Overexpression of Ovine Insulin-Like Growth Factor-l Stimulates Autonomous Autocrine or Paracrine Growth in Bovine Mammary-Derived Epithelial Cells
Donato Thomas
Romagnolo, R. Michael Akers, Eric A. Wong, B. McFadden, and Jeffrey D. Turner
Pat L. Boyle,
Department of Dairy and Animal Science Lactation Physiology and Biotechnology Laboratories Virginia Polytechnic Institute and State University (D.R., R.M.A., E.A.W., P.L.B.) Blacksburg, Virginia 24061-0315 Department of Animal Science University of Idaho (T.B.M.) Moscow, Idaho 83843 Department of Animal Science McGill University (J.D.T.) Montreal, Quebec, Canada 9X 1CO
INTRODUCTION
To test the hypothesis that insulin-like growth factorI (IGF-I) affects the growth of bovine mammary epithelial cells through an autocrine and/or paracrine pathway, a cell line (MD-IGF-I) was’ originated from MAC-T cells by cotransfection with a construct containing the cDNA for an ovine exon 2-encoded prepro-IGF-I under control of the mouse mammary tumor virus-long terminal repeat promoter. Clone MDIGF-I contained multiple copies of the plasmid integrated into the genome, expressed the highest level of IGF-I mRNA, and secreted radioimmunoactive IGF-I into the medium. The mitogenic activity of MDIGF-I cells was stimulated 80% by dexamethasone (DEX). The total DNA in MD-IGF-I cells was 2.5fold higher than that in parental MAC-T cells in the presence of DEX. Conditioned medium from MD-IGF-I cells, induced with DEX, stimulated [3H]thymidine incorporation into DNA of MAC-T cells and uninduced MD-IGF-I cells. These data provide evidence that IGF-I was secreted into medium by MD-IGF-I cells. It is suggested that IGF-I can stimulate the growth of mammary epithelial cells by an autocrine and/or paracrine mode of action. The MD-IGF-I cell line may be a suitable system to study translational and posttranslational modifications of IGF-I peptides. (Molecular Endocrinology 6: 1774-1780, 1992)
Insulin-like growth factor-l (IGF-I) plays an important role in mediating the growth and development of a number of tissues (1). The liver largely contributes to the pool of circulating IGF-I, which exerts an endocrine effect on peripheral target cells. Although intracellularly produced growth factors may associate with the plasma membrane in a “juxtacrine” pathway (2) or via a strict “intracrine” mode of action (3) a molecular basis for an autocrine or paracrine function of IGF-I has been postulated (4). Levels of IGF-I mRNA in the mammary gland of adult rats were lower than those in other tissues. Nevertheless, it is thought that the endocrine as well as autocrine or paracrine pathways probably contribute to the development of the mammary gland in a coordinate fashion during different stages of development (5, 6). Certainly, there are several demonstrations that exogenous IGF-I stimulates the proliferation of mammary tissue
from peripubertal and pregnant rodents (7) and ruminants (g-10). Also, both primary (11) and clonal (1214) mammary epithelial cells respond to IGF-I. However, is not known to what degree mammary tissue contributes to its own development by autocrine and/or paracrine production of IGF-I. As endogenous expression of IGF-I in the mammary gland is relatively low, mammary epithelial cells represent an excellent system for inducing overexpression of IGF-I to investigate the presence of autocrine and/or paracrine pathways as well as intracellular routing and
0888-8809/92/1774-l 780503 00/o Molecular Endocrmalogy CopyrIght 0 1992 by The Endocme Society
1774
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 18:56 For personal use only. No other uses without permission. . All rights reserved.
Overexpression
of IGF-I In Mammary
Epithelial
1775
Cells
posttranslational modification of IGF-I peptides (15). Because the physiological significance of IGF-I is functionally determined at the molecular and cellular levels, we tested the hypothesis of whether the growth of transformed mammary epithelial cells (MD-IGF-I) was stimulated by the synthesis and secretion of IGF-I originating from an ovine exon 2-encoded prepro-IGF-I (olGF-I) cDNA.
RESULTS An exon 2-containing olGF-I cDNA, encoding a 33amino acid leader peptide, was cloned into expression vector pMSG. The resulting plasmid, designated pMMTV-IGF-I (MMTV, mouse mammary tumor virus), contained the olGF-I cDNA under the control of the MMTV-long terminal repeat (MMTV-LTR) promoter (Fig. 1). Expression vector pMMTV-IGF-I contained a transcription initiation site 134 basepairs (bp) up-stream and a simian virus-40 (SV40) polyadenylation site 822 bp down-stream of the cloned olGF-I cDNA. Thus, the predicted length of the olGF-I transcript from pMMTVIGF-I was approximately 1.65 kilobases (kb). The MMTV-LTR promoter was shown to be active and inducible by dexamethasone (DEX) in the bovine mammary epithelial cell line MAC-T after transient transfection studies with plasmid pMSG-CAT (data not shown). Stable MAC-T transformants containing pMMTVIGF-I were isolated in Dulbecco’s Modified Eagle’s Medium (DMEM) plus hygromycin-B (HYG-B) after cotransformation with a plasmid conferring HYG-B resistance. The presence of integrated copies of pMMTVIGF-I was confirmed by Southern blotting (data not shown). Stable transformants contained single or multiple copies of pMMTV-IGF-I. One clone, MD-IGF-I, contained approximately 30 copies of pMMTV-IGF-I integrated in a tandem array and was used for further analysis.
Northern blot of total RNA extracted from MAC-T and MD-IGF-I cells showed (Fig. 2) that a RNA species of approximately 1.7 kb hybridized strongly when the MD-IGF-I cells were induced with DEX. The length of this transcript is in agreement with the predicted size of processed mRNAs transcribed from pMMTV-IGF-I. Densitometric analysis of autoradiograms showed a 40fold increase in the intensity of hybridizing RNA species compared to that in MAC-T cells. Even in the absence of DEX, MD-IGF-I cells demonstrated detectable quantities of this RNA species. To investigate the mitogenic actions of endogenously produced recombinant olGF-I, cells were plated in se-
1234
-
20s
1.7-
8) 100.0 s $ iii B Et E
ISY DMEM+O.l/rM
dex
80.0 60.0
;
40.0
4 ii
20.0
MAC-T
MD-IGF-I
CELL TYPE
Fig. 2. Northern Analysis of Total RNA from Parental MAC-T and MD-IGF-I Cells Cells were incubated for 24 h in the presence (0.1 PM) or absence of DEX. A, Total RNA was extracted and separated
(15 pg) on a 1% agarose gel in 2.2 Fig. 1. Cloning of olGF-I cDNA into the Expression Vector pMSG Expression of olGF-I cDNA was under the control of the MMTV-LTR promoter. The approximate sites of initiation of transcription (+l), start codon (ATG), leader sequence mature peptide (B), E-peptide (105 bp), 3’-untranslated region (3’ UTR; 271 bp), SV40 early splice, and SV40 poly(A) sites are indicated. The lower pane/ shows the predicted length of mRNA transcripts from the pMMTV-IGF-I cDNA construct.
M
formaldehyde and
transferred to nitrocellulose membrane. Blots were probed with a 32P-labeled IGF-I cDNA probe and washed at high stringency. RNA samples were from MAC-T cells with no DEX (lane 1) and with 0.1 PM DEX (lane 2) and MD-IGF-I cells with no DEX (lane 3) and with 0.1 PM DEX (lane 4). Migrations of
the 28s and 18s ribosomal RNAs are indicated. B, Relative expression of IGF-I mRNA was estimated by densitometric analysis of Northern blots. Values are expressed as a percentage of the maximal intensity.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 18:56 For personal use only. No other uses without permission. . All rights reserved.
MOL 1776
ENDO.
1992
Vo16No.11
lective medium in the absence of fetal calf serum (FCS), but in the presence of 0.0, 0.1, and 2.5 PM DEX. Total DNA from MAC-T and MD-IGF-I cells is depicted in Fig. 3. MD-IGF-I cells (Fig. 38) had the greatest growth response in association with induction of the MMTV-
AJ MAC-l
2.5 I hzs 0
‘2‘ 2.0 z < .30 1.5 I 0 a 6 c
1.0
DMEM DMEM+O.lpM DMEM+2.5pM
dex dex
A.llll
0.5
0.0
24
48
:
72
TIME (hr)
6)
MD-IGF-I
2.5T c
2.0
: 3
1.5
DMEM+O.lpM DMEM+Z.SpM
0
dex dex
s 0 1.0 i 6 I- 0.5
LTR promoter, and by 48 h, total DNA was 47-54% higher than that in parental MAC-T cells (Fig. 3A). By 72 h, total DNA from MD-IGF-I cells treated with 0.1 and 2.5 PM DEX was 80% higher than that in uninduced MD-IGF-I cells, but showed 2.4- and 2.5fold increases (P < 0.01) compared to that from DEX-treated MAC-T cells. Induction of MD-IGF-I cells with 0.1 PM DEX up to 240 h (Fig. 3C) provided further evidence of the responsiveness of MD-IGF-I cells to glucocorticoid stimulation (P < 0.01) and their capacity for continued growth. MD-IGF-I cells responded to exogenous IGF-I (100 rig/ml) with a 2-fold increase (P < 0.01) in total DNA in DMEM alone (Fig. 4). More importantly, treatment with 0.1 FM DEX caused a mitogenic response comparable to that promoted by 100 rig/ml exogenous IGF-I. Because binding proteins can interfere with RlAs for IGF-I, conditioned media from MAC-T and MD-IGF-I cells were extracted to confirm the secretion of IGF-I into the medium. The average IGF-I concentrations from two independent experiments are summarized in Table 1. IGF-I in conditioned medium from MAC-T cells ranged from 0.75-l .45 rig/ml (11.2-21.7 ng/106 cells). The average concentration of IGF-I for MD-IGF-I cells ranged from 1.85-7.10 rig/ml (27.7-l 06.5 rig/l O6 cells) in the absence and presence of DEX. To investigate whether the endogenously produced recombinant olGF-I was capable of stimulating the proliferation of MAC-T and MD-IGF-I cells, conditioned media from both cell types were used to study effects
3.0
24
72
40
CJ x -5 s
2.5
2
1.5
;1 E
1 .o
T
I DMEM tXXY DMEM+O.lpM
dex
2.0
TIME (hr) MD-IGF-I
0 5.0
c
4.5 T 4.0
Tw
3.5
3
3.0
2
I BJ
OMEM [email protected]
dex
0.5 0.0 - IGF-I
2.5 2.0
a E
1.5 1 .o 0.5 0.0 40
96
144 TIME (hr)
192
240
Fig. 3. Effect of DEX on Growth of MAC-T and MD-IGF-I Ceils In A and B, MAC-T and MD-IGF-I cells were cultured in DMEM (H), DMEM with 0.1 PM DEX (H), and DMEM with 2.5 PM DEX (Cl) for 24, 48, and 72 h. In C, MD-IGF-I cells were cultured for up to 240 h in DMEM (B) and DMEM plus 0.1 PM DEX (Q. At the end of each incubation, cells were harvested, and total DNA (micrograms per well) was determined. Bars represent the mean f SEM values from six samples, assayed in duplicate.
+ IGF-I
Fig. 4. Response of Bovine Mammary Epithelial Cells (MDIGF-I) to Exogenous IGF-I MD-IGF-I cells were cultured for 72 h in DMEM (m) and DMEM with 0.1 PM DEX (Q in the presence (100 rig/ml) or absence of IGF-I. At the end of the incubation period, cells were harvested, and total DNA (micrograms per well) was determined. Bars represent the mean f SEM values from four samples, assayed in duplicate.
Table 1. RIA IGF-I of Conditioned MAC-T and MD-IGF-I Cells after
Medium 72 h
ng IGF-l/ml Treatment
DMEM DMEM
+ DEX
from
Parental
ng IGF-I/I 0’ Cells
MAC-T
MD-IGF-I
MAC-T
MD-IGF-I
1.45 0.75
1.85 7.10
21.7 11.25
27.7 106.5
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 18:56 For personal use only. No other uses without permission. . All rights reserved.
Overexpression
of IGF-I in Mammary
Epithelial
Cells
on [3H]thymidine incorporation (Fig. 5). Figure 5A demonstrates that conditioned medium from MD-IGF-I ceils stimulated (P < 0.01) labeling of MAC-T cells to a greater extent than conditioned medium from MAC-T cells, particularly when MD-IGF-I cells were induced with DEX. Similarly, conditioned medium from MD-IGFI or MAC-T cells cultured with DEX markedly stimulated
1777
thymidine uptake into MD-IGF-I (Fig. 56). The response of MD-IGF-I ceils to medium from MAC-T cells incubated with DEX is due to induction of MMTV-IGF-I by residual DEX. While the addition of FCS increased thymidine incorporation by both cell types (Fig. 5C), the somewhat reduced uptake with conditioned medium, compared with fresh medium, probably reflected some medium exhaustion, since conditioned samples were not supplemented with fresh medium before testing.
MAC-T
A)
DMEM+O.lpM
DISCUSSION
dex
60.0 “0 x % -:
400
MD-IGF-I
MAC-T Conditioned
media
MD-IGF-I
8)
DMEM+O.lpM
dex
60.0 "0 x t -5
40.0
20.0
MAC-T
MD-IGF-I
Conditioned
media
Cl 200.0 175.0
f
I 'Si 0
DMEM O.l/lM dex 5% FCS
150.0+ 0
a
125.0
2
100.0
t
L? E 8
50.0 25.0
\
75.0 0.0 i MAC-T
1 MD-IGF-I
Fig. 5. Thymidine Incorporatron into DNA of MAC-T and MDIGF-I Cells MAC-T and MD-IGF-I cells were cultured for 72 h in DMEM (W) and DMEM with 0.1 PM DEX (RI). At the end of the incubatron period, conditioned medium was harvested and used to culture MAC-T (A) and MD-IGF-I (B) cells. [3H]Thymidine incorporation was determined after 16 h with a 2-h pulse. C, Thymidine uptake by MAC-T and MD-IGF-I cells in DMEM, DMEM plus 0.1 FM DEX, and DMEM plus 5% FCS. Bars represent the mean f SEM Of four ValueS.
The mitogenic effect of IGF-I is well documented for a large number of cell types. Among them, bovine mammary epithelial cells have been shown to respond to exogenous IGF-I (8, 9, 14). As expression of IGF-I by the mammary gland in adult rats (16) and pigs (17) was very low compared to that in other tissues, it has been suggested that local production of IGF-I may limit the growth of mammary epithelial cells, so that the pool of circulating IGF-I is relatively more important. On the other hand, efforts to stimulate local mammary tissue production of IGF-I might offer a mechanism for enhanced mammary development and milk production. To investigate whether IGF-I can stimulate the growth of bovine mammary epithelial cells by an autocrine and/ or paracrine mode of action, MAC-T cells were transformed with a construct containing an ovine IGF-I cDNA under control of the MMTV-LTR promoter. The MMTVLTR has been previously used in mammalian systems to induce expression of cloned cDNAs in the presence of glucocorticoids (18, 19). The presence of the transgene in MD-IGF-I cells increased steady state levels of IGF-I mRNA when the MMTV promoter was induced with DEX. Not surprisingly, Northern blot analysis of total RNA from MAC-T cells showed that endogenous production of IGF-I mRNA was very low, although production of IGF-I mRNA by bovine mammary tissue has been previously reported (20, 21). Because cell proliferation and thymidine incorporation by MD-IGF-I cells induced with DEX were comparably higher than that by MAC-T cells, it is suggested that the olGF-I transgene supported enhanced mitogenic activity. Although DEX stimulated the activity of the MMTV-LTR promoter, no difference in total DNA was observed between 0.1-2.5 PM DEX. MAC-T cells, on the other hand, did not show a growth response to DEX. Analysis of the growth response of MD-IGF-I cells for up to 10 days provided further evidence of their accelerated mitogenic activity compared with that of parental MAC-T cells. Although it could be argued that faster growth of MD-IGF-I ceils was due to an intracellular mechanism associated with IGF-I production, the fact that IGF-I was found in conditioned medium of MD-IGF-I cells from two independent experiments suggested that this cell line secreted recombinant olGF-I. Rates of IGF-I production (Table 1) are similar to those reported for
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 18:56 For personal use only. No other uses without permission. . All rights reserved.
MOL 1778
ENDO.
1992
expression of recombinant human IGF-I in Chinese hamster ovary and mouse L-cells (22). While dexamethasone caused a reduction in the steady state concentration of endogenous IGF-I mRNA in several tissues (23, 24), recent reports (25, 26) suggested that glucocorticoids reduced IGF-I bioactivity not by causing a reduction of serum IGF-I, but, rather, by sequestering IGF-I due to stimulated local production of IGF-l-binding proteins (IGFBPs). Although we did not measure whether changes in free IGF-I were concomitant with variations in the production of IGFBPs, it is possible that the DEX treatment may have affected the production of IGFBPs by MD-IGF-I cells, thereby influencing the availability of IGF-I (27, 28). Autonomous growth by transformed phenotypes has been previously used to investigate the presence of autocrine and/or paracrine pathways for the effects of human basic fibroblast growth factor in hamster kidneyderived cells (29) and IGF-II in MCF-7 cells (30). Similarly, introduction of the olGF-I gene in MAC-T cells triggered autonomous growth due to the secretion of recombinant IGF-I. If it is accepted that fundamental prerequisites for autocrine and/or paracrine stimulation are endogenous production of a growth factor, its secretion, corresponding presence of cell membrane receptors, and subsequent sQmulation of cell growth (5, 31), then our data suggest that the autocrine and/or paracrine pathways for IGF-I were active in these mammary epithelial MD-IGF-I cells. A related study (15) illustrated that the parent MAC-T cells exibit a proliferative response to exogenous IGF-I and express IGF-I receptors. In this study we used an exon 2-encoded preproIGF-I form of IGF-I encoding a 33-amino acid leader peptide. Consequently, one could ask whether the IGFI secreted in the medium represented the intact form of IGF-I or the truncated Des(l-3)IGF-I form, which has been shown to be biologically more active (27, 32). We plan to test this hypothesis by purification and sequence analysis of the IGF-I protein secreted by the MD-IGF-I cell line. We propose that the MD-IGF-I cell line may be a suitable system to study translational and posttranslational modifications leading to the production of various IGF-I peptides.
MATERIALS MMTV-LTR
AND METHODS Plasmids
A 0.7-kb fragment encodlng an exon 2 (previously exon 1A) (33) prepro-IGF-I cDNA containing a 33-amino acid leader signal peptlde was isolated from clone A21 by a Bglll restriction digest and purified from a low melting agaiose iel using the Geneclean system (BIO 101, Inc., La Jolla, CA). The ends of the fragmeni were illled in using T4 DNA polymerase, generating blunt ends. The blunted olGF-I Insert was cloned tnto the Smal sate of the 7.6.kb expresslon vector pMSG (Pharmacia, LKB Biotechnology, Inc., Piscataway, NJ). Thus, the olGF-I cDNA was placed under the control of the MMTV-LTR promoter contalned In the expression vector pMSG. To evaluate the activity of the MMTV promoter, bovine mammary epithelial
Vo16No.11
MAC-T cells (34) were pMSG-CAT (Pharmacia).
transiently
transfected
Development
MD-IGF-I
Transformants
of Stable
with
plasmid
MAC-T cells were plated at a density of 2 x lo5 in 60-mm tissue culture plates and cotransfected, using the calcium phosphate precipitation procedure, with the construct pMMTVIGF-I and a plasmid containing a cassette for resistance to HYG-B (35). Cells were alvcerol shocked for 2 min at 37 C. After incubaiion for 48 h ai37 C, cells were placed in selective medium containing 200 pg/ml HYG-B (Sigma, St. Louis, MO). Resistant cells were selected after 14 days, and single cells were cloned by the limiting dilution procedure using 96-well plates. For Southern blot analysis, genomic DNA was extracted (36), digested with EcoRI, and run on 1% agarose gels. The Southern blot was probed with a 0.7-kb olGF-I cDNA labeled by nick translation. Northern
Analysis
MD-IGF-I clones were plated in loo-mm tissue culture plates in DMEM-10% FCS with 200 pg/ml HYG-B. When cells were at 85% confluency, medium was removed, and cells were washed with Dulbecco’s PBS. Then, fresh DMEM medium containinq 0.1 UM DEX to induce the MMTV-LTR oromoter was added. Ceils were trypsinized after 24 h, and tbtal RNA was extracted using a guanidinium thiocyanate isolation procedure (37) and separated on a 1% agarose gel containing 2.2 M formaldehyde. Northern blots (Nitroplus, MSI, Inc., Westborough, MA) were hybridized overnight at 42 C and washed at high stringency according to the manufacturer’s instruction (MSI). Growth Medium
of MD-IGF-I
Cells
and Detection
of IGF-I
in
A series of experiments was designed to monitor the mitogenic activity of the MD-IGF-I cell line. For all growth experiments, MD-IGF-I and control MAC-T cells were plated at a density of 2 x lo4 cells In 24-well tissue culture plates in DMEM with 10% FCS. After 24 h, medium was removed, and cells were washed with PBS. In a first experiment, the test medium was DMEM without FCS, but with or without 200 pg/ml HYG-B and with or without 0.0, 0.1, and 2.5 pM DEX. Cells were collected 24, 48, and 72 h after induction with DEX, and total DNA per well was measured (38). Six replicates were collected for each treatment, and samples were assayed in duplicate. To test the sensitivity over time of the MD-IGF-I cells to induction with DEX, cells were induced with 0.1 and 2.5 PM DEX up to 240 h. At the end of each treatment, cells were harvested for determination of total DNA. The growth response of MD-IGF-I cells to exogenous human IGF-I (100 no/ml; Boehrinaer Mannheim. Indianaoolis. IN) was asceriained 7; the presence or absence of 0.1 ;M D’EX: Cells were harvested after 72 h, and total DNA was used as an indicator of mitogenic activity. To confirm the secretion of biologlcally active IGF-I in conditioned medium of MD-IGF-I cells, 5 x lo5 MD-IGF-I and MAC-T cells were plated in IOO-mm tissue culture plates for 72 h in the presence and absence of 0.1 PM DEX. At the end of the conditioning period, medium from each treatment was collected and used to culture MAC-T or MD-IGF-I cells, seeded in 24-well culture plates. After 16 h, cells were pulsed for 2 h with [3H]thymidine (ICN Blomedicals, Inc., Costa Mesa, CA), and incorporation was measured, as previously described (14). The presence of IGF-I in the medium was measured by a previously validated RIA procedure (39). Medium was subjected to methanol-formic acid extraction to correct for possible false positive results due to IGF-IBPs. After extraction, the amount of IGF-I associated with BPS in control serum samples was 8.0%, which confirmed that minimal amounts of BPS
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 18:56 For personal use only. No other uses without permission. . All rights reserved.
Overexpression
remarned coefficient 10%.
of IGF-I in Mammary
Epithelral
Cells
1779
in the samples tested by RIA. The RIA rntraassay of variation in control serum samples was less than 13.
Statistics Data are presented as the mean f SEM. The effects of DEX and trme on growth of MAC-T and MD-IGF-I cells were tested using the analysis of variance pr0cedur.e in the Statistical Analysis System (SAS, Cary, NC).
14.
Acknowledgments
15.
Received June 22, 1992. Fievisron received August 12, 1992. Accepted August 20, 1992. Address requests for reprints to: Dr. R. Michael Akers, Department of Darry Science, Lactatron Physiology Laboratory, Vrrgrnia Polytechnic Institute and State University, Blacksburg, Vrrginia 24061-0315.
16.
17.
18.
REFERENCES 19. 1. Srmmen FA 1991 Expression of the insulin-like growth factor-l gene and its products: complex regulation by tissue specific and hormonal factors. Dom Anim Endocrinol 8: 165-l 78 R, Lindquist PB, Nagashima M, Kohr W, Lipari 2. Brachmann T, Napier M, Derynck R 1989 Transmembrane TGF-cu precursors activate EGF/TGF-
Donato Thomas
Romagnolo, R. Michael Akers, Eric A. Wong, B. McFadden, and Jeffrey D. Turner
Pat L. Boyle,
Department of Dairy and Animal Science Lactation Physiology and Biotechnology Laboratories Virginia Polytechnic Institute and State University (D.R., R.M.A., E.A.W., P.L.B.) Blacksburg, Virginia 24061-0315 Department of Animal Science University of Idaho (T.B.M.) Moscow, Idaho 83843 Department of Animal Science McGill University (J.D.T.) Montreal, Quebec, Canada 9X 1CO
INTRODUCTION
To test the hypothesis that insulin-like growth factorI (IGF-I) affects the growth of bovine mammary epithelial cells through an autocrine and/or paracrine pathway, a cell line (MD-IGF-I) was’ originated from MAC-T cells by cotransfection with a construct containing the cDNA for an ovine exon 2-encoded prepro-IGF-I under control of the mouse mammary tumor virus-long terminal repeat promoter. Clone MDIGF-I contained multiple copies of the plasmid integrated into the genome, expressed the highest level of IGF-I mRNA, and secreted radioimmunoactive IGF-I into the medium. The mitogenic activity of MDIGF-I cells was stimulated 80% by dexamethasone (DEX). The total DNA in MD-IGF-I cells was 2.5fold higher than that in parental MAC-T cells in the presence of DEX. Conditioned medium from MD-IGF-I cells, induced with DEX, stimulated [3H]thymidine incorporation into DNA of MAC-T cells and uninduced MD-IGF-I cells. These data provide evidence that IGF-I was secreted into medium by MD-IGF-I cells. It is suggested that IGF-I can stimulate the growth of mammary epithelial cells by an autocrine and/or paracrine mode of action. The MD-IGF-I cell line may be a suitable system to study translational and posttranslational modifications of IGF-I peptides. (Molecular Endocrinology 6: 1774-1780, 1992)
Insulin-like growth factor-l (IGF-I) plays an important role in mediating the growth and development of a number of tissues (1). The liver largely contributes to the pool of circulating IGF-I, which exerts an endocrine effect on peripheral target cells. Although intracellularly produced growth factors may associate with the plasma membrane in a “juxtacrine” pathway (2) or via a strict “intracrine” mode of action (3) a molecular basis for an autocrine or paracrine function of IGF-I has been postulated (4). Levels of IGF-I mRNA in the mammary gland of adult rats were lower than those in other tissues. Nevertheless, it is thought that the endocrine as well as autocrine or paracrine pathways probably contribute to the development of the mammary gland in a coordinate fashion during different stages of development (5, 6). Certainly, there are several demonstrations that exogenous IGF-I stimulates the proliferation of mammary tissue
from peripubertal and pregnant rodents (7) and ruminants (g-10). Also, both primary (11) and clonal (1214) mammary epithelial cells respond to IGF-I. However, is not known to what degree mammary tissue contributes to its own development by autocrine and/or paracrine production of IGF-I. As endogenous expression of IGF-I in the mammary gland is relatively low, mammary epithelial cells represent an excellent system for inducing overexpression of IGF-I to investigate the presence of autocrine and/or paracrine pathways as well as intracellular routing and
0888-8809/92/1774-l 780503 00/o Molecular Endocrmalogy CopyrIght 0 1992 by The Endocme Society
1774
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 18:56 For personal use only. No other uses without permission. . All rights reserved.
Overexpression
of IGF-I In Mammary
Epithelial
1775
Cells
posttranslational modification of IGF-I peptides (15). Because the physiological significance of IGF-I is functionally determined at the molecular and cellular levels, we tested the hypothesis of whether the growth of transformed mammary epithelial cells (MD-IGF-I) was stimulated by the synthesis and secretion of IGF-I originating from an ovine exon 2-encoded prepro-IGF-I (olGF-I) cDNA.
RESULTS An exon 2-containing olGF-I cDNA, encoding a 33amino acid leader peptide, was cloned into expression vector pMSG. The resulting plasmid, designated pMMTV-IGF-I (MMTV, mouse mammary tumor virus), contained the olGF-I cDNA under the control of the MMTV-long terminal repeat (MMTV-LTR) promoter (Fig. 1). Expression vector pMMTV-IGF-I contained a transcription initiation site 134 basepairs (bp) up-stream and a simian virus-40 (SV40) polyadenylation site 822 bp down-stream of the cloned olGF-I cDNA. Thus, the predicted length of the olGF-I transcript from pMMTVIGF-I was approximately 1.65 kilobases (kb). The MMTV-LTR promoter was shown to be active and inducible by dexamethasone (DEX) in the bovine mammary epithelial cell line MAC-T after transient transfection studies with plasmid pMSG-CAT (data not shown). Stable MAC-T transformants containing pMMTVIGF-I were isolated in Dulbecco’s Modified Eagle’s Medium (DMEM) plus hygromycin-B (HYG-B) after cotransformation with a plasmid conferring HYG-B resistance. The presence of integrated copies of pMMTVIGF-I was confirmed by Southern blotting (data not shown). Stable transformants contained single or multiple copies of pMMTV-IGF-I. One clone, MD-IGF-I, contained approximately 30 copies of pMMTV-IGF-I integrated in a tandem array and was used for further analysis.
Northern blot of total RNA extracted from MAC-T and MD-IGF-I cells showed (Fig. 2) that a RNA species of approximately 1.7 kb hybridized strongly when the MD-IGF-I cells were induced with DEX. The length of this transcript is in agreement with the predicted size of processed mRNAs transcribed from pMMTV-IGF-I. Densitometric analysis of autoradiograms showed a 40fold increase in the intensity of hybridizing RNA species compared to that in MAC-T cells. Even in the absence of DEX, MD-IGF-I cells demonstrated detectable quantities of this RNA species. To investigate the mitogenic actions of endogenously produced recombinant olGF-I, cells were plated in se-
1234
-
20s
1.7-
8) 100.0 s $ iii B Et E
ISY DMEM+O.l/rM
dex
80.0 60.0
;
40.0
4 ii
20.0
MAC-T
MD-IGF-I
CELL TYPE
Fig. 2. Northern Analysis of Total RNA from Parental MAC-T and MD-IGF-I Cells Cells were incubated for 24 h in the presence (0.1 PM) or absence of DEX. A, Total RNA was extracted and separated
(15 pg) on a 1% agarose gel in 2.2 Fig. 1. Cloning of olGF-I cDNA into the Expression Vector pMSG Expression of olGF-I cDNA was under the control of the MMTV-LTR promoter. The approximate sites of initiation of transcription (+l), start codon (ATG), leader sequence mature peptide (B), E-peptide (105 bp), 3’-untranslated region (3’ UTR; 271 bp), SV40 early splice, and SV40 poly(A) sites are indicated. The lower pane/ shows the predicted length of mRNA transcripts from the pMMTV-IGF-I cDNA construct.
M
formaldehyde and
transferred to nitrocellulose membrane. Blots were probed with a 32P-labeled IGF-I cDNA probe and washed at high stringency. RNA samples were from MAC-T cells with no DEX (lane 1) and with 0.1 PM DEX (lane 2) and MD-IGF-I cells with no DEX (lane 3) and with 0.1 PM DEX (lane 4). Migrations of
the 28s and 18s ribosomal RNAs are indicated. B, Relative expression of IGF-I mRNA was estimated by densitometric analysis of Northern blots. Values are expressed as a percentage of the maximal intensity.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 18:56 For personal use only. No other uses without permission. . All rights reserved.
MOL 1776
ENDO.
1992
Vo16No.11
lective medium in the absence of fetal calf serum (FCS), but in the presence of 0.0, 0.1, and 2.5 PM DEX. Total DNA from MAC-T and MD-IGF-I cells is depicted in Fig. 3. MD-IGF-I cells (Fig. 38) had the greatest growth response in association with induction of the MMTV-
AJ MAC-l
2.5 I hzs 0
‘2‘ 2.0 z < .30 1.5 I 0 a 6 c
1.0
DMEM DMEM+O.lpM DMEM+2.5pM
dex dex
A.llll
0.5
0.0
24
48
:
72
TIME (hr)
6)
MD-IGF-I
2.5T c
2.0
: 3
1.5
DMEM+O.lpM DMEM+Z.SpM
0
dex dex
s 0 1.0 i 6 I- 0.5
LTR promoter, and by 48 h, total DNA was 47-54% higher than that in parental MAC-T cells (Fig. 3A). By 72 h, total DNA from MD-IGF-I cells treated with 0.1 and 2.5 PM DEX was 80% higher than that in uninduced MD-IGF-I cells, but showed 2.4- and 2.5fold increases (P < 0.01) compared to that from DEX-treated MAC-T cells. Induction of MD-IGF-I cells with 0.1 PM DEX up to 240 h (Fig. 3C) provided further evidence of the responsiveness of MD-IGF-I cells to glucocorticoid stimulation (P < 0.01) and their capacity for continued growth. MD-IGF-I cells responded to exogenous IGF-I (100 rig/ml) with a 2-fold increase (P < 0.01) in total DNA in DMEM alone (Fig. 4). More importantly, treatment with 0.1 FM DEX caused a mitogenic response comparable to that promoted by 100 rig/ml exogenous IGF-I. Because binding proteins can interfere with RlAs for IGF-I, conditioned media from MAC-T and MD-IGF-I cells were extracted to confirm the secretion of IGF-I into the medium. The average IGF-I concentrations from two independent experiments are summarized in Table 1. IGF-I in conditioned medium from MAC-T cells ranged from 0.75-l .45 rig/ml (11.2-21.7 ng/106 cells). The average concentration of IGF-I for MD-IGF-I cells ranged from 1.85-7.10 rig/ml (27.7-l 06.5 rig/l O6 cells) in the absence and presence of DEX. To investigate whether the endogenously produced recombinant olGF-I was capable of stimulating the proliferation of MAC-T and MD-IGF-I cells, conditioned media from both cell types were used to study effects
3.0
24
72
40
CJ x -5 s
2.5
2
1.5
;1 E
1 .o
T
I DMEM tXXY DMEM+O.lpM
dex
2.0
TIME (hr) MD-IGF-I
0 5.0
c
4.5 T 4.0
Tw
3.5
3
3.0
2
I BJ
OMEM [email protected]
dex
0.5 0.0 - IGF-I
2.5 2.0
a E
1.5 1 .o 0.5 0.0 40
96
144 TIME (hr)
192
240
Fig. 3. Effect of DEX on Growth of MAC-T and MD-IGF-I Ceils In A and B, MAC-T and MD-IGF-I cells were cultured in DMEM (H), DMEM with 0.1 PM DEX (H), and DMEM with 2.5 PM DEX (Cl) for 24, 48, and 72 h. In C, MD-IGF-I cells were cultured for up to 240 h in DMEM (B) and DMEM plus 0.1 PM DEX (Q. At the end of each incubation, cells were harvested, and total DNA (micrograms per well) was determined. Bars represent the mean f SEM values from six samples, assayed in duplicate.
+ IGF-I
Fig. 4. Response of Bovine Mammary Epithelial Cells (MDIGF-I) to Exogenous IGF-I MD-IGF-I cells were cultured for 72 h in DMEM (m) and DMEM with 0.1 PM DEX (Q in the presence (100 rig/ml) or absence of IGF-I. At the end of the incubation period, cells were harvested, and total DNA (micrograms per well) was determined. Bars represent the mean f SEM values from four samples, assayed in duplicate.
Table 1. RIA IGF-I of Conditioned MAC-T and MD-IGF-I Cells after
Medium 72 h
ng IGF-l/ml Treatment
DMEM DMEM
+ DEX
from
Parental
ng IGF-I/I 0’ Cells
MAC-T
MD-IGF-I
MAC-T
MD-IGF-I
1.45 0.75
1.85 7.10
21.7 11.25
27.7 106.5
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 18:56 For personal use only. No other uses without permission. . All rights reserved.
Overexpression
of IGF-I in Mammary
Epithelial
Cells
on [3H]thymidine incorporation (Fig. 5). Figure 5A demonstrates that conditioned medium from MD-IGF-I ceils stimulated (P < 0.01) labeling of MAC-T cells to a greater extent than conditioned medium from MAC-T cells, particularly when MD-IGF-I cells were induced with DEX. Similarly, conditioned medium from MD-IGFI or MAC-T cells cultured with DEX markedly stimulated
1777
thymidine uptake into MD-IGF-I (Fig. 56). The response of MD-IGF-I ceils to medium from MAC-T cells incubated with DEX is due to induction of MMTV-IGF-I by residual DEX. While the addition of FCS increased thymidine incorporation by both cell types (Fig. 5C), the somewhat reduced uptake with conditioned medium, compared with fresh medium, probably reflected some medium exhaustion, since conditioned samples were not supplemented with fresh medium before testing.
MAC-T
A)
DMEM+O.lpM
DISCUSSION
dex
60.0 “0 x % -:
400
MD-IGF-I
MAC-T Conditioned
media
MD-IGF-I
8)
DMEM+O.lpM
dex
60.0 "0 x t -5
40.0
20.0
MAC-T
MD-IGF-I
Conditioned
media
Cl 200.0 175.0
f
I 'Si 0
DMEM O.l/lM dex 5% FCS
150.0+ 0
a
125.0
2
100.0
t
L? E 8
50.0 25.0
\
75.0 0.0 i MAC-T
1 MD-IGF-I
Fig. 5. Thymidine Incorporatron into DNA of MAC-T and MDIGF-I Cells MAC-T and MD-IGF-I cells were cultured for 72 h in DMEM (W) and DMEM with 0.1 PM DEX (RI). At the end of the incubatron period, conditioned medium was harvested and used to culture MAC-T (A) and MD-IGF-I (B) cells. [3H]Thymidine incorporation was determined after 16 h with a 2-h pulse. C, Thymidine uptake by MAC-T and MD-IGF-I cells in DMEM, DMEM plus 0.1 FM DEX, and DMEM plus 5% FCS. Bars represent the mean f SEM Of four ValueS.
The mitogenic effect of IGF-I is well documented for a large number of cell types. Among them, bovine mammary epithelial cells have been shown to respond to exogenous IGF-I (8, 9, 14). As expression of IGF-I by the mammary gland in adult rats (16) and pigs (17) was very low compared to that in other tissues, it has been suggested that local production of IGF-I may limit the growth of mammary epithelial cells, so that the pool of circulating IGF-I is relatively more important. On the other hand, efforts to stimulate local mammary tissue production of IGF-I might offer a mechanism for enhanced mammary development and milk production. To investigate whether IGF-I can stimulate the growth of bovine mammary epithelial cells by an autocrine and/ or paracrine mode of action, MAC-T cells were transformed with a construct containing an ovine IGF-I cDNA under control of the MMTV-LTR promoter. The MMTVLTR has been previously used in mammalian systems to induce expression of cloned cDNAs in the presence of glucocorticoids (18, 19). The presence of the transgene in MD-IGF-I cells increased steady state levels of IGF-I mRNA when the MMTV promoter was induced with DEX. Not surprisingly, Northern blot analysis of total RNA from MAC-T cells showed that endogenous production of IGF-I mRNA was very low, although production of IGF-I mRNA by bovine mammary tissue has been previously reported (20, 21). Because cell proliferation and thymidine incorporation by MD-IGF-I cells induced with DEX were comparably higher than that by MAC-T cells, it is suggested that the olGF-I transgene supported enhanced mitogenic activity. Although DEX stimulated the activity of the MMTV-LTR promoter, no difference in total DNA was observed between 0.1-2.5 PM DEX. MAC-T cells, on the other hand, did not show a growth response to DEX. Analysis of the growth response of MD-IGF-I cells for up to 10 days provided further evidence of their accelerated mitogenic activity compared with that of parental MAC-T cells. Although it could be argued that faster growth of MD-IGF-I ceils was due to an intracellular mechanism associated with IGF-I production, the fact that IGF-I was found in conditioned medium of MD-IGF-I cells from two independent experiments suggested that this cell line secreted recombinant olGF-I. Rates of IGF-I production (Table 1) are similar to those reported for
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 18:56 For personal use only. No other uses without permission. . All rights reserved.
MOL 1778
ENDO.
1992
expression of recombinant human IGF-I in Chinese hamster ovary and mouse L-cells (22). While dexamethasone caused a reduction in the steady state concentration of endogenous IGF-I mRNA in several tissues (23, 24), recent reports (25, 26) suggested that glucocorticoids reduced IGF-I bioactivity not by causing a reduction of serum IGF-I, but, rather, by sequestering IGF-I due to stimulated local production of IGF-l-binding proteins (IGFBPs). Although we did not measure whether changes in free IGF-I were concomitant with variations in the production of IGFBPs, it is possible that the DEX treatment may have affected the production of IGFBPs by MD-IGF-I cells, thereby influencing the availability of IGF-I (27, 28). Autonomous growth by transformed phenotypes has been previously used to investigate the presence of autocrine and/or paracrine pathways for the effects of human basic fibroblast growth factor in hamster kidneyderived cells (29) and IGF-II in MCF-7 cells (30). Similarly, introduction of the olGF-I gene in MAC-T cells triggered autonomous growth due to the secretion of recombinant IGF-I. If it is accepted that fundamental prerequisites for autocrine and/or paracrine stimulation are endogenous production of a growth factor, its secretion, corresponding presence of cell membrane receptors, and subsequent sQmulation of cell growth (5, 31), then our data suggest that the autocrine and/or paracrine pathways for IGF-I were active in these mammary epithelial MD-IGF-I cells. A related study (15) illustrated that the parent MAC-T cells exibit a proliferative response to exogenous IGF-I and express IGF-I receptors. In this study we used an exon 2-encoded preproIGF-I form of IGF-I encoding a 33-amino acid leader peptide. Consequently, one could ask whether the IGFI secreted in the medium represented the intact form of IGF-I or the truncated Des(l-3)IGF-I form, which has been shown to be biologically more active (27, 32). We plan to test this hypothesis by purification and sequence analysis of the IGF-I protein secreted by the MD-IGF-I cell line. We propose that the MD-IGF-I cell line may be a suitable system to study translational and posttranslational modifications leading to the production of various IGF-I peptides.
MATERIALS MMTV-LTR
AND METHODS Plasmids
A 0.7-kb fragment encodlng an exon 2 (previously exon 1A) (33) prepro-IGF-I cDNA containing a 33-amino acid leader signal peptlde was isolated from clone A21 by a Bglll restriction digest and purified from a low melting agaiose iel using the Geneclean system (BIO 101, Inc., La Jolla, CA). The ends of the fragmeni were illled in using T4 DNA polymerase, generating blunt ends. The blunted olGF-I Insert was cloned tnto the Smal sate of the 7.6.kb expresslon vector pMSG (Pharmacia, LKB Biotechnology, Inc., Piscataway, NJ). Thus, the olGF-I cDNA was placed under the control of the MMTV-LTR promoter contalned In the expression vector pMSG. To evaluate the activity of the MMTV promoter, bovine mammary epithelial
Vo16No.11
MAC-T cells (34) were pMSG-CAT (Pharmacia).
transiently
transfected
Development
MD-IGF-I
Transformants
of Stable
with
plasmid
MAC-T cells were plated at a density of 2 x lo5 in 60-mm tissue culture plates and cotransfected, using the calcium phosphate precipitation procedure, with the construct pMMTVIGF-I and a plasmid containing a cassette for resistance to HYG-B (35). Cells were alvcerol shocked for 2 min at 37 C. After incubaiion for 48 h ai37 C, cells were placed in selective medium containing 200 pg/ml HYG-B (Sigma, St. Louis, MO). Resistant cells were selected after 14 days, and single cells were cloned by the limiting dilution procedure using 96-well plates. For Southern blot analysis, genomic DNA was extracted (36), digested with EcoRI, and run on 1% agarose gels. The Southern blot was probed with a 0.7-kb olGF-I cDNA labeled by nick translation. Northern
Analysis
MD-IGF-I clones were plated in loo-mm tissue culture plates in DMEM-10% FCS with 200 pg/ml HYG-B. When cells were at 85% confluency, medium was removed, and cells were washed with Dulbecco’s PBS. Then, fresh DMEM medium containinq 0.1 UM DEX to induce the MMTV-LTR oromoter was added. Ceils were trypsinized after 24 h, and tbtal RNA was extracted using a guanidinium thiocyanate isolation procedure (37) and separated on a 1% agarose gel containing 2.2 M formaldehyde. Northern blots (Nitroplus, MSI, Inc., Westborough, MA) were hybridized overnight at 42 C and washed at high stringency according to the manufacturer’s instruction (MSI). Growth Medium
of MD-IGF-I
Cells
and Detection
of IGF-I
in
A series of experiments was designed to monitor the mitogenic activity of the MD-IGF-I cell line. For all growth experiments, MD-IGF-I and control MAC-T cells were plated at a density of 2 x lo4 cells In 24-well tissue culture plates in DMEM with 10% FCS. After 24 h, medium was removed, and cells were washed with PBS. In a first experiment, the test medium was DMEM without FCS, but with or without 200 pg/ml HYG-B and with or without 0.0, 0.1, and 2.5 pM DEX. Cells were collected 24, 48, and 72 h after induction with DEX, and total DNA per well was measured (38). Six replicates were collected for each treatment, and samples were assayed in duplicate. To test the sensitivity over time of the MD-IGF-I cells to induction with DEX, cells were induced with 0.1 and 2.5 PM DEX up to 240 h. At the end of each treatment, cells were harvested for determination of total DNA. The growth response of MD-IGF-I cells to exogenous human IGF-I (100 no/ml; Boehrinaer Mannheim. Indianaoolis. IN) was asceriained 7; the presence or absence of 0.1 ;M D’EX: Cells were harvested after 72 h, and total DNA was used as an indicator of mitogenic activity. To confirm the secretion of biologlcally active IGF-I in conditioned medium of MD-IGF-I cells, 5 x lo5 MD-IGF-I and MAC-T cells were plated in IOO-mm tissue culture plates for 72 h in the presence and absence of 0.1 PM DEX. At the end of the conditioning period, medium from each treatment was collected and used to culture MAC-T or MD-IGF-I cells, seeded in 24-well culture plates. After 16 h, cells were pulsed for 2 h with [3H]thymidine (ICN Blomedicals, Inc., Costa Mesa, CA), and incorporation was measured, as previously described (14). The presence of IGF-I in the medium was measured by a previously validated RIA procedure (39). Medium was subjected to methanol-formic acid extraction to correct for possible false positive results due to IGF-IBPs. After extraction, the amount of IGF-I associated with BPS in control serum samples was 8.0%, which confirmed that minimal amounts of BPS
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 18:56 For personal use only. No other uses without permission. . All rights reserved.
Overexpression
remarned coefficient 10%.
of IGF-I in Mammary
Epithelral
Cells
1779
in the samples tested by RIA. The RIA rntraassay of variation in control serum samples was less than 13.
Statistics Data are presented as the mean f SEM. The effects of DEX and trme on growth of MAC-T and MD-IGF-I cells were tested using the analysis of variance pr0cedur.e in the Statistical Analysis System (SAS, Cary, NC).
14.
Acknowledgments
15.
Received June 22, 1992. Fievisron received August 12, 1992. Accepted August 20, 1992. Address requests for reprints to: Dr. R. Michael Akers, Department of Darry Science, Lactatron Physiology Laboratory, Vrrgrnia Polytechnic Institute and State University, Blacksburg, Vrrginia 24061-0315.
16.
17.
18.
REFERENCES 19. 1. Srmmen FA 1991 Expression of the insulin-like growth factor-l gene and its products: complex regulation by tissue specific and hormonal factors. Dom Anim Endocrinol 8: 165-l 78 R, Lindquist PB, Nagashima M, Kohr W, Lipari 2. Brachmann T, Napier M, Derynck R 1989 Transmembrane TGF-cu precursors activate EGF/TGF-
