GASTROENTEROLOGY
LIVER, PANCREAS,
1992;103:1851-1859
AND BILIARY TRACT
Decreased Expression of Transforming Growth Factor a During Differentiation of Human Pancreatic Cancer Cells AGNES ESTIVAL, PASCAL and FRANCOIS CLEMENTE
CLERC,
NICOLE
INSERM U 151, Institut L. Bugnard, Centre Hospitalier Rockefeller University, New York, New York
The relationship between cell differentiation and transforming growth factor 01(TGF-a) expression in human pancreatic cancer cells was analyzed in Capan 1 cells. These cells differentiate either spontaneously or after butyrate treatment. During differentiation (spontaneous or butyrate induced), TGF-a messenger RNA (mRNA) levels decreased, whereas the TGF-fil mRNA levels remained unchanged. TGF-a was present in cells as proTGF-a, which decreased after butyrate treatment. Secretion of TGFa was not found. Under the two conditions of differentiation, the membrane-bound protein kinase C activity was also reduced. Conversely, long-term phorbol ester treatment increased both membranebound protein kinase C activity (260%) and TGF-a mRNA level (500%), a not significant increase of TGF-j31 mRNA was observed. However, phorbol l&myristate-13-acetate did not induce TGF-a synthesis or secretion. These data suggest that expression of TGF-a can be reduced in cancer cells; they also suggest the existence of a relationship between TGF-a expression and cell differentiation. In addition, the protein kinase C-induced TGF-a mRNA level was not followed by the increase of TGF-a biosynthesis, suggesting a translational control. Finally, the expression of TGF-a and -plmessengers appears to be differently regulated. ransforming growth factor (Y (TGF-a) was initially purified from transformed cells.’ Later on, TGF-a messenger RNA (mRNA) was found in various tissues during embryonic development’ and at low levels in some normal adult cell types.3,4 It is generally accepted that its biological effects are mediated through epidermal growth factor (EGF) receptors.’ TGF-a promotes cell proliferation in vivo and in vitro, and its involvement in the autocrine control of cancer cell growth has been strongly suggested in several reports.68
T
VAYSSE,
Universitaire
JAMES
Rangueil, Toulouse,
P. TAM, France; and
It should be emphasized that TGF-a is not a product of any oncogene but rather is synthesized by tumor cells following the activation of oncogenes.’ On the other hand, in normal cells, the expression of TGF-a appears to be somehow related to their differentiation state. Indeed, it is highly expressed during normal embryonic development when cells are poorly differentiated, whereas it is weakly represented in normal, fully differentiated cells. Thus, the level of TGF-a expression by cancer cells could be related, to a certain degree, to their differentiation state and be reduced by cell differentiation. The possibility of decreasing TGF-a expression in human pancreatic cancer cells and the mechanisms responsible for that decrease are of great interest for their therapeutic implications. To test the hypothesis of a differentiation-related TGF-a expression in pancreatic cancer cells, we followed the expression of TGF-a during cell differentiation. We analyzed TGF-a and -fll mRNAs in parallel because the TGF-Pl is expressed by different cell types whether they are differentiated or not.‘*” Because it has recently been shown that TGF-a is increased by protein kinase C activation,” we also analyzed the modulation of this enzyme during cell differentiation. The human pancreatic cancer cell line Capan 1 was chosen for this study because of its interesting properties. It has been shown that Capan 1 cells spontaneously acquire more differentiated phenotypes during cell agingl’ and also under the action of sodium butyrate.13p14 Differentiation is evidenced morphologically by cell polarization and by apical junction complexes, and biochemically by increased alkaline phosphatase and aminopeptidase N levels.14 Our results show that during differentiation of Capan 1 cells, both TGF-a expression and activation of protein kinase C decrease. 0
1992 by the American Gastroenterological 0016-5085/92/$3.00
Association
1852
ESTIVAL ET AL.
GASTROENTEROLOGY Vol. 103, No. 6
Materials and Methods Materials RPM1 1640 and fetal calf serum (FCS) were obtained from Gibco (Grand Island, NY). Phorbol 12-myristate-13-acetate (PMA), 4a-phorbol 12,13didecanoate, histone Hl, phenylmethylsulfonyl fluoride (PMSF), and sodium butyrate were purchased from Sigma (St. Louis, MO) and the ion exchanger DE52 cellulose from Whatman (Maidstone, England). The Capan 1 cell line was kindly provided by Dr. J. Fogh (Sloan-Kettering Institute for Cancer Research, New York, NY). [y-32P]adenosine triphosphate ([y-32P]ATP) and [35S]cysteine were obtained from Amersham (Les Ulis, France). Cell Culture
and Cell Treatments
The Capan 1 cells were plated in 180-cm2 flasks at a density of lo* cells/cm’ and grown in RPM1 1640 supplemented with 10% FCS. After a 16-hour attachment period, media were replaced with fresh medium and cells were treated for 3 days with PMA (50 nmol/L) or 4a-phorbol12,13-didecanoate (50 nmol/L) or sodium butyrate (2 mmol/L). The different determinations were performed the fourth day after seeding; the day before, media were replaced with fresh medium containing the drugs. For the proliferation studies, the Capan 1 cells were grown in the presence of the drugs, and the medium containing the drugs was replaced every other day. Cells were counted after trypsinization using a Coulter Counter (model ZM; Coulter Electronics, Margency, France). Protein
Kinase
C Assay
Protein kinase C activity was assayed as previously describedI by measuring the incorporation of 32P from [y32PI ATP into histone Hl. A control assay was performed in the absence of diacylglycerol in the assay tube. The soluble and particulate cell fractions containing the protein kinase C were obtained after a 100,OOOg centrifugation then chromatographied on ion-exchanger DE52 cellulose and recovered in the chromatographic fraction eluted by 50 mmol/L NaCl before any enzyme activity assay.15 RNA Extraction
and Messenger
Quantification
Total RNA was extracted according to the method of Chirgwin et a1.16Twenty micrograms of total RNA was electrophoretically fractionated on 1% agarose gel in denaturing conditions and then transferred onto nitrocellulose filters.17 The concentration of individual mRNA samples was determined by slot-blot hybridization analysis. Complementary DNA (cDNA) probes were 32P-labeled (sp act, lo7 cpm/ug) by nick-translation (Amersham kit). Hybridizations were carried out in a mixture of 50% formamide, 6X SSC, 0.5% sodium dodecyl sulfate (SDS), 5X Denhardt’s, 10 mmol/L ethylenediamine tetraacetic acid (EDTA), and 100 ug/mL salmon sperm DNA at 42°C for 16 hours. Filters were washed in 2X SSC and 0.1% SDS one time for 20 minutes at room temperature and two times for 20 minutes in 1X SSC and 0.1% SDS at 45“C. Filters were air-dried and then exposed to Kodak X-Omat AR films
(Rochester, New York). Spots were quantified using an LKB laser densitometer (Pharmacia, France). Human-specific probes for TGF-a and -pl were kindly provided by Dr R. Derynck (Genentech, San Francisco, CA). The TGF-a cDNA probe contained the complete coding sequence, and the TGF-Pl cDNA was a 1050-base pair-long probee Immunoprecipitation Electrophoresis
of TGF-a
and
Gel
The immunoprecipitation was performed according to the method of Bringman et al.“’ Two polyclonal antibodies were used. The first was raised against human TGF-a peptide and the second against the intracellular Cterminal sequence of human TGF-a precursors.5 TGF-a was characterized in cells by incorporation studies. Cells were incubated for 16 hours in serum-free culture media containing 100 uCi/mL of [35S]cysteine then washed three times with phosphate-buffered saline (PBS) and harvested by scraping in PBS containing 1% NP-40,10 mmol/L EDTA, and 1 mmol/L PMSF at 4’C. Samples were diluted with 50 mmol/L Tris-HCl, 0.15 mol/L NaCl, 2 mmol/L EDTA, 1% NP-40, 0.1% SDS, 0.1% sodium azide, and 1 mg/mL bovine serum albumin at pH 8.0 and then incubated in parallel with 10 uL of purified polyclonal antibodies for 16 hours at 4°C. The protein A-sepharose suspension (Pharmacia LKB, Uppsala, Sweden) was then added and the mixture incubated for 2 hours at 25°C under rotary agitation. After five washings, beads were boiled at 100°C for 5 minutes in 100 uL of 2X concentrated electrophoresis sample buffer.lg Beads were eliminated by centrifugation and samples analyzed by electrophoresis according to the method of Laemmlilg in denaturing 15% polyacrylamide gels, which were dried and exposed to Kodak X-Omat AR films. TGF-a produced by cells cultured in the presence of butyrate or PMA was analyzed and quantified by Western blot. Cells were kept in serum-free media 24 hours before their recovery. Cells were washed and lysed as for incorporation studies. Culture media were clarified by centrifugation, acidified by acetic acid to a final concentration of 0.1 mol/L, dialyzed in Spectrapore tubing (3000-kilodalton cutoff) against 100 volumes of 0.1 mol/L acetic acid and 0.1 mmol/L PMSF at 4”C, and lyophilized.” Samples of cells and culture media were immunoprecipitated as described above. Immunoprecipitates were subjected to electrophoresis and electroblotted onto nitrocellulose membranes. Membranes were preincubated for 1 hour at 25’C in PBS containing 5% nonfat dry milk and 0.5% Tween 20, then incubated overnight with anti-TGF-a or anti-c-terminal sequence of proTGF-a, diluted in PBS containing 0.5% Tween 20. Membranes were then rinsed three times with PBS containing 0.5% Tween 20 and incubated for 1 hour at 25°C with anti-rabbit antibodies coupled to peroxydase. After three more washes, protein bands were detected by chemiluminescence on autoradiography films (ECL, Amersham) and quantified by laser densitometry. Control experiments were carried out in parallel by using rabbit nonimmune sera for immunoprecipitations. Human TGF-a of 5.6 kilodaltons was also used as a control during electrophoresis and immunoblotting.
December
TGF-a
1992
Aminopeptidase
N Assay
a
AND CELL DIFFERENTIATION
1853
b
Aminopeptidase N activity was determined as previously described14 using L-alanine-p-nitroanilide as a substrate at a concentration of 1.5 mmol/L. Proteins Protein levels were measured according to the method of Lowry et al.” Results Expression
25.7
of TGFs by Capan 1 Cells
TGF mRNAs and protein. Total RNA was electrophoretically fractionated in agarose gel (1%) in denaturing conditions and analyzed by blot hybridization. Northern blots analysis revealed the existence of a single band corresponding to TGF-a mRNA of 4.7 kilobases or TGF-Pl mRNA of 2.4 kb (Figure 1). Immunoprecipitation of proteins extracted from cell membranes with antibodies directed against either TGF-a or the C-terminal peptide of TGF-a precursors gave the same result. Both showed the presence of two TGF-a proteins of 16 and 26 kilodaltons, respectively (Figure 2). They correspond to the high-molecular-weight TGF-a precursors already described as membrane-anchored proforms.” Modulation of TGF-a and $1 expressions during cell aging. Figure 3 shows the proliferation curve of Capan 1 cells. Because these cells acquire more differentiated phenotypes with cell aging,12 it was possible to follow the expression of the TGFs during spontaneous differentiation. TGF-a mRNA levels were elevated shortly after seeding (on day 2) when the level of aminopeptidase N, the enzymatic marker of differentiation,14 was still low (5.0 f 0.5 mU/mg protein; n = 4). The messenger level decreased 2 days later (Figure 4A; day 4) during the increase in aminopeptidase N activity (10.0 rt 1.0 mU/mg protein; n = 4) and then remained constant during the final pe-
Figure 1. Northern blot analysis of TGF-a and -fll mRNAs in Capan 1 cells. Total RNA (20 pg) was electrophoresed on 1% agarose-6% formaldehyde gel, transferred onto nitrocellulose, and hybridized with “P-labeled human cDNA probes. Left, TGFfil mRNA, -2.4 kb. Right, TGF-a mRNA, -4.7 kb. Sizes were estimated using standards (internal ribosomal RNAs and commercial RNA ladder).
-
18.4
-
14.3
Figure 2. Immunoprecipitation of TGF-a. Cell membranes were immunoprecipitated with anti-TGFa antiserum (a) or preimmune serum (b) and analyzed by electrophoresis. Two bands of approximately 16 and 26 kilodaltons were visible. The positions of the molecular weight markers are shown at right.
riod of the exponential phase and at confluence (data not shown). The age-related decrease of TGF-a mRNA concentration was not dependent on either cell density or proliferation rate. Indeed, cells seeded at different concentrations (from 15,000 to 45,000 cells/cm2) expressed comparable high levels of TGF-a mRNA 2 days after seeding (Figure 5). Aminopeptidase N activity was also unmodified by cell density (data not shown). On the other hand, the role of growth rate was studied by seeding cells at high (45,000 cells/ cm2) or low (15,000 cells/cm2) density, which respectively increased and decreased cell proliferation. TGF-a mRNA concentration was only found to decrease when aminopeptidase N activity increased. Therefore, the expression of TGF-a mRNA did not appear to be related to cell growth but rather to cell differentiation. TGF-Pl mRNA concentrations were not modified during cell differentiation (Figure 4B). Modulation of TGF-a and $31 mRNA expressions by butyrate treatment. Sodium butyrate, a strong differentiating agent, caused an increased expression of more differentiated morphological and biochemical phenotypes’4 and, furthermore, arrested cell proliferation at a concentration of 2 mmol/L (Figure 3). A 3-day exposure to sodium butyrate was necessary to achieve its full differentiating effect, and aminopeptidase N activity increased to 15 f 0.9 mU/mg protein (n = 4). Cells were then cultured for 3 days in the presence of sodium butyrate, and the TGF-a mRNA level was measured on day 4. A 50% decrease in TGF-a mRNA concentration (Figure 4C) was found compared with that of control cells at the same age
1854
GASTROENTEROLOGYVol.103,No.6
ESTIVALET AL.
lyzed the modulation of the protein kinase C activity during cell differentiation. During cell aging. Shortly after seeding (24 hours), cells showed the highest level of protein kinase C activity in the particulate fraction (Figure 7), which represents the membrane-translocated enzyme involved in kinase activity.21-23 Increased cell aging led
100
days
1
I
1
2
I
3
1
4
I
5
I
6
A
I
50
7
Figure 3. Proliferation of Capan 1 cells in the absence (control, 0) or presence of PMA (50 nmol/L; ?? ) or butyrate (2 mmol/L; 0). Cells were seeded at 10,000 cells/cm’ and counted at the indicated times (n = 8).
(day 4). On the other hand, the level of mRNA in controls at day 4 had already decreased because of spontaneous differentiation (Figure 4.A). Therefore, after butyrate treatment, the final TGF-a mRNA level was only 25% that found on the second day of culture. By contrast, TGF-Pl mRNA concentration was not modified by sodium butyrate (Figure 4D). Modulation of TGF-a protein by butyrate treatment. Capan 1 was cultured for 3 days in the presence of sodium butyrate (2 mmol/L). Cells were then recovered and media were concentrated by lyophilization. Butyrate decreased the 16-kilodalton proTGF-a concentration by 50% and the 26-kilodalton proTGF-a concentration by 90% compared with untreated cells (Figure 6). In culture media, the secretory form of -6 kilodaltons was not found either in control or in butyrate-treated cells; only high-molecular-weight forms were detectable, at low levels and chiefly in controls. On the other hand, antibodies directed against intracellular and extracellular sequences of TGF-a precursors identified the same protein bands of 26 and 16 kilodaltons. These data suggest there is no growth factor secretion in controls or butyrate-treated cells. Thus, differentiation of Capan 1 cells, either spontaneous or induced by sodium butyrate, leads to a decrease of TGF-a expression. Protein Kinase C Activity During Cell Differentiation Because protein kinase C activity implicated in TGF-a mRNA expression,”
has been we ana-
1 a
0
L
L 2
4
days
100
C
D
a
B
50
0
c
But
Figure 4. Modulation of TGF-a and -fll mRNAs during spontaneous and butyrate-induced differentiation. Upper panel shows the modulation of TGF-a (A) and TGF-01 (B) mRNAs during cell aging (2 and 4 days after the seeding). Lower panel shows the levels of TGF-a (C) and TGF-pl (II) mRNAs after butyrate treatment (2 mmol/L) compared with those of controls of the same age (4 days after the seeding). C, control; But, butyrate-treated cells. Total RNA (2-6 pg for each sample) was analyzed by slotblot. Results are expressed as percent of control and are the means of three independent experiments (SD = 10%).
December
1992
I
TGF-a AND CELL DIFFERENTIATION
45 15 30 000 000 \
5
’
1
\ 375
:/ ’
.
\
\
‘\ \
Figure 5. Effect of cell density on TGF-a mRNA expression. Caand pan 1 cells were seeded at three densities (15,00&3O,OOO, 45,000 cells/cm*). mRNA levels were determined by slot-blot analysis the second day after seeding. The figure is representative of two independent experiments.
3
Figure 6. Effect of butyrate (2 mmol/L) and PMA (50 nmol/L) on TGF-a expression. After 3 days of treatment, cells were lysed and 100 pg of cell protein was immunoprecipitated with antiTGF-a antibodies and analyzed by Western blot. The figure is representative of three independent experiments. c, controls: but, butyrate-treated cells: pma, PMA-treated cells.
\
*
\
I
\ \
-\
‘\ *
1
to a progressive decrease of the particulate activity, and on day 4 the activity was 10% the value found at 24 hours. At the same time, cytosolic protein kinase C activity (which accounts for -90% of the total) was increased, but this change was less marked. The decrease of the particulate fraction suggests a reduced activation of protein kinase C during cell differentiation. During butyrate treatment. As for TGF-a experiments, cells were treated for 3 days with 2 mmol/L butyrate, and protein kinase C activity was determined on day 4. The enzyme level was globally decreased. As expected, its level in the particulate fraction was much more affected than that in the soluble one (50% and 20% decreases, respectively) (Figure 8). Because in untreated cells the enzyme level in the particulate fraction on the fourth day of culture was still lower than that on the second day, butyrate produced a stronger effect on the protein kinase C activity than that observed during spontaneous differentiation, as found for the TGF-a mRNA. Taken together, these results show that differentiation of Capan 1 cells decreased the levels of both
.
.-.
*
1855
2
3
---_--_y
4 days
Figure 7. Modulation of protein kinase C during cell aging. Cells were seeded at 10,000 cells/cm’, and at the indicated times protein kinase C activity was determined both in cytosolic (0) and particulate (*) fractions (n = 5 for each point; SD = ~10%).
TGF-a and membrane-bound ity.
protein kinase C activ-
Effects of PMA on Protein Kinase C Activation and TGF Expression It was of interest to analyze in this cell line whether activation of protein kinase C increases the expression of TGF-a mRNA and protein without affecting the expression of TGF-Pl mRNA. Eflect of PMA on cell proliferation. PMA slightly affected cell proliferation (Figure 3). A significant decrease in cell density was observed at PMA concentrations of >50 nmol/L; this was found to be a result of cell detachment (data not shown). Modification of protein kinase C levels by PMA. Cell treatment by PMA induced a rapid translocation of the enzyme (within 30 minutes) from cytosol to membranes (Figure 9A), as observed in many cell types. *5~22~23 This translocation-activation process was also accompanied by a decrease in overall protein kinase C activity; such a decrease has been reported to be a result of the down-regulation of the enzyme.22~23 Dose-response curves showed a maximal effect for PMA at -100 nmol/L (Figure 9B). A concentration of 50 nmol/L was used for long-term cell treatments because, as previously mentioned for cell proliferation, 100 nmol/L PMA was too high for long-term studies. After a 3-day treatment with PMA (Figure 8), the protein kinase C level in the particulate fraction was still high (260% of controls the same day of culture), and the level in the soluble fraction remained low (50% of the control value). High enzyme activity at the membrane level suggests that the rapid down-regulation of protein kinase C by
1656 ESTIVAL ET AL.
260
GASTROENTEROLOGY Vol. 103, No. 6
r
260
0
Cont.
m
PMA
Isi
But.
comparable low levels. Furthermore, immunoprecipitation did not detect specific protein bands possessing only the C-terminal sequence of TGF-a precursors. Thus, the cleavage of proTGF-a was probably not responsible for the low level of growth factor found in cells.
Discussion
:ytosol
-
100
-
50
We have studied the expression of TGF-a and -pl mRNAs in Capan 1, a human pancreatic cancer cell line. We have shown that these cells constitutively expressed both TGF messengers. Immunoprecipitation analysis of cell extracts revealed the presence of TGF-a peptides only as high-molecularweight transmembrane precursors. The low-molecular-weight form of -6 kilodaltons, corresponding to the cleavage product of these precursors by extracellular proteolytic enzymes,” has been found in culture media of some cell types (normal or tumoral)
membrane
Figure 8. Effect of butyrate or PMA treatment on protein kinase C activity. After a S-day treatment with butyrate (2 mmol/L; but.) or PMA (50 nmol/L), kinase levels were determined for both cytosolic and particulate (membrane) fractions (cont., control). Protein kinase C levels were expressed as a percent of that of untreated cells the same day of culture. Data represent the mean values of five independent experiments (SD was always 40%).
phorbol esters was not further increased during the S-day treatment. Effect ofPMA on TGF mRNA levels. After 3 days of phorbol ester treatment, when protein kinase C activity was still elevated in the particulate fraction, a 500% increase in TGF-a mRNA level was observed, but only a slight increase in that of TGF-Pl mRNA was found (Figure 10). This result suggests that PMA directly or indirectly enhanced the level of TGF-a mRNA. This effect was specific because the 4a-phorbol-12,13_didecanoate, a nonactivating phorbol ester, did not induce TGF-a mRNA expression (Figure 11). These data show that the activation of protein kinase C induced the expression of TGF-a mRNA without affecting that of TGF-Pl. Effect ofPMA on TGF-a protein expression. After a S-day treatment with 50 nmol/L PMA, Capan 1 cells expressed lower levels of proTGF-a compared with controls (Figure 6). Concentration of the 16-kilodalton form was found to have decreased by 30% and that of the 26-kilodalton form by 70%. Analysis of culture media showed only the presence of high-molecular-weight forms as for untreated cells, and at
A
15
30
60 Time (min)
I
b
1500
B
”
0.02
0.2 PMA (PM)
2
Figure 9. Modifications of protein kinase C levels in Capan 1 cells by PMA. (A) Kinetic studies of protein kinase C translocation by PMA (ZOOnmol/L). (I?) Dose-dependent effect of PMA on protein kinase C translocation. Cells were treated at the indicated doses for 30 minutes. The kinase activity was assayed in particulate (a) and cytosolic (0) fractions. The data shown represent the means of three independent experiments.
December
TGF-a AND CELL DIFFERENTIATION
1992
n
0
L
500 1
P 100
50
I TGF
ci
TGF /3
Figure. 10. Modulations of TGF-a and -gl mRNA expressions by PMA. Cells were treated for 3 days with PMA (60 nmol/L). Total RNA (2-6 pg for each sample) was analyzed by slot-blot analysis. Variations in mRNA contents are given as degree of change relative to controls. Cl, controls; phorbol ester. Values are means of three independent experiments.
secreting TGF-a.24 This low-molecular-weight form was never detected in concentrated serum-free media of Capan 1 cells; only low amounts of uncleaved high-molecular-weight TGF-a forms were found, probably resulting from lysis of some cells. Furthermore, we did not find protein bands specifically immunoprecipitated only by the antibody directed against the C-terminal sequence of proTGF-a, which remains in cells after TGF-a secretion.“*25 These data seem to exclude the disappearance of secreted TGFa by extensive proteolytic cleavage. The results obtained suggest that Capan 1 cells, as other cell types,24 do not secrete significant amounts of TGF-a. Capan 1 cells have been shown to acquire more differentiated phenotypes spontaneously during cell aging. Under these conditions, TGF-l31 mRNA levels were unmodified, whereas levels of TGF-a mRNA decreased by 50% and aminopeptidase N activity, the enzymatic marker of differentiation,14 increased. Sodium butyrate, a strong differentiating agent,“-14 was able to arrest cell proliferation at a concentration of 2 mmol/L. In the presence of 2 mmol/L sodium butyrate, TGF-fil mRNA levels remained constant; by contrast, aminopeptidase N activity increased 1.5 times and the level of TGF-a mRNA decreased two times compared with the values
1857
found during spontaneous differentiation. At the same time, lower amounts of proTGF-a were found in cell extracts. As in controls, no TGF-a secretion was detectable. These results indicate that differentiation of Capan 1 cells (spontaneous or induced) is accompanied by a decrease in TGF-a expression. The possibility that the decrease in TGF-a expression could be responsible for cell differentiation was not supported by the results obtained after treating cells with EGF or TGF-a, because these treatments were not found to exert any effect on spontaneous cell differentiation (data not shown). Reports showing that phorbol esters increased TGF-a mRNA expression in human keratinocytes” and in the human pancreatic cancer cell line T3M426 led us to analyze the activation state of protein kinase C during cell differentiation. The level of protein kinase C activity in the membrane fraction is thought to reflect the activation state of the enzyme.21-23 In both differentiating conditions tested, membrane-bound protein kinase C activity had decreased. For instance, its level was reduced by 50% at day 4 compared with day 2, and a stronger decrease occurred during butyrate treatment. Taken together, these data show that during differentiation Capan 1 cells present lower levels of TGF-a mRNA and protein as well as a decreased activation state of protein kinase C. We then assessed whether Capan 1 cells, as other cell lines,“~‘” responded to PMA by increasing TGF-a mRNA levels as a result of the protein kinase C activation Treatment of Capan 1 cells with the active form of phorbol ester induced the translocation of protein kinase C to membranes; the level of membrane-bound enzyme activity increased 2.6-fold and that of the TGF-a mRNA fivefold. That seems to reflect the involvement of this enzyme in the regula-
cont. PMA 4aPDD 1 Figure 11. Specificity of phorbol esters for TGF-a mRNA induction. Total RNA (6 Fg) from control cells (cont.) and from cells treated with 50 nmol/L PMA or with 60 nmol/L 4n-phorbol12,1%didecanoate (4aPDD, an inactive phorbol ester) were blotted onto nitrocellulose and hybridized with the human cDNA TGF-a probe. The figure shows a representative autoradiograph of slot-blot experiments.
1858
ESTIVAL ET AL.
tion of TGF-a mRNA as described in the other cell lines analyzed. However, cell content in high-molecular-weight TGF-a had not increased. Furthermore, secretion of the 6-kilodalton TGF-a was not evident. Indeed, this form was undetectable in culture media, and the antibodies used directed against either the intracellular or the extracellular proTGF-a sequences immunoprecipitated the same cell proteins. The nonparallelism between messenger level and TGF-(r synthesis has been observed in many cell types’ and also in human pancreatic cancer cell lines8 The mechanism underlying this nonparallelism is not yet clearly established and suggests the existence of translational controls of TGF-a biosynthesis. Because in the course of this study TGF-Bl mRNA concentration remained at a constant level whatever the protein kinase C activity, different regulatory mechanisms seem to take place in TGF-a and -Bl mRNA expressions in this pancreatic cancer cell line. Our results show for the first time that a pancreatic cancer cell line expresses lower levels of TGF-a mRNA and protein during differentiation. TGF-a was not found to be secreted by Capan 1 cells in the experimental conditions used. However, it has been shown that the membrane-anchored high-molecular-weight TGF-a binds to EGF receptors on adjacent cells and leads to signal transduction as the low-molecular-weight formSz4 Because Capan 1 cells express EGF receptors and are growth stimulated by EGF,27 the decrease in proTGF-a expression could lead to a decrease in the autocrine control of growth via EGF receptors. The involvement of protein kinase C is suggested by a decrease of protein kinase activity during differentiation and, on the contrary, an increase of TGF-a mRNA levels under PMA stimulation; however, that relationship still needs to be confirmed. It will be of great interest to extend the findings observed in Capan 1 to other pancreatic cell lines to better define the influence of cell differentiation on the autocrine control of human pancreatic cancer cell growth by TGF-a. References Todaro GJ, Lee DC, Webb NR, Rose TM, Brown JP. Rat type a transforming growth factor: structure and possible function as a membrane receptor. In: Feramisco J, Ozanne B, Stiles C (eds). Growth factors and transformation. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1985:51-58. Twardzik DR. Differential expression of transforming growth factor-a during prenatal development of the mouse. Cancer Res 1985;45:5413-5416. Lee DC, Rose TM, Webb NR, Todaro GJ. Cloning and sequence analysis of a cDNA for rat transforming growth factora. Nature 1985;313:489-491. Kudlow JE, Leung AWC, Kobrin MS, Paterson AJ, Asa SL. Transforming growth factor-a in the mammalian brain. J Biol Chem 1989:264:3880-3883.
GASTROENTEROLOGY Vol. 103, No. 6
5. Tam JP, Sheikh MA, Solomon DS, Ossowski L. Efficient synthesis of human type a transforming growth factor: its physical and biological characterization. Proc Nat1 Acad Sci USA 1986;83:8082-8086. 6. Tam JP. Physiological effects of transforming growth factor in the newborn mouse. Science 1985;229:673-675. 7. Heldin CH, Westermark B. Growth factors as transforming proteins. Eur J Biochem 1989;184:487-496. 8. Smith JJ, Derynck R, Korc M. Production of transforming growth factor-a in human pancreatic cancer cells: evidence for a superagonist autocrine cycle. Proc Nat1 Acad Sci USA 1987;84:7567-7570. 9. Derynck R, Goeddel DV, Ullrich A, Gutterman JU, Williams RD, Bringman TS, Berger WH. Synthesis of messenger for transforming growth factor a and p and the epidermal growth factor receptor by human tumors. Cancer Res 1987;47:707712. 10. Sporn MB, Roberts AB, Wakefield LM, de Combrugghe B. Some recent advances in the chemistry and biology of transforming growth factor-beta. J Cell Biol 1987;105:1039-1045. 11. Pittelkow MR, Lindquist PB, Abraham RT, Graves-Deal R, Derynck R, Coffey RJ Jr. Induction of transforming growth factor-a expression in human keratinocytes by phorbol esters. J Biol Chem 1989;264:5164-5171. 12. Levrat JH, Palevody C, Ratovo DG, Hollande E. Differentiation of the human pancreatic adenocarcinoma cell line (CAPAN1)in culture and co-culture with fibroblasts, dome formation. lnt J Cancer 1984;34:177-185. 13. Bloom EJ, Siddiqui B, Hicks JW, Kim YS. Effect of sodium butyrate, a differentiating agent, on cell surface glycoconjugates of human pancreatic cell lines. Pancreas 1989;4:59-64. 14. Bensaadi N, Clemente F, Vaysse N. Modulation of enzymatic activities during spontaneous and induced differentiation in a human pancreatic adenocarcinoma cell line Capan 1. lnt J Pancreatol 1989;4:391-406. 15. Anderson WB, Estival A, Tapiovaara H, Gopala Krishna R. Altered subcellular distribution of protein kinase C (phorbol ester receptor). Possible role in tumor promotion and the regulation of cell growth. Advances in cyclic nucleotide and protein phosphorylation research. 1985;19:287-306. 16. Chirgwin JM, Przybyla AE, MacDonald RJ, Rutter WJ. Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 1979;18:5294-5299. 17. Thomas PS. Hybridization of denatured RNA and small DNA fragments transferred to nitrocellulose. Proc Nat1 Acad Sci USA 1980;77:5201-5205. 18. Bringman TS, Lindquist PB, Derynck R. Different transforming growth factor-a species are derived from a glycosylated and palmitoylated transmembrane precursor. Cell 1987;48: 429-440. 19. Laemli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970;227:680685. 20. Lowry OH, Rosebrough NJ, Farr AL, Randall NJ. Protein measurement with the folin phenol reagent. J Biol Chem 1951;193:265-275. 21. Nishizuka Y. The molecular heterogeneity of protein kinase C and its implication for cellular regulation. Nature 1988;334: 661-665. 22. Shoji M, Girard PR, Charp PA, Koeffler HP, Vogler WR, Kuo JF. Effects of phorbol ester on translocation and down regulation of protein kinase C and phosphorylation of endogenous proteins in human acute myeloid leukemia cell line KG-l and its phorbol ester-resistent subline GF-la. Cancer Res 1987;47:6363-6370, 23. Bornes C, Eppenberg V, Wyss R, Fabbro D. Continuous synthesis of two protein kinase C-related proteins after down regula-
December
1992
tion by phorbol esters. Proc Nat1 Acad Sci USA 1988;85:21102114. 24.
Massague J. Transforming growth factor-u. A model for membrane-anchored growth factors. J Biol Chem 1990;265:2139321396. 25. Pandiella A, Massague J. Cleavage of the membrane precursor for transforming growth factor a is a regulated process. Proc Nat1 Acad Sci USA 1991;88:1726-1730, 26. Glinsmann-Gibson BJ, Korc M. Regulation of transforming growth factor-u mRNA expression in T3M4 human pancreatic carcinoma cells. Pancreas 1991;6:142-149. 27. Oustry P, Estival A, Pradayrol L, Vaysse N, Clemente F. Two
TGF-a AND CELL DIFFERENTIATION
1859
subclasses of EGF receptors in the human pancreatic cancer cell lines Capan 1 and MIA PaCa 2. Int J Pancreatol 1990:6:119-128.
Received December 26, 1990. Accepted June 25,1992. Address requests for reprints to: Francois Clemente, M.D., INSERM U 151, Centre Hospitalier Universitaire Rangueil Bat L 3, 31054 Toulouse Cedex, France. Supported by grants from the Federation Nationale des Centres de Lutte contre le Cancer (FNCLCC), ARC (6165), and the National Institutes of Health (CA36544) (to J.P.T.).
LIVER, PANCREAS,
1992;103:1851-1859
AND BILIARY TRACT
Decreased Expression of Transforming Growth Factor a During Differentiation of Human Pancreatic Cancer Cells AGNES ESTIVAL, PASCAL and FRANCOIS CLEMENTE
CLERC,
NICOLE
INSERM U 151, Institut L. Bugnard, Centre Hospitalier Rockefeller University, New York, New York
The relationship between cell differentiation and transforming growth factor 01(TGF-a) expression in human pancreatic cancer cells was analyzed in Capan 1 cells. These cells differentiate either spontaneously or after butyrate treatment. During differentiation (spontaneous or butyrate induced), TGF-a messenger RNA (mRNA) levels decreased, whereas the TGF-fil mRNA levels remained unchanged. TGF-a was present in cells as proTGF-a, which decreased after butyrate treatment. Secretion of TGFa was not found. Under the two conditions of differentiation, the membrane-bound protein kinase C activity was also reduced. Conversely, long-term phorbol ester treatment increased both membranebound protein kinase C activity (260%) and TGF-a mRNA level (500%), a not significant increase of TGF-j31 mRNA was observed. However, phorbol l&myristate-13-acetate did not induce TGF-a synthesis or secretion. These data suggest that expression of TGF-a can be reduced in cancer cells; they also suggest the existence of a relationship between TGF-a expression and cell differentiation. In addition, the protein kinase C-induced TGF-a mRNA level was not followed by the increase of TGF-a biosynthesis, suggesting a translational control. Finally, the expression of TGF-a and -plmessengers appears to be differently regulated. ransforming growth factor (Y (TGF-a) was initially purified from transformed cells.’ Later on, TGF-a messenger RNA (mRNA) was found in various tissues during embryonic development’ and at low levels in some normal adult cell types.3,4 It is generally accepted that its biological effects are mediated through epidermal growth factor (EGF) receptors.’ TGF-a promotes cell proliferation in vivo and in vitro, and its involvement in the autocrine control of cancer cell growth has been strongly suggested in several reports.68
T
VAYSSE,
Universitaire
JAMES
Rangueil, Toulouse,
P. TAM, France; and
It should be emphasized that TGF-a is not a product of any oncogene but rather is synthesized by tumor cells following the activation of oncogenes.’ On the other hand, in normal cells, the expression of TGF-a appears to be somehow related to their differentiation state. Indeed, it is highly expressed during normal embryonic development when cells are poorly differentiated, whereas it is weakly represented in normal, fully differentiated cells. Thus, the level of TGF-a expression by cancer cells could be related, to a certain degree, to their differentiation state and be reduced by cell differentiation. The possibility of decreasing TGF-a expression in human pancreatic cancer cells and the mechanisms responsible for that decrease are of great interest for their therapeutic implications. To test the hypothesis of a differentiation-related TGF-a expression in pancreatic cancer cells, we followed the expression of TGF-a during cell differentiation. We analyzed TGF-a and -fll mRNAs in parallel because the TGF-Pl is expressed by different cell types whether they are differentiated or not.‘*” Because it has recently been shown that TGF-a is increased by protein kinase C activation,” we also analyzed the modulation of this enzyme during cell differentiation. The human pancreatic cancer cell line Capan 1 was chosen for this study because of its interesting properties. It has been shown that Capan 1 cells spontaneously acquire more differentiated phenotypes during cell agingl’ and also under the action of sodium butyrate.13p14 Differentiation is evidenced morphologically by cell polarization and by apical junction complexes, and biochemically by increased alkaline phosphatase and aminopeptidase N levels.14 Our results show that during differentiation of Capan 1 cells, both TGF-a expression and activation of protein kinase C decrease. 0
1992 by the American Gastroenterological 0016-5085/92/$3.00
Association
1852
ESTIVAL ET AL.
GASTROENTEROLOGY Vol. 103, No. 6
Materials and Methods Materials RPM1 1640 and fetal calf serum (FCS) were obtained from Gibco (Grand Island, NY). Phorbol 12-myristate-13-acetate (PMA), 4a-phorbol 12,13didecanoate, histone Hl, phenylmethylsulfonyl fluoride (PMSF), and sodium butyrate were purchased from Sigma (St. Louis, MO) and the ion exchanger DE52 cellulose from Whatman (Maidstone, England). The Capan 1 cell line was kindly provided by Dr. J. Fogh (Sloan-Kettering Institute for Cancer Research, New York, NY). [y-32P]adenosine triphosphate ([y-32P]ATP) and [35S]cysteine were obtained from Amersham (Les Ulis, France). Cell Culture
and Cell Treatments
The Capan 1 cells were plated in 180-cm2 flasks at a density of lo* cells/cm’ and grown in RPM1 1640 supplemented with 10% FCS. After a 16-hour attachment period, media were replaced with fresh medium and cells were treated for 3 days with PMA (50 nmol/L) or 4a-phorbol12,13-didecanoate (50 nmol/L) or sodium butyrate (2 mmol/L). The different determinations were performed the fourth day after seeding; the day before, media were replaced with fresh medium containing the drugs. For the proliferation studies, the Capan 1 cells were grown in the presence of the drugs, and the medium containing the drugs was replaced every other day. Cells were counted after trypsinization using a Coulter Counter (model ZM; Coulter Electronics, Margency, France). Protein
Kinase
C Assay
Protein kinase C activity was assayed as previously describedI by measuring the incorporation of 32P from [y32PI ATP into histone Hl. A control assay was performed in the absence of diacylglycerol in the assay tube. The soluble and particulate cell fractions containing the protein kinase C were obtained after a 100,OOOg centrifugation then chromatographied on ion-exchanger DE52 cellulose and recovered in the chromatographic fraction eluted by 50 mmol/L NaCl before any enzyme activity assay.15 RNA Extraction
and Messenger
Quantification
Total RNA was extracted according to the method of Chirgwin et a1.16Twenty micrograms of total RNA was electrophoretically fractionated on 1% agarose gel in denaturing conditions and then transferred onto nitrocellulose filters.17 The concentration of individual mRNA samples was determined by slot-blot hybridization analysis. Complementary DNA (cDNA) probes were 32P-labeled (sp act, lo7 cpm/ug) by nick-translation (Amersham kit). Hybridizations were carried out in a mixture of 50% formamide, 6X SSC, 0.5% sodium dodecyl sulfate (SDS), 5X Denhardt’s, 10 mmol/L ethylenediamine tetraacetic acid (EDTA), and 100 ug/mL salmon sperm DNA at 42°C for 16 hours. Filters were washed in 2X SSC and 0.1% SDS one time for 20 minutes at room temperature and two times for 20 minutes in 1X SSC and 0.1% SDS at 45“C. Filters were air-dried and then exposed to Kodak X-Omat AR films
(Rochester, New York). Spots were quantified using an LKB laser densitometer (Pharmacia, France). Human-specific probes for TGF-a and -pl were kindly provided by Dr R. Derynck (Genentech, San Francisco, CA). The TGF-a cDNA probe contained the complete coding sequence, and the TGF-Pl cDNA was a 1050-base pair-long probee Immunoprecipitation Electrophoresis
of TGF-a
and
Gel
The immunoprecipitation was performed according to the method of Bringman et al.“’ Two polyclonal antibodies were used. The first was raised against human TGF-a peptide and the second against the intracellular Cterminal sequence of human TGF-a precursors.5 TGF-a was characterized in cells by incorporation studies. Cells were incubated for 16 hours in serum-free culture media containing 100 uCi/mL of [35S]cysteine then washed three times with phosphate-buffered saline (PBS) and harvested by scraping in PBS containing 1% NP-40,10 mmol/L EDTA, and 1 mmol/L PMSF at 4’C. Samples were diluted with 50 mmol/L Tris-HCl, 0.15 mol/L NaCl, 2 mmol/L EDTA, 1% NP-40, 0.1% SDS, 0.1% sodium azide, and 1 mg/mL bovine serum albumin at pH 8.0 and then incubated in parallel with 10 uL of purified polyclonal antibodies for 16 hours at 4°C. The protein A-sepharose suspension (Pharmacia LKB, Uppsala, Sweden) was then added and the mixture incubated for 2 hours at 25°C under rotary agitation. After five washings, beads were boiled at 100°C for 5 minutes in 100 uL of 2X concentrated electrophoresis sample buffer.lg Beads were eliminated by centrifugation and samples analyzed by electrophoresis according to the method of Laemmlilg in denaturing 15% polyacrylamide gels, which were dried and exposed to Kodak X-Omat AR films. TGF-a produced by cells cultured in the presence of butyrate or PMA was analyzed and quantified by Western blot. Cells were kept in serum-free media 24 hours before their recovery. Cells were washed and lysed as for incorporation studies. Culture media were clarified by centrifugation, acidified by acetic acid to a final concentration of 0.1 mol/L, dialyzed in Spectrapore tubing (3000-kilodalton cutoff) against 100 volumes of 0.1 mol/L acetic acid and 0.1 mmol/L PMSF at 4”C, and lyophilized.” Samples of cells and culture media were immunoprecipitated as described above. Immunoprecipitates were subjected to electrophoresis and electroblotted onto nitrocellulose membranes. Membranes were preincubated for 1 hour at 25’C in PBS containing 5% nonfat dry milk and 0.5% Tween 20, then incubated overnight with anti-TGF-a or anti-c-terminal sequence of proTGF-a, diluted in PBS containing 0.5% Tween 20. Membranes were then rinsed three times with PBS containing 0.5% Tween 20 and incubated for 1 hour at 25°C with anti-rabbit antibodies coupled to peroxydase. After three more washes, protein bands were detected by chemiluminescence on autoradiography films (ECL, Amersham) and quantified by laser densitometry. Control experiments were carried out in parallel by using rabbit nonimmune sera for immunoprecipitations. Human TGF-a of 5.6 kilodaltons was also used as a control during electrophoresis and immunoblotting.
December
TGF-a
1992
Aminopeptidase
N Assay
a
AND CELL DIFFERENTIATION
1853
b
Aminopeptidase N activity was determined as previously described14 using L-alanine-p-nitroanilide as a substrate at a concentration of 1.5 mmol/L. Proteins Protein levels were measured according to the method of Lowry et al.” Results Expression
25.7
of TGFs by Capan 1 Cells
TGF mRNAs and protein. Total RNA was electrophoretically fractionated in agarose gel (1%) in denaturing conditions and analyzed by blot hybridization. Northern blots analysis revealed the existence of a single band corresponding to TGF-a mRNA of 4.7 kilobases or TGF-Pl mRNA of 2.4 kb (Figure 1). Immunoprecipitation of proteins extracted from cell membranes with antibodies directed against either TGF-a or the C-terminal peptide of TGF-a precursors gave the same result. Both showed the presence of two TGF-a proteins of 16 and 26 kilodaltons, respectively (Figure 2). They correspond to the high-molecular-weight TGF-a precursors already described as membrane-anchored proforms.” Modulation of TGF-a and $1 expressions during cell aging. Figure 3 shows the proliferation curve of Capan 1 cells. Because these cells acquire more differentiated phenotypes with cell aging,12 it was possible to follow the expression of the TGFs during spontaneous differentiation. TGF-a mRNA levels were elevated shortly after seeding (on day 2) when the level of aminopeptidase N, the enzymatic marker of differentiation,14 was still low (5.0 f 0.5 mU/mg protein; n = 4). The messenger level decreased 2 days later (Figure 4A; day 4) during the increase in aminopeptidase N activity (10.0 rt 1.0 mU/mg protein; n = 4) and then remained constant during the final pe-
Figure 1. Northern blot analysis of TGF-a and -fll mRNAs in Capan 1 cells. Total RNA (20 pg) was electrophoresed on 1% agarose-6% formaldehyde gel, transferred onto nitrocellulose, and hybridized with “P-labeled human cDNA probes. Left, TGFfil mRNA, -2.4 kb. Right, TGF-a mRNA, -4.7 kb. Sizes were estimated using standards (internal ribosomal RNAs and commercial RNA ladder).
-
18.4
-
14.3
Figure 2. Immunoprecipitation of TGF-a. Cell membranes were immunoprecipitated with anti-TGFa antiserum (a) or preimmune serum (b) and analyzed by electrophoresis. Two bands of approximately 16 and 26 kilodaltons were visible. The positions of the molecular weight markers are shown at right.
riod of the exponential phase and at confluence (data not shown). The age-related decrease of TGF-a mRNA concentration was not dependent on either cell density or proliferation rate. Indeed, cells seeded at different concentrations (from 15,000 to 45,000 cells/cm2) expressed comparable high levels of TGF-a mRNA 2 days after seeding (Figure 5). Aminopeptidase N activity was also unmodified by cell density (data not shown). On the other hand, the role of growth rate was studied by seeding cells at high (45,000 cells/ cm2) or low (15,000 cells/cm2) density, which respectively increased and decreased cell proliferation. TGF-a mRNA concentration was only found to decrease when aminopeptidase N activity increased. Therefore, the expression of TGF-a mRNA did not appear to be related to cell growth but rather to cell differentiation. TGF-Pl mRNA concentrations were not modified during cell differentiation (Figure 4B). Modulation of TGF-a and $31 mRNA expressions by butyrate treatment. Sodium butyrate, a strong differentiating agent, caused an increased expression of more differentiated morphological and biochemical phenotypes’4 and, furthermore, arrested cell proliferation at a concentration of 2 mmol/L (Figure 3). A 3-day exposure to sodium butyrate was necessary to achieve its full differentiating effect, and aminopeptidase N activity increased to 15 f 0.9 mU/mg protein (n = 4). Cells were then cultured for 3 days in the presence of sodium butyrate, and the TGF-a mRNA level was measured on day 4. A 50% decrease in TGF-a mRNA concentration (Figure 4C) was found compared with that of control cells at the same age
1854
GASTROENTEROLOGYVol.103,No.6
ESTIVALET AL.
lyzed the modulation of the protein kinase C activity during cell differentiation. During cell aging. Shortly after seeding (24 hours), cells showed the highest level of protein kinase C activity in the particulate fraction (Figure 7), which represents the membrane-translocated enzyme involved in kinase activity.21-23 Increased cell aging led
100
days
1
I
1
2
I
3
1
4
I
5
I
6
A
I
50
7
Figure 3. Proliferation of Capan 1 cells in the absence (control, 0) or presence of PMA (50 nmol/L; ?? ) or butyrate (2 mmol/L; 0). Cells were seeded at 10,000 cells/cm’ and counted at the indicated times (n = 8).
(day 4). On the other hand, the level of mRNA in controls at day 4 had already decreased because of spontaneous differentiation (Figure 4.A). Therefore, after butyrate treatment, the final TGF-a mRNA level was only 25% that found on the second day of culture. By contrast, TGF-Pl mRNA concentration was not modified by sodium butyrate (Figure 4D). Modulation of TGF-a protein by butyrate treatment. Capan 1 was cultured for 3 days in the presence of sodium butyrate (2 mmol/L). Cells were then recovered and media were concentrated by lyophilization. Butyrate decreased the 16-kilodalton proTGF-a concentration by 50% and the 26-kilodalton proTGF-a concentration by 90% compared with untreated cells (Figure 6). In culture media, the secretory form of -6 kilodaltons was not found either in control or in butyrate-treated cells; only high-molecular-weight forms were detectable, at low levels and chiefly in controls. On the other hand, antibodies directed against intracellular and extracellular sequences of TGF-a precursors identified the same protein bands of 26 and 16 kilodaltons. These data suggest there is no growth factor secretion in controls or butyrate-treated cells. Thus, differentiation of Capan 1 cells, either spontaneous or induced by sodium butyrate, leads to a decrease of TGF-a expression. Protein Kinase C Activity During Cell Differentiation Because protein kinase C activity implicated in TGF-a mRNA expression,”
has been we ana-
1 a
0
L
L 2
4
days
100
C
D
a
B
50
0
c
But
Figure 4. Modulation of TGF-a and -fll mRNAs during spontaneous and butyrate-induced differentiation. Upper panel shows the modulation of TGF-a (A) and TGF-01 (B) mRNAs during cell aging (2 and 4 days after the seeding). Lower panel shows the levels of TGF-a (C) and TGF-pl (II) mRNAs after butyrate treatment (2 mmol/L) compared with those of controls of the same age (4 days after the seeding). C, control; But, butyrate-treated cells. Total RNA (2-6 pg for each sample) was analyzed by slotblot. Results are expressed as percent of control and are the means of three independent experiments (SD = 10%).
December
1992
I
TGF-a AND CELL DIFFERENTIATION
45 15 30 000 000 \
5
’
1
\ 375
:/ ’
.
\
\
‘\ \
Figure 5. Effect of cell density on TGF-a mRNA expression. Caand pan 1 cells were seeded at three densities (15,00&3O,OOO, 45,000 cells/cm*). mRNA levels were determined by slot-blot analysis the second day after seeding. The figure is representative of two independent experiments.
3
Figure 6. Effect of butyrate (2 mmol/L) and PMA (50 nmol/L) on TGF-a expression. After 3 days of treatment, cells were lysed and 100 pg of cell protein was immunoprecipitated with antiTGF-a antibodies and analyzed by Western blot. The figure is representative of three independent experiments. c, controls: but, butyrate-treated cells: pma, PMA-treated cells.
\
*
\
I
\ \
-\
‘\ *
1
to a progressive decrease of the particulate activity, and on day 4 the activity was 10% the value found at 24 hours. At the same time, cytosolic protein kinase C activity (which accounts for -90% of the total) was increased, but this change was less marked. The decrease of the particulate fraction suggests a reduced activation of protein kinase C during cell differentiation. During butyrate treatment. As for TGF-a experiments, cells were treated for 3 days with 2 mmol/L butyrate, and protein kinase C activity was determined on day 4. The enzyme level was globally decreased. As expected, its level in the particulate fraction was much more affected than that in the soluble one (50% and 20% decreases, respectively) (Figure 8). Because in untreated cells the enzyme level in the particulate fraction on the fourth day of culture was still lower than that on the second day, butyrate produced a stronger effect on the protein kinase C activity than that observed during spontaneous differentiation, as found for the TGF-a mRNA. Taken together, these results show that differentiation of Capan 1 cells decreased the levels of both
.
.-.
*
1855
2
3
---_--_y
4 days
Figure 7. Modulation of protein kinase C during cell aging. Cells were seeded at 10,000 cells/cm’, and at the indicated times protein kinase C activity was determined both in cytosolic (0) and particulate (*) fractions (n = 5 for each point; SD = ~10%).
TGF-a and membrane-bound ity.
protein kinase C activ-
Effects of PMA on Protein Kinase C Activation and TGF Expression It was of interest to analyze in this cell line whether activation of protein kinase C increases the expression of TGF-a mRNA and protein without affecting the expression of TGF-Pl mRNA. Eflect of PMA on cell proliferation. PMA slightly affected cell proliferation (Figure 3). A significant decrease in cell density was observed at PMA concentrations of >50 nmol/L; this was found to be a result of cell detachment (data not shown). Modification of protein kinase C levels by PMA. Cell treatment by PMA induced a rapid translocation of the enzyme (within 30 minutes) from cytosol to membranes (Figure 9A), as observed in many cell types. *5~22~23 This translocation-activation process was also accompanied by a decrease in overall protein kinase C activity; such a decrease has been reported to be a result of the down-regulation of the enzyme.22~23 Dose-response curves showed a maximal effect for PMA at -100 nmol/L (Figure 9B). A concentration of 50 nmol/L was used for long-term cell treatments because, as previously mentioned for cell proliferation, 100 nmol/L PMA was too high for long-term studies. After a 3-day treatment with PMA (Figure 8), the protein kinase C level in the particulate fraction was still high (260% of controls the same day of culture), and the level in the soluble fraction remained low (50% of the control value). High enzyme activity at the membrane level suggests that the rapid down-regulation of protein kinase C by
1656 ESTIVAL ET AL.
260
GASTROENTEROLOGY Vol. 103, No. 6
r
260
0
Cont.
m
PMA
Isi
But.
comparable low levels. Furthermore, immunoprecipitation did not detect specific protein bands possessing only the C-terminal sequence of TGF-a precursors. Thus, the cleavage of proTGF-a was probably not responsible for the low level of growth factor found in cells.
Discussion
:ytosol
-
100
-
50
We have studied the expression of TGF-a and -pl mRNAs in Capan 1, a human pancreatic cancer cell line. We have shown that these cells constitutively expressed both TGF messengers. Immunoprecipitation analysis of cell extracts revealed the presence of TGF-a peptides only as high-molecularweight transmembrane precursors. The low-molecular-weight form of -6 kilodaltons, corresponding to the cleavage product of these precursors by extracellular proteolytic enzymes,” has been found in culture media of some cell types (normal or tumoral)
membrane
Figure 8. Effect of butyrate or PMA treatment on protein kinase C activity. After a S-day treatment with butyrate (2 mmol/L; but.) or PMA (50 nmol/L), kinase levels were determined for both cytosolic and particulate (membrane) fractions (cont., control). Protein kinase C levels were expressed as a percent of that of untreated cells the same day of culture. Data represent the mean values of five independent experiments (SD was always 40%).
phorbol esters was not further increased during the S-day treatment. Effect ofPMA on TGF mRNA levels. After 3 days of phorbol ester treatment, when protein kinase C activity was still elevated in the particulate fraction, a 500% increase in TGF-a mRNA level was observed, but only a slight increase in that of TGF-Pl mRNA was found (Figure 10). This result suggests that PMA directly or indirectly enhanced the level of TGF-a mRNA. This effect was specific because the 4a-phorbol-12,13_didecanoate, a nonactivating phorbol ester, did not induce TGF-a mRNA expression (Figure 11). These data show that the activation of protein kinase C induced the expression of TGF-a mRNA without affecting that of TGF-Pl. Effect ofPMA on TGF-a protein expression. After a S-day treatment with 50 nmol/L PMA, Capan 1 cells expressed lower levels of proTGF-a compared with controls (Figure 6). Concentration of the 16-kilodalton form was found to have decreased by 30% and that of the 26-kilodalton form by 70%. Analysis of culture media showed only the presence of high-molecular-weight forms as for untreated cells, and at
A
15
30
60 Time (min)
I
b
1500
B
”
0.02
0.2 PMA (PM)
2
Figure 9. Modifications of protein kinase C levels in Capan 1 cells by PMA. (A) Kinetic studies of protein kinase C translocation by PMA (ZOOnmol/L). (I?) Dose-dependent effect of PMA on protein kinase C translocation. Cells were treated at the indicated doses for 30 minutes. The kinase activity was assayed in particulate (a) and cytosolic (0) fractions. The data shown represent the means of three independent experiments.
December
TGF-a AND CELL DIFFERENTIATION
1992
n
0
L
500 1
P 100
50
I TGF
ci
TGF /3
Figure. 10. Modulations of TGF-a and -gl mRNA expressions by PMA. Cells were treated for 3 days with PMA (60 nmol/L). Total RNA (2-6 pg for each sample) was analyzed by slot-blot analysis. Variations in mRNA contents are given as degree of change relative to controls. Cl, controls; phorbol ester. Values are means of three independent experiments.
secreting TGF-a.24 This low-molecular-weight form was never detected in concentrated serum-free media of Capan 1 cells; only low amounts of uncleaved high-molecular-weight TGF-a forms were found, probably resulting from lysis of some cells. Furthermore, we did not find protein bands specifically immunoprecipitated only by the antibody directed against the C-terminal sequence of proTGF-a, which remains in cells after TGF-a secretion.“*25 These data seem to exclude the disappearance of secreted TGFa by extensive proteolytic cleavage. The results obtained suggest that Capan 1 cells, as other cell types,24 do not secrete significant amounts of TGF-a. Capan 1 cells have been shown to acquire more differentiated phenotypes spontaneously during cell aging. Under these conditions, TGF-l31 mRNA levels were unmodified, whereas levels of TGF-a mRNA decreased by 50% and aminopeptidase N activity, the enzymatic marker of differentiation,14 increased. Sodium butyrate, a strong differentiating agent,“-14 was able to arrest cell proliferation at a concentration of 2 mmol/L. In the presence of 2 mmol/L sodium butyrate, TGF-fil mRNA levels remained constant; by contrast, aminopeptidase N activity increased 1.5 times and the level of TGF-a mRNA decreased two times compared with the values
1857
found during spontaneous differentiation. At the same time, lower amounts of proTGF-a were found in cell extracts. As in controls, no TGF-a secretion was detectable. These results indicate that differentiation of Capan 1 cells (spontaneous or induced) is accompanied by a decrease in TGF-a expression. The possibility that the decrease in TGF-a expression could be responsible for cell differentiation was not supported by the results obtained after treating cells with EGF or TGF-a, because these treatments were not found to exert any effect on spontaneous cell differentiation (data not shown). Reports showing that phorbol esters increased TGF-a mRNA expression in human keratinocytes” and in the human pancreatic cancer cell line T3M426 led us to analyze the activation state of protein kinase C during cell differentiation. The level of protein kinase C activity in the membrane fraction is thought to reflect the activation state of the enzyme.21-23 In both differentiating conditions tested, membrane-bound protein kinase C activity had decreased. For instance, its level was reduced by 50% at day 4 compared with day 2, and a stronger decrease occurred during butyrate treatment. Taken together, these data show that during differentiation Capan 1 cells present lower levels of TGF-a mRNA and protein as well as a decreased activation state of protein kinase C. We then assessed whether Capan 1 cells, as other cell lines,“~‘” responded to PMA by increasing TGF-a mRNA levels as a result of the protein kinase C activation Treatment of Capan 1 cells with the active form of phorbol ester induced the translocation of protein kinase C to membranes; the level of membrane-bound enzyme activity increased 2.6-fold and that of the TGF-a mRNA fivefold. That seems to reflect the involvement of this enzyme in the regula-
cont. PMA 4aPDD 1 Figure 11. Specificity of phorbol esters for TGF-a mRNA induction. Total RNA (6 Fg) from control cells (cont.) and from cells treated with 50 nmol/L PMA or with 60 nmol/L 4n-phorbol12,1%didecanoate (4aPDD, an inactive phorbol ester) were blotted onto nitrocellulose and hybridized with the human cDNA TGF-a probe. The figure shows a representative autoradiograph of slot-blot experiments.
1858
ESTIVAL ET AL.
tion of TGF-a mRNA as described in the other cell lines analyzed. However, cell content in high-molecular-weight TGF-a had not increased. Furthermore, secretion of the 6-kilodalton TGF-a was not evident. Indeed, this form was undetectable in culture media, and the antibodies used directed against either the intracellular or the extracellular proTGF-a sequences immunoprecipitated the same cell proteins. The nonparallelism between messenger level and TGF-(r synthesis has been observed in many cell types’ and also in human pancreatic cancer cell lines8 The mechanism underlying this nonparallelism is not yet clearly established and suggests the existence of translational controls of TGF-a biosynthesis. Because in the course of this study TGF-Bl mRNA concentration remained at a constant level whatever the protein kinase C activity, different regulatory mechanisms seem to take place in TGF-a and -Bl mRNA expressions in this pancreatic cancer cell line. Our results show for the first time that a pancreatic cancer cell line expresses lower levels of TGF-a mRNA and protein during differentiation. TGF-a was not found to be secreted by Capan 1 cells in the experimental conditions used. However, it has been shown that the membrane-anchored high-molecular-weight TGF-a binds to EGF receptors on adjacent cells and leads to signal transduction as the low-molecular-weight formSz4 Because Capan 1 cells express EGF receptors and are growth stimulated by EGF,27 the decrease in proTGF-a expression could lead to a decrease in the autocrine control of growth via EGF receptors. The involvement of protein kinase C is suggested by a decrease of protein kinase activity during differentiation and, on the contrary, an increase of TGF-a mRNA levels under PMA stimulation; however, that relationship still needs to be confirmed. It will be of great interest to extend the findings observed in Capan 1 to other pancreatic cell lines to better define the influence of cell differentiation on the autocrine control of human pancreatic cancer cell growth by TGF-a. References Todaro GJ, Lee DC, Webb NR, Rose TM, Brown JP. Rat type a transforming growth factor: structure and possible function as a membrane receptor. In: Feramisco J, Ozanne B, Stiles C (eds). Growth factors and transformation. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1985:51-58. Twardzik DR. Differential expression of transforming growth factor-a during prenatal development of the mouse. Cancer Res 1985;45:5413-5416. Lee DC, Rose TM, Webb NR, Todaro GJ. Cloning and sequence analysis of a cDNA for rat transforming growth factora. Nature 1985;313:489-491. Kudlow JE, Leung AWC, Kobrin MS, Paterson AJ, Asa SL. Transforming growth factor-a in the mammalian brain. J Biol Chem 1989:264:3880-3883.
GASTROENTEROLOGY Vol. 103, No. 6
5. Tam JP, Sheikh MA, Solomon DS, Ossowski L. Efficient synthesis of human type a transforming growth factor: its physical and biological characterization. Proc Nat1 Acad Sci USA 1986;83:8082-8086. 6. Tam JP. Physiological effects of transforming growth factor in the newborn mouse. Science 1985;229:673-675. 7. Heldin CH, Westermark B. Growth factors as transforming proteins. Eur J Biochem 1989;184:487-496. 8. Smith JJ, Derynck R, Korc M. Production of transforming growth factor-a in human pancreatic cancer cells: evidence for a superagonist autocrine cycle. Proc Nat1 Acad Sci USA 1987;84:7567-7570. 9. Derynck R, Goeddel DV, Ullrich A, Gutterman JU, Williams RD, Bringman TS, Berger WH. Synthesis of messenger for transforming growth factor a and p and the epidermal growth factor receptor by human tumors. Cancer Res 1987;47:707712. 10. Sporn MB, Roberts AB, Wakefield LM, de Combrugghe B. Some recent advances in the chemistry and biology of transforming growth factor-beta. J Cell Biol 1987;105:1039-1045. 11. Pittelkow MR, Lindquist PB, Abraham RT, Graves-Deal R, Derynck R, Coffey RJ Jr. Induction of transforming growth factor-a expression in human keratinocytes by phorbol esters. J Biol Chem 1989;264:5164-5171. 12. Levrat JH, Palevody C, Ratovo DG, Hollande E. Differentiation of the human pancreatic adenocarcinoma cell line (CAPAN1)in culture and co-culture with fibroblasts, dome formation. lnt J Cancer 1984;34:177-185. 13. Bloom EJ, Siddiqui B, Hicks JW, Kim YS. Effect of sodium butyrate, a differentiating agent, on cell surface glycoconjugates of human pancreatic cell lines. Pancreas 1989;4:59-64. 14. Bensaadi N, Clemente F, Vaysse N. Modulation of enzymatic activities during spontaneous and induced differentiation in a human pancreatic adenocarcinoma cell line Capan 1. lnt J Pancreatol 1989;4:391-406. 15. Anderson WB, Estival A, Tapiovaara H, Gopala Krishna R. Altered subcellular distribution of protein kinase C (phorbol ester receptor). Possible role in tumor promotion and the regulation of cell growth. Advances in cyclic nucleotide and protein phosphorylation research. 1985;19:287-306. 16. Chirgwin JM, Przybyla AE, MacDonald RJ, Rutter WJ. Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 1979;18:5294-5299. 17. Thomas PS. Hybridization of denatured RNA and small DNA fragments transferred to nitrocellulose. Proc Nat1 Acad Sci USA 1980;77:5201-5205. 18. Bringman TS, Lindquist PB, Derynck R. Different transforming growth factor-a species are derived from a glycosylated and palmitoylated transmembrane precursor. Cell 1987;48: 429-440. 19. Laemli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970;227:680685. 20. Lowry OH, Rosebrough NJ, Farr AL, Randall NJ. Protein measurement with the folin phenol reagent. J Biol Chem 1951;193:265-275. 21. Nishizuka Y. The molecular heterogeneity of protein kinase C and its implication for cellular regulation. Nature 1988;334: 661-665. 22. Shoji M, Girard PR, Charp PA, Koeffler HP, Vogler WR, Kuo JF. Effects of phorbol ester on translocation and down regulation of protein kinase C and phosphorylation of endogenous proteins in human acute myeloid leukemia cell line KG-l and its phorbol ester-resistent subline GF-la. Cancer Res 1987;47:6363-6370, 23. Bornes C, Eppenberg V, Wyss R, Fabbro D. Continuous synthesis of two protein kinase C-related proteins after down regula-
December
1992
tion by phorbol esters. Proc Nat1 Acad Sci USA 1988;85:21102114. 24.
Massague J. Transforming growth factor-u. A model for membrane-anchored growth factors. J Biol Chem 1990;265:2139321396. 25. Pandiella A, Massague J. Cleavage of the membrane precursor for transforming growth factor a is a regulated process. Proc Nat1 Acad Sci USA 1991;88:1726-1730, 26. Glinsmann-Gibson BJ, Korc M. Regulation of transforming growth factor-u mRNA expression in T3M4 human pancreatic carcinoma cells. Pancreas 1991;6:142-149. 27. Oustry P, Estival A, Pradayrol L, Vaysse N, Clemente F. Two
TGF-a AND CELL DIFFERENTIATION
1859
subclasses of EGF receptors in the human pancreatic cancer cell lines Capan 1 and MIA PaCa 2. Int J Pancreatol 1990:6:119-128.
Received December 26, 1990. Accepted June 25,1992. Address requests for reprints to: Francois Clemente, M.D., INSERM U 151, Centre Hospitalier Universitaire Rangueil Bat L 3, 31054 Toulouse Cedex, France. Supported by grants from the Federation Nationale des Centres de Lutte contre le Cancer (FNCLCC), ARC (6165), and the National Institutes of Health (CA36544) (to J.P.T.).
