Heparin and Hormonal Regulation of mRNA Synthesis ... - NCBI - NIH

2 downloads 59 Views 2MB Size Report
of insulin, glucagon, and growth hormone in combination with heparin; TGF(i was regulated ...... growth factor II mRNA expression in human breast cancer.
Vol. 11, No. 1

MOLECULAR AND CELLULAR BIOLOGY, Jan. 1991, p. 108-116 0270-7306/91/010108-09$02.00/0 Copyright © 1991, American Society for Microbiology

Heparin and Hormonal Regulation of mRNA Synthesis and Abundance of Autocrine Growth Factors: Relevance to Clonal Growth of Tumors ISABEL ZVIBEL, ELAINE HALAY, AND LOLA M. REID*

Departments of Molecular Pharmacology and Microbiology and Immunology, Albert Einstein College of Medicine, Bronx, New York 10461 Received 8 November 1989/Accepted 12 October 1990

Highly sulfated, heparinlike species of heparan sulfate proteoglycans, with heparinlike glycosaminoglycan chains, are extracellular matrix components that are plasma membrane bound in growth-arrested liver cells. Heparins were found to inhibit the growth and lower the clonal growth efficiency of HepG2, a minimally deviant, human hepatoma cell line. Heparan sulfates, closely related glycosaminoglycans present in the extracellular matrix around growing liver cells, had no effect on the growth rate or clonal growth efficiency of HepG2 cells. Neither heparins nor heparan sulfates had any effect on the growth rate or clonal growth efficiency of two poorly differentiated, highly metastatic hepatoma cell lines, SK-Hep-l and PLC/PRF/5. Heparin's inhibition of growth of HepG2 cells correlated with changes in the mRNA synthesis and abundance of insulinlike growth factor II (IGF II) and transforming growth factor beta (TGF,I). HepG2 cells expressed high basal levels of mRNAs encoding IGF II and TGFP that were inducible, through transcriptional and posttranscriptional mechanisms, to higher levels by specific heparin-hormone combinations. For both IGF II and TGF,, the regulation was multifactorial. Transcriptionally, IGF II was regulated by the additive effects of insulin, glucagon, and growth hormone in combination with heparin; TGF(i was regulated primarily by the synergistic effects of insulin and growth hormone in combination with heparin. Posttranscriptionally, the mRNA abundance of the IGF II 4.5- and 3.7-kb transcripts was affected by insulin. Heparin induction of all IGF II transcripts was also dependent on triiodotyronine and prolactin, but it is unknown whether their induction by heparin was via transcriptional or posttranscriptional mechanisms. Heparin-insulin combinations regulated TGFPi posttranscriptionally. The poorly differentiated hepatoma cell lines PLC/PRF/5 and SK-Hep-l either did not express or constitutively expressed low basal levels of IGF I, IGF II, and TGFIA, whose mRNA synthesis and abundance showed no response to any heparin-hormone combination. We discuss the data as evidence that matrix chemistry is a variable determining the expression of autocrine growth factor genes and the biological responses to them. HepG2 cells, which are highly differentiated and nonmetastatic (29), were obtained from Barbara Knowles (Wistar Institute, Philadelphia, Pa.). Culture conditions. (i) Substrates. Cells were plated directly onto 100- or 150-mm tissue culture plastic dishes (Falcon). (ii) Media. The hepatomas were cultured in RPMI 1640 (GIBCO, Grand Island, N.Y.) supplemented with penicillin (100 ,ug/ml) and streptomycin (100 ,ug/ml). This medium was further supplemented with 10% fetal bovine serum (GIBCO Hyclone) to produce serum-supplemented medium (SSM) or with a defined mixture of trace elements, hormones, and growth factors (see below) to produce a serum-free, hormonally defined medium (HDM). The HDM used for the hepatomas was that designed for growth of hepatoma cells on tissue culture plastic and described in detail elsewhere (11, 19). It contains insulin (100 ,ug/ml; Sigma), transferrin (10 ,ug/ml; Sigma), glucagon (10 ,ug/ml; Sigma), hydrocortisone (108 M; Sigma), triiodotyronine (10-' M; Sigma), prolactin (2 mU/ml; Sigma), growth hormone (GH) (10 ,uU/ml; Sigma), linoleic acid bound to bovine serum albumin (10 ,ug/ml; Pentax), zinc (10-10 M), selenium (3 x 10-1o M), and copper (10-' M). The trace elements were obtained from Johnson Matthey Chemicals (London, England). Sources for heparins and heparan sulfates. Reference standards for bovine lung-derived heparins and for bovine lungderived heparan sulfates were provided by Larry Rosenberg,

Previous studies from our laboratory (10) have indicated that the extracellular matrix contains a set of factors affecting the clonal growth efficiency of tumor cells, a characteristic important for the ability of tumor cells to metastasize and colonize specific tissues. The matrix components responsible for this phenomenon include species of highly sulfated, heparinlike heparan sulfate proteoglycans or their glycosaminoglycan chains, heparinlike heparan sulfates (10). We are now trying to elucidate the mechanism by which heparins and heparinlike heparan sulfates or their proteoglycan forms can differentially affect the clonal growth efficiency of metastatic versus nonmetastatic carcinomas. Our working hypothesis, tested in these studies, has been that these heparins or heparinlike molecules regulate the synthesis of autocrine growth factors thought to be involved in low-density growth.

passage

MATERIALS AND METHODS Human hepatoma cell lines. PLC/PRF/5, established by Alexander et al. (1), expresses some liver-specific functions (29) and was obtained from I. Millman (Fox Chase Cancer Center, Philadelphia, Pa.). SK-Hep-1, a relatively anaplastic cell line (15), was a gift from Jorgen Fogh (Sloan Kettering Institute, Walker Laboratories, New York, N.Y.). Early*

Corresponding author. 108

VOL . 1 l, 1991

HEPARIN AND HORMONAL REGULATION OF mRNA SYNTHESIS

who obtained them from Martin Matthews and J. A. Ci-

fonelli, University of Chicago. Commercially available bo-

vine lung-derived heparins were obtained from Sigma. The biological activity of reference standards was quite reproducible, but that of commercially available heparins varied widely from batch to batch. Nevertheless, the scarcity of glycosaminoglycan reference standards forces investigators to use the commercial heparins of weaker and variable specific activity. These findings are analogous to those reported by Castellot et al. (5) in their structure-function analysis of heparin effects on inhibition of growth of smooth muscle cells. The data shown are from Sigma's bovine lung-derived heparin, lot 53F-0532. The cultures were treated with 20 to 50 p.g of this heparin per ml, a concentration previously shown to regulate tissue-specific gene expression in normal and neoplastic liver cells (18, 47). Clonal growth assays. Cells were plated in triplicate for each condition at 100, 103, 104, and 105 on 60-mm tissue culture plates and in SSM. After 18 h, the plates were gently rinsed and refed. The test media consisted of SSM with and without either bovine lung-derived heparin (50 ,ug/ml) or bovine lung-derived heparan sulfate (50 p.g/ml). The cultures were incubated for 2 to 3 weeks with weekly medium changes and then stained with 1% acridine orange. The number of colonies per plate was counted. Clonal growth efficiency was calculated as (number of colonies/number of cells seeded) x 100. The experiment was replicated five times. Growth curves. Cells were plated in triplicate for each condition at 105 cells per 60-mm dish in SSM. After 24 h, the plates were rinsed with phosphate-buffered saline (PBS) and fed SSM with or without bovine lung-derived heparin (50 jig/ml) or bovine lung-derived heparan sulfate (50 ,ug/ml). Medium changes were done twice weekly. Triplicate plates per condition were used for cell counts on days 1, 3, 7, 10, and 14. The average cell numbers per day were plotted on semilog paper. The doubling time was determined from the slope of the curve during log-phase growth of the cells. The growth curves were repeated three times. Molecular hybridization assays. Nuclear transcript run-on assays and Northern (RNA) blots (see below) were used to determine the synthesis and abundance of mRNAs encoding various autocrine growth factors. The hepatomas were plated onto 100- or 150-mm culture dishes under the conditions specified, and the medium was changed 6 h after plating and then again after 48 h. The cells were assayed after % h in culture. In each experiment, cells were pooled from two to three dishes per culture condition. The experiments were run at least three times. The data from all autoradiograms of either run-on assays or Northern blots were scanned with an Quantimet densitometer (model 920; Manufacturer's Cambridge Instrument). The data for each of the autocrine growth factor genes were normalized to that of a common gene used as an internal control. Dihydrofolate reductase (DHFR) was the internal control for the nuclear transcript run-on assays; 18S RNA was the internal control for the Northern blots. The expression of these two genes was found not to alter under the experimental conditions used. Northern blots. The cells were washed twice with 10 ml of cold PBS, removed from culture dishes with a rubber policeman, and pelleted; cytoplasmic RNA was isolated by the isotonic buffer-Nonidet P-40 lysis method, followed by phenol-chloroform and chloroform extractions (31). RNA samples were resolved by electrophoresis through 1% agarose, submerged-slab, denaturing formaldehyde gels in MOPS buffer (31). RNA was transferred to GeneScreen

109

(New England Nuclear), and the RNA-containing filters were prehybridized and then hybridized with the appropriate probes. The cDNA clones complementary to specific mRNAs (listed below) were radioactively labeled by primer extension as described by Feinberg and Vogelstein (14). [32P]dCTP (specific activity, 3,200 Ci/mmol) was included to obtain a specific activity of 8 x 108 to 12 x 108 cpm/,Lg of DNA. DNAs used in hybridization included those complementary to mRNAs encoding insulinlike growth factor I (IGF I) and insulinlike growth factor II (IGF II), transforming growth factor alpha (TGFa), transforming growth factor beta (TGF,B), and 18S RNA. The cDNA probe for human IGF I was plasmid phigf 1, containing a human liver-derived cDNA with a PstI insert of 662 bp of prepro-IGF I (2). The cDNA encoding human prepro-IGF II was plasmid phigf 2, containing an internal PstI insert of 1,090 bp (2, 3). The plasmid for mouse TGFa contained an EcoRI insert of 925 bp. These three plasmids were the kind gifts of G. I. Bell (Howard Hughes Institute, University of Chicago). The murine TGFPi was an EcoRI cDNA insert of 1,600 bp (a kind gift of R. Derynck, Genentech, South San Francisco, Calif.). The cDNA encoding DHFR derived from hamster cells contained an insert of 1,900 bp and was a gift of P. Melera (University of Maryland, Baltimore). Nuclear transcript run-on assays. The method is a modification of that described by Clayton and Darnell (7). After 96 h in culture, one or two 150-mm plates were scraped with a rubber policeman into PBS; the cells were washed twice with PBS and centrifuged at 800 x g for 5 min. Then 5 ml of isotonic buffer (140 mM NaCl, 10 mM Tris [pH 7.4], 1 mM MgCl2) and 0.5% (final concentration) Nonidet P-40 were added for 5 min on ice. The nuclei were pelleted by centrifugation at 1,000 x g for 5 min, and the supernatant was used for cytoplasmic RNA extraction. The nuclei were washed with 5 ml of 10 mM Tris (pH 8.1)-20% glycerol-140 mM KCI-5 mM MgCl2-1 mM MnCl2-14 mM P-mercaptoethanol, resuspended in 1 ml of the same buffer, and frozen in liquid nitrogen. On the day of the experiment, the nuclei were thawed at 37°C, centrifuged at 1,000 x g for 5 min, and suspended in 200 IlI of complete reaction buffer (10 mM Tris [pH 8.0], 140 mM KCI, 5 mM MgCl2, 14 mM dithiothreitol, 1 mM each ATP, GTP, and CTP, 20% glycerol, 100 ,ug of creatine phosphokinase per ml, 10mM creatine phosphate, and 1 mCi of [32P]UTP [specific activity, 3,000 Ci/mmol; Amersham] per ml). The pellets were incubated in the complete reaction buffer at 30°C for 15 min with gentle shaking. Then 1.5 ml of 500 mM NaCl-10 mM Tris (pH 7.4-S50 mM MgCl2 was added along with DNase so that the final concentration was 100 U/ml. Then EDTA was added to a final concentration of 10 mM (pH 8.0), sodium dodecyl sulfate (SDS) was added to a final concentration of 0.5%, and proteinase K was added to a final concentration of 0.4 mg/ml. The samples were incubated for 30 min at 37°C, after which 4.5 ml of ETS buffer (10 mM Tris [pH 7.5], 0.5% SDS, 10 mM EDTA) was added. The RNA was purified from the nuclei by standard hot phenol-chloroform extractions, followed by ethanol precipitation. To eliminate unincorporated label, the samples were trichloroacetic acid precipitated, followed by at least two rounds of ethanol precipitation. The same number of counts per minute from each RNA sample was added to hybridization bags containing nitrocellulose filters to which 5-,ug amounts of different cDNA plasmids were bound. The filters had been prehybridized overnight and were then hybridized

110

MOL. CELL. BIOL.

ZVIBEL ET AL. TABLE 1. Influence of heparin and heparan sulfate on growth of hepatoma cellsa

Cell line

HepG2b

Plating density

102

103 104 105 SK-Hep-1

102

10-3 105 PLC/PRF/5

102

103 105

Control

2.8 ± 1.7 (7 3) 14 5 Too many Too many 28 ± 5 Too many Too many 7.3 ± 2.7 46 6.8 Too many

to count to count to count to count to count

Doubling time (h)

colonies No. ofHP

HS

0 (5 2) 0 64 ± 11 Too many to count 21 ± 3.8 Too many to count Too many to count 15 ± 3.9 53 7.2 Too many to count

5.3 ± 2.3 (10 ± 3) 12 ± 3 Too many to count Too many to count 22 ± 5 Too many to count Too many to count 15.3 + 4 55 ± 7.5 Too many to count

Control

HP

HS

29±2

58±6

31±4

34±4

31±4

28±3

28+3

26±2

24±4

a Data are from three to five experiments for the clonal growth efficiency assays and three experiments for growth rates (doubling times). HP, Bovine lung-derived heparin (NIH reference standard obtained from M. Matthews and J. A. Cifonelli), used at 50 ,ug/ml. HS, Bovine lung-derived heparan sulfate (reference standard obtained from M. Matthews and J. A. Cifonelli), used at 50 to 100 Rg/ml. b Data are from experiments with the original heparin-sensitive subline of HepG2. Data in parentheses (102 seeding density) are for the same subline after responsiveness to heparins had waned.

for 48 h at 37°C in a buffer containing 5 x SSC (SSC is 0.15 M NaCl plus 0.015 M sodium citrate), 0.1% SDS, lx Denhardt solution, 50% formamide, 50 mM sodium phosphate, and 500 ,ug of yeast tRNA per ml. The nitrocellulose filters were washed four times at 60°C for 20 min each time in 1 x SSC-0.1% SDS and then twice at 60°C for 45 min each time in 0.2x SSC-0.1% SDS. They were incubated 30 min at 37°C with 10 gig of RNase A per ml in 2x SSC and then washed for 1 h at 37°C with 2x SSC. The filters were exposed to Kodak X-Omat film with two intensifying screens at -700C. RESULTS Heparin effects on clonal growth efficiency of metastatic and nonmetastatic hepatoma cells. Heparins, derived from bovine lung (Table 1) or porcine intestine (data not shown), were found to inhibit the growth of HepG2 cells plated at cell densities above 104/60-mm dish and to eliminate HepG2 survival at all cell densities at or below 103/60-mm plate. Heparins had no effect on the growth rate or clonal growth efficiency of PLC/PRF/5, SK-Hep-1, or the American Type Culture Collection-derived subline of HepG2. Bovine lungderived heparan sulfates had no effect on growth at any density of any of the cell lines. Autocrine growth factors produced by metastatic and nonmetastatic hepatoma cell lines. To test our hypothesis that heparins affect production of autocrine growth factors, we screened for various autocrine growth factors in three human hepatoma cell lines: HepG2, a minimally deviant cell line, and two poorly differentiated hepatoma cell lines, PLC/PRF/5 and SK-Hep-1. Northern blots of the cytoplasmic RNA from the cell lines cultured for 96 h in SSM were hybridized with cDNA probes for a battery of autocrine growth factors (Fig. 1). HepG2 expressed significant basal levels of IGF II and TGF, and lower levels of IGF I. The IGF I probe bound to a transcript of 5.3 kb in all three cell lines, an mRNA size identical to that reported for human fetal liver (20). In the adult human liver, the mRNA sizes observed for IGF I are 7.7, 5.3, and 0.9 kb. Similarly, the IGF II mRNA species present in HepG2 cells were the fetal ones: 6.0-, 4.5-, 3.7-, and 2.2-kb species. However, the 2.2-kb transcript was not expressed consistently. In the human adult liver, the size of IGF II mRNA has been shown to be 5.3 kb (2). Only TGF, was expressed at significant

levels in PLC/PRF/5 and SK-Hep-1 cells; only faint bands for IGF II were observed. Heparin modulation of autocrine growth factor expression in the human hepatoma cell lines. Heparin treatment of the cells altered the expression of the autocrine growth factors in HepG2 but not in PLC/PRF/5 and SK-Hep-1 cells. Representative findings from one of three experiments are shown in Fig. 2. HepG2, PLC/PRF/5, and SK-Hep-1 cells were cultured in SSM or in HDM in the presence or absence of bovine lung heparin (20 jig/ml). Steady-state mRNA levels of IGF II and TGF, were increased in HepG2 cells cultured with heparin (Fig. 2). By contrast, heparin had no effect on any of the autocrine growth factors' mRNA synthesis or abundance in either SK-Hep-1 or PLC/PRF/5 cells (Fig. 2). Thus, only the HepG2 cell line proved regulatable by heparin.

After 1.5 years of studies on the HepG2 subline in the laboratory (passage number unknown), the heparin regulatability waned. All sublines of HepG2 tested from the American Type Culture Collection and from a number of laboratories also proved insensitive or relatively insensitive to heparins. However, an early-passage (passage 82) HepG2 U- Ln LL-

c\J L

C-D

0:L.

-,

_) a-

28S- '-

cr a.

0

cL.j IX I CL c 0

Lr) ILL

c C'%JcL C:D D

Co U 0

W

-

0.

I )

r

CL c)

W'1-"I

18S-

GFI

IGFIE

TGFa

TGF

FIG. 1. Northern blots with cytoplasmic RNA (20 ,ug per lane) extracted from hepatoma cells grown for 4 days in SSM. Cytoplasmic RNA was extracted and run in 1% agarose-formaldehyde gels, blotted to GeneScreen, and hybridized with [32P]dCTP primer extension-labeled cDNA coding for the specified autocrine growth factor. Shown are representative blots from one of five experiments.

HEPARIN AND HORMONAL REGULATION OF mRNA SYNTHESIS

VOL . 1 l, 1991 A

PLC/PRF/5

HepG2

SK/Hep-1

A

II

II

I

^ -a

0

CD

0

kb 6.0-

-C

-a

-a

LI

VI

m~ c1r) ::

m I I .I .I ,ii E'E'c

4.5 -

ill

I

I

0

I

I

0

3.7 -

6.0 4.5 3.7 Heparin:

-

(209g/rnI)

+

SSM

-

+

-

-

+

HDM

SSm

HOM

+

+

SSM

+

HDM

B

B

"*""ww"_wwwu 18S Bss

CL

C

L

U ()

0.

I

N.(L

__

0 4

0.

0)

-

l_ CD

c1c.

I

=

+

+

=: kb 2.5 -

--

_ _-

_0 _

6

HepG2 SKHepI FIG. 2. Northern blots of cytoplasmic RNA from human hepatoma cell lines cultured for 4 days in SSM or HDM in the presence or absence of bovine lung heparin (hep) (20 ,ug/ml). Cytoplasmic RNA was extracted after 4 days, and 20 ,ug of cytoplasmic RNA per lane was run in a 1% agarose-formaldehyde gel and hybridized with primer extension-labeled cDNA probes. Shown are representative blots from one of three experiments. (A) IGF II probe; (B) the same blot after stripping and rehybridizing with an 18S RNA probe; (C), TGF,3 probe.

subline obtained from Barbara Knowles (Wistar Institute) showed heparin sensitivity even greater than that of the subline with which we had originally been working. Except for the medium controls (see Fig. SC and 6), the data shown are either from our original subline or from this earlypassage subline. Effects of hormones and heparin on steady-state mRNA levels for IGF II and TGFji in HepG2 cells. Other studies in our laboratory (18, 37, 47) have shown that heparin can affect mRNA synthesis and abundance of liver-specific genes by acting in concert with hormones, each gene being regulated by two to three specific hormones. We investigated whether autocrine growth factor genes are similarly affected by heparin. Replicate plates of HepG2 cells were grown for 4 days in HDM or in HDM in which individual hormones were omitted and in the presence or absence of bovine lung-derived heparin (20 to 50 ,ug/ml). In parallel, we tested the effects of each of the hormones added alone and with or without heparin in serum-free RPMI 1640. The RNAs from these cultures were hybridized with radiolabeled probes for IGF II and TGFP. We detected three IGF II mRNA transcripts of 6.0, 4.5, and 3.7 kb and only sometimes a 2.2-kb species. The 6.0-, 4.5-, and 3.7-kb transcripts were

FIG. 3. Northern blots of cytoplasmic RNA from the human hepatoma cell line HepG2 cultured in HDM from which different hormones were omitted. Bovine lung heparin (20 ,ug/ml) was added to half of the cultures incubated with each of these defined media. The Northern blots were prepared by using 5 ,ug of cytoplasmic RNA per lane and hybridized with 32P-labeled cDNAs. The blots are representative of those from three experiments. (A) IGF II probe; (B) the same blot stripped and rehybridized with an 18S RNA probe; (C) TGFI probe. hep, Bovine lung heparin; ins, insulin; gluc, glucagon; T3, triiodotyronine; pro, prolactin.

each regulated by distinct hormone-heparin combinations (Fig. 3A and 4A). The addition of heparin to complete HDM resulted in an increase in the abundance of all three mRNA transcripts of IGF II. The omission of insulin from HDM eliminated the heparin induction of the 6.0-kb transcript. The lack of insulin also resulted in an increase in the basal expression of the 4.5-kb and especially of the 3.7-kb transcripts; the addition of heparin did not induce the mRNA abundance further. The omission of glucagon from HDM had no significant effect on either the basal or heparininducible levels of the three mRNA species encoding IGF II. Deletion of triiodotyronine, GH, or prolactin from HDM did not affect the basal levels of expression but eliminated the heparin-induced mRNA accumulation of all three mRNA transcripts of IGF II. TGF, mRNA levels (Fig. 3C and 4B) were also increased by heparin supplementation of HDM. In cultures in HDM from which GH was omitted, the TGF, basal level was elevated, and heparin addition did not induce it further. In cultures in HDM from which insulin was omitted, the basal level of TGF, was reduced and heparin induction was

MOL. CELL. BIOL.

ZVIBEL ET AL.

112

IGF I

A

0.4 0.3 0.2

CL)

C (L

3.7kb Transcript

H

~~~~~~~~~~~~~~~~~~~~C_

.

L

DHFR IGF II

H H [triri

I

t

A

.. ..

0.1

TGFP

-

S.

_

-

B to 0.3 =

H

H

I

4.5kb Transcript

)

CDL

0.2 -c

.- 0.1

(/) c

a

a:

H

0.4 0.3 0.2 0.1

+

ep

H kb 6.0TrHnscript

DHFR IG FII TGF,

H~~H H~~~~~O Conttrol arin

|

ig/mi) I;~~~~~~~~~~

_s

Hepc c

HDM -Insulin -Glucogon B

03

H

0.1

0.2

-

HDM

-Insulin

-Glucagon -Growth Hormone

FIG. 4. Histograms showing the relative mRNA abundances of the TGF3 and the three IGF II transcripts in HepG2 cells cultured in HDM with or without specific hormone deletions and with or without heparin. The histograms derive from densitometry readings of autoradiograms from one of three replicate experiments. The data were normalized to 18S RNA expression under the same conditions.

eliminated. HepG2 cells cultured in the presence of insulin alone in a serum-free medium had high mRNA abundance for TGF,, and addition of heparin increased TGFO and mRNA abundance even more. Effects of heparin and hormones on the transcription rates of IGF II and TGFI mRNAs in HepG2 cells. Both TGF,3 and IGF II were regulated by the additive or synergistic effects of multiple hormones in combination with heparin. By contrast, RPMI 1640 alone or RPMI 1640 with heparin showed no significant effect on transcription rates of IGF II or TGFP (Fig. 5C). Heparin treatment of cells cultured in complete HDM resulted in a three- to fourfold increase in the mRNA synthesis rate. Heparin-insulin combinations produced a small increase in the transcription rates of IGF II mRNA (Fig. 5A and B; Fig. 6A). Additive effects were observed in cells treated with heparin and with insulin combined with

+

+

en

U)

c.

c

c

C

c

+,

+

-

-

c

+:

4

+

_r

-

-

_w

_

_

RPM1i

R PM i. heF9

_

_-

ifj.F II

TGF

(SOg/ml)

U~~~~~~~~~~~

4

u)

oGEM

-T3 -Growth -Prolactin Hormone 0 Control X Heparin

TGF-0 0.4

4-

U)

:.

FIG. 5. Nuclear transcription run-on assays of HepG2 cells cultured for 4 days. The blots are representative of those from one of three experiments. (A) Expression in cells in HDM and in derivative HDM from which specific hormones were deleted. Half of the cultures maintained in each medium were also treated with bovine lung heparin (50 ,ug/ml). Note the expression of DHFR used as an internal control. (B) Expression in cells in serum-free RPMI 1640 supplemented with one or two hormones at a time in the presence or absence of bovine lung heparin (50 ,ug/ml). (C) Expression in HepG2 cells (the American Type Culture Collection subline) cultured in serum-free RPMI 1640 (no hormones or growth factors) alone or supplemented with bovine lung heparin (50 ,ug/ml). These control experiments could not be repeated with the early-passage (passage 82) HepG2 subline obtained from Barbara Knowles since that subline did not survive in RPMI 1640 without hormones. a-tub, a-Tubulin. Other abbreviations are as for Fig. 3.

glucagon and especially with GH (Fig. 5A and B; Fig. 6A). In experiments in which hormones were deleted one by one from HDM (Fig. 5A), the omission of either insulin, glucagon, or GH caused a decrease in the inductive effect of heparin on the transcription rate of IGF II compared with that in cells in complete HDM (Fig. 5A). Multiple hormones also affected the transcription rate of TGFP. The transcription rate of TGFI mRNA (Fig. 5A and B; Fig. 6B) was induced weakly (twofold) by heparin in combination with insulin plus glucagon. However, the peak transcription rates (13-fold induction) were observed in heparin-treated cells in complete HDM or in a serum-free medium supplemented only with insulin and GH (Fig. 6B). Deletion of insulin from the complete HDM resulted in a loss of heparin induction of the transcription rate. Deletion of glucagon resulted in a small reduction in the transcription rate.

HEPARIN AND HORMONAL REGULATION OF mRNA SYNTHESIS

VOL . 1 l, 1991

IGF 11 Relative Transcription Rate

7 1A

EI Control

_ Heparin (lung)

5

4-

3

*Ins

+Ins

+Gluc

+Ins

+GH

HDM HDM-InsHDM-GlucHDM-GH

RPMI

TGF B

*lns

*lns

*lns

*Gluc

+GH

HDM

HDM-lnsHDM-GlucHDM-GH

RPMI

FIG. 6. Histograms showing the effects of heparin-hormone combinations on the transcription rates of mRNAs for the genes indicated on top. The data were normalized to DHFR expression under the same conditions. The run-on assays were performed three times with similar results. Representative experiments

were

chosen

and used for calculating the histograms. The data for RPMI 1640 with or without heparin were from an independent experiment using the original subline of HepG2; these controls could not be repeated with the early-passage (passage 82) subline since that subline did not survive in RPMI 1640 without hormones. The calculations were made as follows: (optical density for each autocrine growth factor gene optical density for pGEM)/(optical density for DHFR under the same condition optical density for pGEM). (A) IGF II; (B) TGF,B. Ins, Insulin; Gluc, glucagon. -

DISCUSSION

Highly sulfated heparan sulfate proteoglycans, with heparinlike glycosaminoglycan chains, are plasma membraneassociated extracellular matrix components of growth-arrested liver cells (13, 26-28). Heparins were found to inhibit growth and to lower the clonal growth efficiency of a minimally deviant hepatoma cell line, HepG2. Heparan sulfates, similar to the poorly sulfated glycosaminoglycan chains on heparan sulfate proteoglycans produced by growing liver cells (13, 41), had no effect on HepG2 growth or clonal growth efficiency. Neither heparins nor heparan sulfates had any effect on the growth or clonal growth efficiency of two highly metastatic hepatoma cell lines, PLC/PRF/5 and SK-Hep-1. The insensitivity of the poorly differentiated hepatomas to heparin regulation probably results from their degradation by tumor cell-derived glycosidases (33). Heparins have been shown to promote the growth of some cell types such as endothelia (16, 52) and to inhibit the

113

proliferation of others such as smooth muscle cells and epithelia (5, 6, 26, 38, 39, 54). The mitogenic effect on endothelia was due to fibroblast growth factor, which remained fully active after binding with high affinity to a pentasaccharide sequence in the heparin species made by endothelia (52). The heparin inhibition of growth of smooth muscle cells has been found to be due to a saccharide sequence in certain heparins that binds to and results in down-regulation of the epidermal growth factor receptor without effects on the insulin receptor or platelet-derived growth factor (5, 38, 39, 54). In recent studies, Conrad and associates have suggested that heparin, by itself, can also regulate cell growth through unique changes in its chemical structure that occur in a density-dependent fashion (13) and through its translocation to the nucleus of the cell (26). Heparin effects on the expression of autocrine growth factors in metastatic versus nonmetastatic carcinoma cell lines. Since autocrine growth factors are thought to be critical for low-density growth of tumors (46), we tested the hypothesis that heparins inhibit growth of minimally deviant tumors by regulating the synthesis of autocrine growth factors in the cells. As predicted by our hypothesis, autocrine growth factor synthesis was regulatable by heparins in the HepG2 cell line but not in two metastatic cell lines, SK-Hep-1 and PLC/PRF/5. Contradicting the hypothesis were the data showing that the highest basal levels of the autocrine growth factors were evident in the highly differentiated, nonmetastatic HepG2 cells, and heparins induced the autocrine growth factors in HepG2 cells to even higher levels. Thus, heparins, which we previously (10) showed caused cell differentiation and inhibition of growth in normal hepatocytes and in minimally deviant tumor cells, also result in greatly elevated levels of specific autocrine growth factors. Heparin effects on autocrine growth factors are through potentiation of hormonal regulation. In past reports, there has been a paucity of data to identify circulating hormones that might regulate IGF II or TGFPi (36, 53). In adult liver, IGF I but not IGF II is under transcriptional control by GH (32). In HepG2, heparin regulation of synthesis of autocrine growth factors was via potentiation of hormone effects. TGF, and each of the transcripts of IGF II required the presence of specific hormones for heparin induction of mRNA synthesis or abundance. Moreover, the heparin effects were neutral, stimulatory, or inhibitory on expression of TGFi or of specific transcripts of IGF II, depending on which heparin-hormone combination was used. Heparin in the presence of GH and insulin was stimulatory for TGF,B expression and involved predominantly a transcriptional control mechanism. By contrast, the heparin effect on IGF II mRNA levels involved both transcriptional and posttranscriptional mechanisms. Therefore, the specificity of the influence of heparin is dictated by the specific hormones. The differential regulation of the abundance of the three IGF II transcripts by distinct groups of hormones and growth factors (Fig. 3 and 4) could offer an explanation for why the different transcripts are found in specific tissues and in tissues of different developmental stages. The human IGF II gene is transcribed from three promoters, which are both developmentally regulated (8) and tissue specific (25, 44). The factors affecting IGF II mRNA synthesis from these three promoters are not known. Our data are the first to suggest possible regulatory signals. We found that insulin, glucagon, and GH were required for heparin induction of the transcription rate of IGF II. The promoters identified for IGF II are P1, found active only in adult liver and giving rise to a 5.3-kb mRNA; P2, which yields a 6.0-kb as well as a

114

ZVIBEL ET AL.

2.2-kb transcript in fetal tissues; and P3, which yields a 4.8-kb mRNA species that is also present in many fetal tissues. Our studies show that the major transcripts of IGF II in HepG2 cells are 6.0, 4.5, and 3.7 kb. A 2.2-kb transcript was observed in some experiments. Thus, insulin, GH, and glucagon likely affect the P2 and P3 promoters; it is unclear whether the P1 promoter is affected. In addition, elimination of insulin resulted in a relative increase in the mRNA abundance of the 4.5-kb transcript and especially of the 3.7-kb transcript through posttranscriptional mechanisms, suggesting that there may be a negative response element in the coding transcript sensitive to insulin and resulting in lowered mRNA stability. Although seemingly contradictory findings, perhaps the data implicate a feedback loop that maintains a stable level of IGF II. Loss of insulin results in increased stability; presence of insulin results in lowered stability but increased synthesis. Triiodotyronine and prolactin were also found necessary for heparin induction of all three transcripts. However, we have not tested whether their effects are via transcriptional or posttranscriptional mechanisms. Although heparin showed a direct effect on the TGF, mRNA stability, its primary effects were on TGFP mRNA synthesis, again via specific hormone-heparin combinations. The most potent signal for TGF,B transcription was a synergistic effect of insulin, GH, and heparin. Role of IGF H in neoplasia and differentiation. Although the presence of both IGF I and IGF II in tumors and embryonal tissues is well documented (4, 34, 49, 55), the actual data are unclear as to whether IGF II is critically involved in neoplastic transformation, in tumor progression, or in differentiation (4, 9, 20, 30, 32, 34, 42, 45, 48). IGF II mRNA levels are very high in fetal tissues such as embryonic liver (4) but are low in human adult liver (20) and undetectable in rat liver after birth (4). During rat hepatocarcinogenesis, IGF II gene transcription was reactivated from three different promoters, whose activities differed in efficiency in each of the analyzed tumors (51). In normal rat tissues, however, the three promoters were coordinately regulated (50). An alternate explanation is that the IGF II levels in tumors result from expansion of liver progenitor cells expressing IGF II (12, 17, 37) rather than induction due to oncogenic transformation (12). TGFj expression in normal and neoplastic tissues. TGFP is a strong inhibitor of the proliferation of epithelial cells, including hepatocytes (22), but it can have different effects, depending on what other factors are present (40). TGF, is known to antagonize the mitogenic effects of stimulating growth factors, such as the effects of fibroblast growth factor on vascular endothelial cells (21) or the effects of epidermal growth factor on hepatocytes (22) or on myc-transfected fibroblasts (40). One of the TGFi effects is the transcriptional activation of genes for extracellular matrix components and their receptors (23, 24). TGFP inhibition of the differentiation of myoblasts to myotubes is mediated by an increase in collagen I, fibronectin, and integrin receptor expression (23, 24, 35). Recent studies showed that TGF,B can have not only inhibitory but also growth stimulatory effects on highly metastatic cells (43). Perhaps the variability in responses is dictated by which matrix chemistry is induced by TGFP, a fact that should be cell type specific and dependent on synergies with other signals. Do synergies between plasma membrane-associated glycosaminoglycans and growth factors dictate the physiological responses to the growth factors? Our results lead us to

MOL. CELL. BIOL.

hypothesize that the plasma membrane glycosaminoglycan chemistry may influence the signal transduction mechanisms by growth factors such as IGF II and TGF,. At low growth factor concentrations or in the presence of heparan sulfates (produced when cells are in a state of growth), these factors would be mitogens. At high concentrations or in the presence of heparinlike glycosaminoglycans (produced when cells are at high density or in the quiescent state), the factors would be differentiation signals. Highly metastatic tumor cells either do not produce high enough concentrations of the growth factors or never generate the highly sulfated, heparinlike heparan sulfate proteoglycans that are hypothesized to drive the signal transduction process towards a differentiation pathway. This possibility is supported by the fact that tumor progression and the loss of differentiation in hepatomas is associated with loss in the activities of an epimerase and sulfatransferases responsible for the conversion of the poorly sulfated to the highly sulfated forms of heparan sulfate proteoglycans (41). Other growth factors have been shown to have biphasic functions: epidermal growth factor and insulin are both mitogens for adult hepatocytes at low density (11) or in the presence of heparan sulfates. However, they do not affect growth but rather affect the synthesis of liver-specific mRNAs in hepatocytes cultured at high density or in the presence of heparins (19). If our hypothesis proves true, synergies between hormones or growth factor and membrane-associated glycosaminoglycan chemistry could be part of the mechanism of normal liver cell differentiation in vivo as well as a differential regulator of minimally deviant versus metastatic tumors. ACKNOWLEDGMENTS We thank Dinish Williams for excellent technical assistance. Excellent secretarial assistance was given by Rosina Passela. This research was supported primarily by grant 1897 from the Council for Tobacco Research. Funding for some supplies and for technicians who helped with the liver perfusions, with animals, and with glassware washing was through grants from the American Cancer Society (BC-439) and National Institutes of Health (NIH/ NCI P30-CA13330 and AM17702-12). Lola Reid received salary support through a career development award NIH CA00783. Isabel Zvibel received partial salary support through the Molin Foundation. REFERENCES 1. Alexander, J. J., G. MacNab, and R. Saunders. 1978. Studies on in vitro production of hepatitis B surface antigen by a human hepatoma cell line. Perspect. Virol. 10:103-117. 2. Bell, G. I., D. S. Gerhard, N. M. Fong, R. Sanchez-Pescador, and L. B. Rall. 1985. Isolation of the human insulin-like growth factor genes: insulin-like growth factor II and insulin genes are contiguous. Proc. Natl. Acad. Sci. USA 82:6450-6454. 3. Bell, G. I., J. P. Merryweather, R. Sanchez-Pescador, M. M. Stempien, L. Priestley, J. Scott, and L. B. Rail. 1983. Sequence of a cDNA clone encoding human preproinsulin-like growth factor II. Nature (London) 310:775-777. 4. Brown, A. L., D. E. Graham, S. P. Nissley, D. J. Hill, A. J. Strain, and M. M. Rechler. 1986. Developmental regulation of insulin-like growth factor II mRNA in different rat tissues. J. Biol. Chem. 261:13144-13150. 5. Castellot, J. J., Jr., J. Choay, J.-C. Lormeau, M. Petitou, E. Sache, and M. J. Karnovsky. 1986. Structural determinants of the capacity of heparin to inhibit the proliferation of vascular smooth muscle cells. II. Evidence for a pentasaccharide sequence that contains a 3-0 sulfate group. J. Cell Biol. 102:19791984. 6. Castellot, J. J., Jr., D. L. Cochran, and M. J. Karnovsky. 1985. Effect of heparin on vascular smooth muscle cells. I. Cell metabolism. J. Cell. Physiol. 124:21-28.

VOL . 1 l, 1991

HEPARIN AND HORMONAL REGULATION OF mRNA SYNTHESIS

7. Clayton, D. F., and J. E. Darnell, Jr. 1983. Changes in liverspecific compared to common gene transcription during primary culture of mouse hepatocytes. Mol. Cell. Biol. 3:1552-1561. 8. DePagter-Holthuizen, P., M. Jansen, F. M. A. van Schaik, R. van der Kammen, C. OosterwUk, J. L. Van den Brande, and J. S. Sussenback. 1987. The human insulin-like growth factor II gene contains two development-specific promoters. FEBS Lett. 214: 259-264. 9. D'Ercole, A. J., A. D. Stiles, and L. E. Underwood. 1984. Tissue concentrations of somatomedin C: further evidence for multiple sites of synthesis and paracrine or autocrine mechanisms of action. Proc. Natl. Acad. Sci. USA 81:935-939. 10. Doerr, R., I. Zvibel, D. Chiuten, J. D'Olimpio, and L. M. Reid. 1989. Clonal growth of tumors on tissue-specific biomatrices and correlation with organ site specificity of metastases. Cancer Res. 49:384-392. 11. Enat, R., D. M. Jefferson, N. Ruiz-Opazo, Z. Gatmaitan, L. Leinwand, and L. M. Reid. 1984. Hepatocyte proliferation in vitro: its dependence on the use of serum-free, hormonally defined medium and substrata of extracellular matrix. Proc. Natl. Acad. Sci. USA 81:1411-1415. 12. Fausto, N., J. E. Mead, L. Braun, N. L. Thompson, M. Panzica, M. Goyette, G. I. Bell, and P. R. Shank. 1987. Proto-oncogene expression and growth factors during liver regeneration. Symp. Fundamental Cancer Res. 39:69-86. 13. Fedarko, N. S., and H. E. Conrad. 1986. A unique heparan sulfate in the nuclei of hepatocytes: structural changes with the growth state of the cells. J. Cell Biol. 102:587-599. 14. Feinberg, A. P., and B. Vogelstein. 1984. A technique for radio-labeling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 137:266-267. 15. Fogh, J., W. C. Wright, and J. D. Loveless. 1977. Absence of HeLa cell contamination in 169 cell lines derived from human tumors. J. Natl. Cancer Inst. 58:209-213. 16. Folkman, J., and M. Klagsbrun. 1987. Angiogenic factors. Science 235:442-447. 17. Fu, X. X., C. Y. Su, Y. Lee, R. Hintz, L. Biempica, R. Snyder, and C. E. Rogler. 1988. Insulin-like growth factor II expression and oval cell proliferation associated with hepatocarcinogenesis in woodchuck hepatitis virus carriers. J. Virol. 62:3422-3430. 18. Fujita, M., D. C. Spray, H. Choi, J. C. Saez, T. Watanabe, L. C. Rosenberg, E. L. Hertzberg, and L. M. Reid. 1987. Glycosaminoglycans and proteoglycans induce gap junction expression and restore transcription of tissue-specific mRNAs in primary liver cultures. Hepatology 7:1-9. 19. Gatmaitan, Z., D. M. Jefferson, N. Ruiz-Opazo, L. Biempica, I. Arias, G. Dudas, L. Leinwand, and L. M. Reid. 1983. Regulation of growth and differentiation of a rat hepatoma cell line by the synergistic interactions of hormones and collagenous substrata. J. Cell Biol. 97:1179-1190. 20. Han, V. K. M., P. K. Lund, D. C. Lee, and A. J. D'Ercole. 1988. Expression of somatomedin/insulin-like growth factor messenger ribonucleic acids in the human fetus: identification, characterization, and tissue distribution. J. Clin. Endocrinol. Metab. 66:422-429. 21. Hotta, M., and A. Baird. 1987. Differential effects on low density lipoprotein metabolism by transforming growth factor ,B mediates its effects on steroidogenesis in bovine adrenocortical cells in vitro. Endocrinology 121:150-159. 22. Houck, K. A., J. L. Cruise, and G. Michalopoulos. 1988. Norepinephrine modulates the growth-inhibitory effect of transforming growth factor P in primary rat hepatocyte cultures. J. Cell. Physiol. 135:551-555. 23. Ignotz, R. A., and J. Massague. 1986. Transforming growth factor-beta stimulates the expression of fibronectin and collagen and their incorporation into the extracellular matrix. J. Biol. Chem. 261:4337-4345. 24. Ignotz, R. A., and J. Massague. 1987. Cell adhesion protein receptors as targets for transforming growth factor P action. Cell 51:189-197. 25. Irminger, J.-C., K. M. Rosen, R. E. Humbel, and L. VillaKomaroff. 1987. Tissue-specific expression of insulin-like growth factor II mRNAs with distinct 5' untranslated regions.

115

Proc. Natl. Acad. Sci. USA 84:6330-6334. 26. Ishilara, M., N. S. Fedarko, and H. E. Conrad. 1986. Transport of heparan sulfate into the nuclei of hepatocytes. J. Biol. Chem. 261:13575-13580. 27. Kjellen, L., A. Oldberg, and M. Hook. 1981. Cell surface heparan sulfate: mechanisms of proteoglycan-cell association. J. Biol. Chem. 255:10407-10413. 28. Kjellen, L., I. Pettersson, and M. Hook. 1981. Cell-surface heparan sulfate: an intercalated membrane proteoglycan. Proc. Natl. Acad. Sci. USA 78:5371-5375. 29. Knowles, B. B., C. C. Howe, and D. P. Aden. 1980. Human hepatocellular carcinoma cell lines secrete the major plasma proteins and hepatitis B surface antigen. Science 209:497-499. 30. Little, M. H., G. Ablett, and P. J. Smith. 1987. Enhanced expression of insulin-like growth factor II is not a necessary event in Wilms' tumour progression. Carcinogenesis 8:865-868. 31. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual, p. 191-193. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 32. Mathews, L. S., G. Norstedt, and R. D. Palmiter. 1986. Regulation of insulin-like growth factor I gene expression by growth hormone. Proc. Natl. Acad. Sci. USA 83:9343-9347. 33. Nakajima, M., T. Irimura, D. DiFerrante, N. DiFerrante, and G. L. Nicolson. 1983. Heparan sulfate degradation: relation to tumor invasive and metastatic properties of mouse B16 melanoma sublines. Science 220:611-613. 34. Norstedt, G., A. Levinovitz, C. Moller, L. C. Eriksson, and G. Andersson. 1988. Expression of insulin-like growth factor I (IGF-1) and IGF-II mRNA during hepatic development, proliferation and carcinogenesis in the rat. Carcinogenesis 9:209-213. 35. Penttinen, R. P., S. Kobayashi, and P. Bornstein. 1988. Transforming growth factor 13 increases mRNA for matrix proteins both in the presence and in the absence of changes in mRNA stability. Proc. Natl. Acad. Sci. USA 85:1105-1108. 36. Ramasharma, K., and C. Hao Li. 1987. Human pituitary and placental hormones control human insulin-like growth factor II secretion in human granulosa cells. Proc. Natl. Acad. Sci. USA 84:2643-2647. 37. Reid, L. 1990. Stem cell biology, extracellular matrix, and liver differentiation. Curr. Opinions Cell Biol. 2:121-130. 38. Reilly, C. F., L. M. S. Fritze, and R. D. Rosenberg. 1986. Heparin inhibition of smooth muscle cell proliferation: a cellular site of action. J. Cell. Physiol. 129:11-19. 39. Reilly, C. F., L. M. S. Fritze, and R. D. Rosenberg. 1987. Antiproliferative effects of heparin on vascular smooth muscle cells are reversed by epidermal growth factor. J. Cell. Physiol. 131:149-157. 40. Roberts, A. B., M. A. Anzano, L. M. Wakefield, N. S. Roche, D. F. Stern, and M. B. Sporn. 1985. Type 1 transforming growth factor: a bifunctional regulator of cellular growth. Proc. Natl. Acad. Sci. USA 82:119-123. 41. Robinson, J., M. Viti, and M. Hook. 1984. Structure and properties of an under-sulfated heparan sulfate proteoglycan synthesized by a rat hepatoma cell line. J. Cell Biol. 98:946-953. 42. Schmid, C., T. Steiner, and E. R. Froesch. 1983. Preferential enhancement of myoblast differentiation by insulin-like growth factors (IGF I and IGF II) in primary cultures of chicken embryonic cells. FEBS Lett. 161:117-121. 43. Schwarz, L. C., M. C. Gingras, G. Goldberg, A. H. Greenberg, and J. A. Wright. 1988. Loss of growth factor dependence and conversion of transforming growth factor-P 1 inhibition to stimulation in metastatic H-ras-transformed murine fibroblasts. Cancer Res. 48:6999-7003. 44. Shen, S.-J., M. Diamon, C.-Y. Wang, M. Jansen, and J. Ilan. 1988. Isolation of an insulin-like growth factor II cDNA with a unique 5' untranslated region from human placenta. Proc. Natl. Acad. Sci. USA 85:1947-1951. 45. Shinizu, M., F. Torti, and R. A. Roth. 1986. Characterization of the insulin and insulin-like growth factor receptors and responsitivity of a fibroblast/adipocyte cell line before and after differentiation. Biochem. Biophys. Res. Commun. 137:552-558. 46. Sporn, M. B., and A. B. Roberts. 1985. Autocrine growth factors and cancer. Nature (London) 313:745-747.

116

MOL. CELL. BIOL.

ZVIBEL ET AL.

47. Spray, D. C., M. Fujita, J. C. Saez, H. Choi, T. Watanabe, E. Hertzberg, L. C. Rosenberg, and L. M. Reid. 1987. Glycosaminoglycans and proteoglycans induce gap junction synthesis and function in primary liver cultures. J. Cell Biol. 105:541-551. 48. Toilefsen, S. E., J. L. Sadow, and P. Rotwein. 1989. Coordinate expression of insulin-like growth factor II and its receptor during muscle differentiation. Proc. Natl. Acad. Sci. USA 86:1543-1547. 49. Tricoli, J. V., L. B. Rail, C. P. Karakousis, L. Herrera, N. J. Petrelli, G. I. Bell, and T. B. Shows. 1986. Enhanced levels of insulin-like growth factor messenger RNA in human colon carcinomas and liposarcomas. Cancer Res. 46:6169-6173. 50. Ueno, T., K. Takahashi, T. Matsuguchi, H. Endo, and M. Yamamoto. 1988. Transcriptional deviation of the rat insulinlike growth factor II gene initiated at three alternative leaderexons between neonatal tissue and ascites hepatomas. Biochim. Biophys. Acta 950:411-419. 51. Ueno, T., K. Takahashi, T. Matsuguchi, K. Ikejiri, H. Endo, and M. Yamamoto. 1988. Reactivation of rat insulin-like growth

52.

53. 54.

55.

factor II gene during hepatocarcinogenesis. Carcinogenesis 9:1779-1783. Uhlrich, S., 0. Lagente, J. Choay, Y. Courtois, and M. Lenfant. 1986. Structure activity relationship in heparin: stimulation of non-vascular cells by a synthetic heparin pentasaccharide in cooperation with human acidic fibroblast growth factors. Biochem. Biophys. Res. Commun. 139:728-732. Underwood, L. E., and A. J. D'Ercole. 1984. Insulin and insulin-like growth factors (somatomedins) in fetal and neonatal development. Clin. Endocrinol. Metab. 13:69-89. Wright, T. C., Jr., T. V. Johnstone, J. J. Casteilot, and M. J. Karnovsky. 1985. Inhibition of rat cervical epithelial cell growth by heparin and its reversal by EGF. J. Cell. Physiol. 125:499506. Yee, D., K. J. CuHlen, S. Paik, J. F. Perdue, B. Hampton, A. Schwartz, M. E. Lippman, and N. Rosen. 1988. Insulin-like growth factor II mRNA expression in human breast cancer. Cancer Res. 48:6691-6696.