the binding of FGR-s and its
growth inhibitory activity increased in parallel with increasing

concentrations of the ligand. Second, increasing the serum

135

136

concentration in the medium decreased both the binding and the growth
inhibition. Finally, we had previously reported that SV 40 virus-transformed
3T3 fibroblasts showed much less inhibition, when tested with the same growth
inhibitory activity, than the normal counterparts. We have now found that the
level of binding of FGR-s to the SV 40 cells was less than half that of normal
3T3 cells.

We have also found that when target cells were treated with the
FGR-s fraction, which contains two polypeptide chains (Mr = 10,000
and Mr = 13,000), washed to remove nonbound material, and reextracted
with HCl, only the polypeptide with a molecular weight of 10,000 was
recovered. Identical results were also obtained using acetic acid or
$05 for the extraction. Two main possibilities have been considered to
account for these observations: (a) The 10,000 molecular weight
polypeptide but not the 13,000 polypeptide is bound by the target
cells; (D) both polypeptides are bound but the 13,000 polypeptide is
internalized and degraded during the course of the incubation.
Experiments are now underway to distinguish these possibilities.

In any case, we have analyzed the present binding data on the
basis of an interaction between a polypeptide of molecular weight
10,000 and target 3T3 cells. There were approximately 3-4 x 105
‘receptors per cell, with an affinity constant of 5.9 x 107 M"1 at
37° and 2.7 x 107 M‘1 at 4°. This level of binding is comparable
to the binding observed in several other ligand-receptor systems,
particularly the binding of growth stimulatory factors such as EGF
(140,141), PDGF (145), somatomedin-C (146), and insulin (147) to their
respective target cells. In contrast to these systems, however, the

kinetics of the binding of FGR-s reached a plateau much more slowly.

137

Whereas the amount of EGF bound on target cells reached a peak within
one h, the FGR-s binding did not plateau until three to four h.

We had previously reported that the inhibitory action of
Conditioned Medium was more pronounced on dense target cells than on
sparse cells. One possible explanation for this observation was that
the receptor for the inhibitory factor was cryptic at low cell .
densities but appeared as the cells came into contact with each other.
(Our present measurements showed, however, that on a per cell basis,
there was little or no difference in the amount of inhibitory fraction
bound to the target cells. Therefore, the highly potent effect of
conditioned medium on cells at high density must be ascribed to some
other variable, such as the shedding of inhibitor into the assay medium
during the course of the experiment by the target cells themselves.

The most striking observation in the present study is the
inhibition by serum of the binding of FGR—s to its target cells. This
effect was specific inasmuch as BSA, the major protein component of
serum, showed little or no inhibition at concentrations as high as 5
mg/ml. Moreover, highly purified growth factors such as EGF, FGF, and
PDGF can partially mimic the serum effect at concentrations as low as
10-50 ng/ml. These results suggest: (a) growth stimulatory factors
may compete with growth inhibitory factors for cell surface receptors;
or (b) growth stimulatory factors may cause down regulation of
receptors for growth inhibitors.

Recently, Holley and co-workers have reported the partial
purification of cell growth inhibitors isolated from the kidney
epithelial cell line BSC-1 (124). The action of these cell growth

inhibitors was counteracted by the addition of calf serum or EGF. It

138

has been proposed that the growth depressing effects of the isolated
inhibitors and the growth promoting effects of EGF most probably
represent antagonistic intracellular biochemical signals rather than
competition for cell surface receptors. This conclusion was based on
the finding that the binding of 1251-1abe1ed EGF to the BSC-1

cells (110) was not affected by the presence of conditioned medium
containing the inhibitory factors.

Similarly, Glaser and co-workers have shown that a plasma
membrane enriched fraction of 3T3 cells inhibited DNA synthesis in the
same cells (92). This inhibitory activity could be solubilized by
octylglucoside (93). Moreover, the membrane-induced inhibition was
partially reversed by the addition of serum, EGF, or the combination of
FGF plus dexamethasone (148). Using 125I-labeled EGF, it was
shown that the membrane fraction that exhibits growth inhibitory
activity did not block the binding of EGF or cause the down regulation
of EGF receptors on the cell surface (95). In addition, recent data
also indicated that the membranes did not exert their inhibitory effect
by removing growth factors from the medium because membranes prepared
from 3T3 cell mutants that lack EGF receptors (NR-6 cells (149)) still
prevented the initiation of DNA synthesis induced by EGF (95,150). All
of these results argue for a mechanism by which the inhibitory
molecules exert their effect via the generation of a biochemical signal
which is antagonistic to the mitogenic signal induced by growth
stimulatory factors.

Our present data indicate, on the other hand, that serum and
growth factors will inhibit the binding of FGR-s to target cells.

These observations suggest that modulation of cell surface receptor

139

binding, either through direct competition or through down regulation,
by growth stimulatory and growth inhibitory factors may be an important
feature in growth regulation. The results from the converse
experiment, the binding of 125I-labeled growth factors in the

presence and absence of FGR-s should yield further information on the
details of this complex interaction. In any case, the isolation and
chemical characterization of both the membrane-bound inhibitor (92,93)
as well as soluble FGR-s should contribute to our understanding of the

mechanisms of growth control.

BIBLIOGRAPHY

11.

12.

13.
14.
15.

16.

BIBLIOGRAPHY

Prescott, D.M. (1976) Reproduction of Eukaryotic Cells, Academic
Press, New York.

Howard, A. and Pelc, S.R. (1951) Exp. Cell. Res. 2, 178-187.

Pardee, A.B., Dubrow, R., Hamlin, J.L., and Kletzien, R.F. (1978)
Ann. Rev. Biochem. 12, 715-750.

Padilla, G.M., Cameron, I.L., and Zimmerman, A., eds. (1974) Cell
Cycle Controls, Academic Press, New York.

Baserga, R. (1978) J. Cell Physiol. 22, 377-386.
Pardee, A.B. (1974) Proc. Natl. Acad. Sci. U.S.A. 22, 1286-1290.
Prescott, D.M. (1968) Cancer Res. 22, 1815-1820.

Smith, J.A. and Martin, L. (1973) Proc. Natl. Acad. Sci. U.S.A.
29, 1263-1267.

Shields, R. and Martin, L. (1977) J. Cell. Physiol. 22, 345-356.

Baserga, R. (1976) Multiplication and Division in Mammalian Cells
Marcel Dekker, New York.

Gospodarowicz, D. and Moran, J.S. (1976) Ann. Rev. Biochem. 52,
531-558.

Balazs, A. (1979) Control of Cell Proliferation by Endogenous
Inhibitors, Elsevier/North Holland Biomedical Press, Amsterdam.

Stoker, M.G.P. and Ruben, H. (1967) Nature (Lond.) 22g, 171-172.
Kolodny, G.M. and Cross, P.R. (1969) Exp. Cell Res. §2, 423-432.

Todaro, G.J., Lazar, G.K., and Green, H. (1965) J. Cell and Comp.
Physiol. §§, 325-334.

Todaro, G.J. and Green, H. (1963) J. Cell Biol. 21, 299-313.

140

17.
18.
19.

20.
21.

22.

23.

24.

25.

26.
27.

28.

29.

30.
31.

32.
33.

34.
35.
36.

141

Rheinwald, J.G. and Green, H. (1975) Cell 2, 317-330.
Rheinwald, J.G. and Green, H. (1975) Cell 2, 331-340.

Harel, L., Jullien, M., and Demonti, M. (1978) J. Cell Physiol.
22, 327-332.

Yeh, J., and Fisher, H.W. (1969) J. Cell Biol. 29, 382-388.

Lozzio, B.B., Lozzio, C.B., Bamberger, E.G., and Lair, S.V. (1975)
Int. Rev. of Cytol. 22, 1-47.

Rudland, P.S. and Jimenez De Asua (1979) Biochem. Biophys. Acta
222, 91-133.

Jakob, A., Hauri, C. and Froesch, E.R. (1968) J. Clin. Invest. 2,
119-1230

Humbel, R.E., Bunzli, H., Mully, K., Delz, 0., Froesch, E.R.,
Ritschard, W.J. (1971) Proc. Congr. Int. Diabetes Fed. 7th Int.
Congr. Ser. 222, 306-317.

Van Wyk, J.J., Underwood, L.E., Hentz, R.L., Clemens, R.D., Voina,
S.J., and Weaver, R.P. (1974) Rec. Prog. Horn. Res. 22, 259-295.

Goldwasser, E. (1975) Fed. Proc. 22, 2285-2292.

Ross, R., Glamset, J., Kariya, B., and Harker, I. (1974) Proc.
Nat]. Acad. 5C1. U.S.A. ll, 1207'1210.

Antoniades, H.N., Stathakos, D. and Scher, C.D. (1975) Proc. Natl.
Acad. SCio U.S.A. E, 2635‘26390

Pierson, R.W. and Temin, H.M. (1972) J. Cell Physiol. 22,
319-330.

Cohen, S. (1962) J. Biol. Chem. 222, 1555-1562.

Carpenter, G. and Cohen, S. (1979) Ann. Rev. Biochem. 52,
193-216.

Bradshaw, R.A. (1978) Ann. Rev. Biochem. 22, 191-216.

Gospodarowicz, D., Jones, K.L., and Sato, G. (1974) Proc. Natl.
Acad. SCio U.S.A. 7_1, 2295'2299.

Gospodarowicz, D. (1975) J. Biol. Chem. 229, 2515-2520.
Dulak, N.C. and Temin, H.M. (1973) J. Cell Physiol. 22, 161-170.

Austin, P.E., McColloch, E.A., and Till, J.E. (1971) J. Cell
Physiol. 22, 121-134.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.
48.

49.
50.

51.
52.

53.
54.

55.

142
Metcalf, D. (1973) Humoral Control of Growth and Differentiation,
Vol. 1, pp. 91-118, Academic Press, New York.

Jimenez de Asua, L., Clingan, D. and Rudland, P.S. (1975) Proc.
Nat]. Acad. 5C1. U.S.A. lg) 2824'2828.

Rechler, M.M. and Podskalny, J.M. (1976) Nature (Lond.) 222,
134-136.

Rechler, M.M., Podskalny, J.M. (1979) J. Biol. Chem. 222,
3898-3910.

Gospodarowicz, D., Greene, G., and Moran, J.S. (1975) Biochem.
Biophys. Res. Commun. 22, 779-787. ~

Carpenter, G. and Cohen, S. (1979) Ann. Rev. Biochem. 22,
193‘2160

Hollenberg, M.D. and Cuatrecassas, P. (1975) J. Biol. Chem. 222,
3845-3853.

Carpenter, G., Lembach, K.J., M0rrison, M., and Cohen, S. (1975)
J. Biol. Chem. 222, 4297-4304.

Das, M., Miyakawa, T., Fox, F.C., Pruss, R.M., Aharonov, A., and
Herschman, H. (1977) Proc. Natl. Acad. Sci. U.S.A. 22, 2790-2794.

Das, M., and Fox, F.C. (1978) Proc. Natl. Acad. Sci. U.S.A. 22,
2644-2648.

Carpenter, G. and Cohen, S. (1976) J. Cell Biol. 22, 159-171.

Carpenter, G., King, L., and Cohen, S. (1978) Nature (Lond.) 222,
409-410.

Ushiro, H. and Cohen, S. (1980) J. Biol. Chem. 222, 8363-8365.

Hunter, T. and Sefton, B.M. (1980) Proc. Natl. Acad. Sci. U.S.A.
22, 1311-1315.

Langan, T. (1980) Nature (Lond.) 222, 329-330.

Nimmo, H.G. and Cohen, P. (1977) Adv. Cyclic Nucleotide Res. 2,
145-266.

Weiss, P. and Kavanau, J.L. (1957) J. Gen. Physiol. 22, 1-47.

Bullough, W.S. and Laurence, E.B. (1960) Proc. Royal Soc. B 222,
517‘5360

Rytomaa, T. (1973) Nat. Cancer Inst. Manogr. 22, 143-146.

56.
57.

58.

59.

60.

61.
62.

63.
64.

65.
66.

67.

68.

69.

70.

71.

72.

73.

74.

143

Houck, J.C. (1978) J. Reticuloendothelial Soc. 22, 571-581.

Houk, J.C., Weil, R.L. and Shaima, V.K. (1972) Nature New Biol.
259, 210-211.

Simard, A., Corneille, L., Deschamps, Y., and Verly, W.G. (1974)
Proc. Nat]. Acad. SCio U.S.A. fl, 1763-1768.

Gonzalez, R. and Verly, W.G. (1976) Proc. Natl. Acad. Sci. U.S.A.
22, 2196-2201.

Bullough, W.S. and Laurence, E.G. (1974) Eur. J. Cancer 2,
607‘6150

Bullough, w.s. (1975) Bio. Rev. 22, 99-122.

Isaacs, A. and Lindenmann, J. (1957) Proc. Royal Soc. B 222,
258-267.

Zoon, K.C. (1980) Nature (Lond.) 222, 110.

De Maeyer, E., DeMaeyer-Guignard, J., and Vandeputte, M. (1975)
PT‘OCo Nat]. Acad. SCio U.S.A. E, 1753‘17570

Brouty-Boye, D. and letter, B.R. (1980) Science 222, 516-518.

Lindahl, P., Leary, P., and Gresser, I. (1973) Proc. Natl. Acad.
Sci. U.S.A. 22, 2785-2788.

Knight, E. and Korant, B.D. (1977) Biochem. Biophys. Res. Commun.
22, 707-713.

Pfeffer, L.M., Wang, E. and Tamm, I. (1980) J. Cell Biol. 22,
9-17.

Pfeffer, L.M., Wang, E. and Tamm, I. (1980) J. Exp. Med. 222,
469-474.

Pfeffer, L.M., Wang, E. and Tamm, I. (1979) Exp. Cell Res. 222,.
111-120.

Gresser, I., Bouvali, D., Levy, J.P., Brouty-Boye, D. and Thomas,
MOT. (1969) PT‘OCo Nat]. Acad. SCI. U.S.A. _6_3, 51'550

Gresser, I., Brouty-Boye, D., Thomas, M.T. and Macieira-Coello, A.
(1970) Proc. Natl. Acad. Sci. U.S.A. 22, 1052-1056.

Gresser, I. and Tovey, H.G. (1978) Biochem. Biophys. Acta 222,
231-247.

Gresser, I., De Maeyer-Guignard, J., Tovey, H.G., and De Maeyer,
E. (1979) Proc. Natl. Acad. Sci. U.S.A. 22, 5308-5312.

144

75. Duntze, W., Stotzler, D., Bucking-Throm, E. and Kalbetzer, S.

76. Hartwell, L.H. (1973) Bacteriol. Rev. 22, 164-198.
77. Stotzler, D. and Duntze, W. (1976) Eur. J. Biochem. 22, 257-262.

78. Abercrombie, M., and Heaysman, J.E.M. (1954) Exp. Cell Res. 2,
293‘3060

79. Temin, H.M. and Rubin, H. (1958) Virology 2, 669-688.

80. Vogt, M. and Dulbecco, R. (1963) Proc. Natl. Acad. Sci. U.S.A. 22,
171-179.

81. Rubin, H. (1966) Exp. Cell Res. 22, 138-145.
82. Kruse, P.F. and Midema, E. (1965) J. Cell Biol. 22, 273-281.

83. Rubin, A.H., Terasaki, M., and Sanui, H. (1979) Proc. Natl. Acad.
5C7. U.S.A. 19., 3917-3921.

84. Holley, R.W. and Kierman, J.A. (1979) Proc. Natl. Acad. Sci.
U.S.A. -7_1., 2942-2945.

85. Hatten, M.E., Horowitz, A.F. and Burger, M.M. (1977) Exp. Cell
Res. 222, 31-34.

86. Folkman, J. and Moscona, A.A. (1978) Nature (Lond.) 222, 345-349.
87. fisher, H.W. and Yeh, J. (1967) Science lgg, 581-582.

88. Schutz, L. and Mora, P.T. (1968) 0. Cell Physiol. 12, 1-6.

89. Kolandy, G.M. and Cross, P.R. (1969) Exp. Cell. Res. 22, 423-432.
90. Eagle, H. and Levine, E.M. (1967) Nature 222, 1102-1106.

91. Martz, E. and Steinberg, M.S. (1972) J. Cell Physiol. 22,
189-210.

92. Whittenberger, B. and Glaser, L. (1977) Proc. Natl. Acad. Sci.
U.S.A. 22, 2251-2255.

93. Whittenberger, 8., Raben, D., Lieberman, M.A. and Glaser, L.
(1978) Proc. Natl. Acad. Sci. U.S.A. 22, 5457-5461.

94. Whittenberger, B. and Glaser, L. (1978) Nature (Lond.) 222,
821-8230

95. Leidermann, M.A., Rothenberg, P., Raben, D.M. and Glaser, L.
(1980) Biochem. Biophys. Res. Commun. 22, 696-702.

96.

97.

98.

99.

100.

101.
102.

103.

104.
105.
106.
107.
108.
109.

110.

111.

112.

113.

114.

115.

145
Weston, J. and Roth, S. (1969) Cellular Recognition,
Appleton-Century-Crofts, New York, pp. 29-37.

Natraj, C.V. and Dalta, P. (1978) Proc. Natl. Acad. Sci. U.S.A.
22, 6115-6119.

Pariser, R.J. and Cunningham, D.D. (1971) J. Cell Biol. 22,
525-529.

Liederman, M.A., Raben, D.M., Whittenberger, B. and Glaser, L.

Martin, R.G. and Stein, S. (1976) Proc. Natl. Acad. Sci. U.S.A.
22, 1655-1659.

Eagle, H. (1965) Science 222, 42-50.

Holley, R.W. and Kierman, J.A. (1968) Proc. Natl. Acad. Sci.
U0 50A. £9, 300‘3040

Dulbecco, R. and Elkington, J. (1973) Nature (Lond.) 222,
197-199.

Holley, R.W. (1975) Nature (Lond.) 222, 487-490.
Smets, L.A. (1971) Cell Tissue Kinet. 2, 233-240.
Dulbecco, R. (1970) Nature (Lond.) 222, 802-806.
Stoker, M.G.P. (1973) Nature (Lond.) 222, 200-203.
Stoker, M. and Piggott, D. (1974) Cell 2, 207-215.

Lipton, A., Klinger, 1., Paul, D., and Holley, R.W. (1971) Proc.
Nat]. Acad. SClo U.S.A. fl, 2799‘28010

Holley, R.W., Aronour, R., Baldwin, J.H., Brown, K., and Yeh, Y.

Thrash, C.R. and Cunningham, D.D. (1975) J. Cell Physiol. 22,
301-310.

Canagaratna, M.G.P. and Riley, P.A. (1975) J. Cell Physiol. 22,
271-282.

Canagaratna, M.C.P., Chapman, R., Ehrlich, E., Sutton, P.M., and
Riley, P.A. (1977) Differentiation 2, 157-160.

Lipkin, G. and Knecht, M.E. (1974) Proc. Natl. Acad. Sci. U.S.A.
22, 849-853.

Knecht, M.E. and Lipkin, G. (1977) EXp. Cell Res. 222, 15-22.

116.

117.

118.
119.

120.

121.

122.

123.

124.

125.
126.
127.

128.

129.

130.

131.

132.

133.
134.

146

Lipkin, G., Knecht, M.E., and Rosenberg, M. (1978) Cancer Res. 22,
635-643.

Garcia-Giralt, E., Berumen, L., and Macieira-Coelho (1970) J. Nat.
Cancer. Inst. 22, 649-655.

Njeuma, D.L. (1971) Exp. Cell Res. 22, 244-252.

Houck, J.C., Weil, R.L. and Sharma, V.K. (1972) Nature New Biol.
222, 210-211.

Houck, J.C., Kanagalingain, K., Hunt, C., Attallah, A. and Chung,
A. (1977) Science 222, 896-897.

Strobel-Stevens, J.D. and Lacy, J.C. (1981) J. Cell Physiol. 222,
201-207.

Mizel, 5.8., DeLarco, J.E., Todaro, G.J., Fauai, W.L. and
Hilfiker, M.L. (1980) Proc. Natl. Acad. Sci. U.S.A. 22,
2205-2208.

Holley, R.W., Armour, R. and Baldwin, J.H. (1978) Proc. Natl.
Acad. SCI. U.S.A. 22, 1864'18660

Holley, R.W., Bohlen, P., Fava, R., Baldwin, J.H., Kleeman, G. and
Armour, R. (1980) Proc. Natl. Acad. Sci. U.S.A. 22, 5989-5992.

Kawai, S. and Hanafusa, H. (1971) Virology 22, 470-479.
Diamond, L. (1967) Int. J. Cancer. 2, 143-152.

Amano, T., Hamprecht, B., and Kemper, W. (1974) Exp. Cell Res. 2,
399-408.

Rogers, A.W. (1973) Techniques of Autoradiography, Elsevier/North
Holland, Biomedical Press, Amsterdam.

Lowry, 0.H., Rosebrough, N.J., Farr, A.L., and Randall, R.J.
(1951) J. Biol. Chem. 222, 265-275.

McPhie, P. (1971) Methods in Enzymology, Vol. XXII, Academic
Press, New York, p. 25.

Todaro, G.J., Green, H., and Goldberg, B.D. (1964) Proc. Natl.
Acad. Sci. U.S.A. 22, 66-73.

Hunter, W.M. and Greenwood, E.C. (1962) Nature (Lond.) 222,
495-496.

Lammeli, U.K. (1970) Nature (Lond.) 222, 680-685.
Bonner, W.M. and Laskey, R.A. (1974) Eur. J. Biochem. 22, 83-88.

135.

136.

137.
138.
139.
140.
141.

142.

143.

144.
145.

146.

147.

148.

149.

150.

147

Lazarides, E. and Lindberg, U. (1974) Proc. Natl. Acad. Sci.
U0 50A. fl, 4742‘47460

Paul, D., Brown, K.D., Rupriak, H.T. and Ristow (1978) In Vitro
(Rockville) 22, 76-84.

Harel, L. and Jullien, M. (1976) J. Cell Physiol. 22, 253-254.
Lowenstein, W.R. (1979) Biochem. Bi0phys. Acta 222, 1-65.
Schaffner, W. and Weissman, C. (1973) Anal. Biochem. 22, 502-514.
Carpenter, G. and Cohen, S. (1976) J. Cell Biol. 22, 159-171.

Aharonov, A., Press, R.M. and Herschman (1978) J. Biol. Chem. 222,
3970-3977.

Brown, K.D., Yeh, Y. and Holley, R.W. (1979) J. Cell Physiol. 222,
227-238.

Gunthur, G.R., Wang, J.L., Yahara, I., Cunningham, B.A. and
Edelman, G.M. (1973) Proc. Natl. Acad. Sci. 22, 1012-1016.

Heldin, C.-H., Westermark, B. and Wasteson (1979) Proc. Natl.
Acad. Sci. U.S.A. 22, 3722-3726.

Clemmons, D.R., Van Wyk, J.J. and Pledger, W.J. (1980) Proc. Natl.
Acad. U.Svo E, 6644-66480

Raizda, M.K. and Purdue, J.F. (1976) J. Biol. Chem. 222,
6445-6455.

Whittenberger, B., Raben, D. and Glaser, L. (1979) J. Supramol.
Struct. 22, 307-327.

Pruss, R.M., and Herschman, H.R. (1977) Proc. Natl. Acad. U.S.A.
22, 3918-3921.

Peterson, S.W., Vale, R., Das, M. and Fox, C.F. (1978) J.
Supramol. Struct. (Suppl. 2), 126.