EXPERIMENTAL
CELL
RESEARCH
188,129-134
(1990)
The Induction of Vimentin Gene Expression by Sodium Butyrate in Human Promonocytic Leukemia U937 Cells CARLOS RIUS,* CARLOS CABARAS,*~~ AND PATRICIO ALLERGY’ *Centro de Investigaciones Biob’gicas, CSZC Vel&qw.z, 144, E-28006-Madrid, Spain; and j’Departamento de Bioquimica y Biologih Mokcu.!nr II, Facultaa’ de Medicina, Universidad Complutense, E-28040-Madrid, Spain
The administration of 1 n&sodium hutyrate induced the phenotypic differentiation of human promonocytic leukemia U937 cells, as judged by the expression of cD1 lb and cD1 lc antigens, two differentiation-specific surface markers. At the same time, butyrate greatly induced the expression at the mRNA level of the vimentin gene. The increase in the level of this RNA started at 6 h of treatment and reached the maximum at Hour 24. Such an increase was caused at least in part by a stimulation in the rate of gene transcription, as suggested by transcription assays in isolated nuclei. Experiments in the presence of cycloheximide suggested that vimentin induction is probably a direct response to the action of butyrate, not mediated by the prior induction of other gene products. Unlike the case of vimentin, the levels of other RNAs, namely @-actin, ornithine decarboxylase, and c-myc, were not enhanced, but they decreased at different times of treatment with butyrate. Finally, we observed that bntyrate induced also the differentiation of HL60 cells, another human myeloid cell type. Nevertheless, the drug failed to stimulate the expression of o 1990 Academic PIWSS, 1~. vimentin in this cell line.
INTRODUCTION The biological action of sodium butyrate has been a matter of considerable interest in recent years. This agent causes a variety of morphological and functional effects in plant and animal cells, such as increase in cell size [ 11, changes in cell shape and in cytoskeleton structure (see Ref. [2] for review), cell blockade at specific stages of the growth cycle [3-51, and stimulation or inhibition of the expression of specific genes [6,7]. In many cell types butyrate is also a powerful differentiation-inducing agent. This effect is often manifested by the induction of gene products characteristic of the mature phenotype, such as globin in erythroleukemia cells [8], prostaglandin synthetase in mastocytoma P-815 cells [9], and insulin and glucagon in pancreatic islet tumor 1 To whom
reprint
requests
should
be addressed.
cells [lo]. Therefore, butyrate may represent a useful tool with which to study the mechanisms controlling cell differentiation, as well as the regulation of genes the expression of which is associated with this process. In this work we analyze the capacity of sodium butyrate to promote the differentiation of U937 cells, a human promonocytic leukemia cell line [ 111, and to modulate the expression at the mRNA levels of the vimentin gene in this cell type. The work was undertaken for the following reasons: (a) Some established myeloid leukemia cell lines represent useful models for the study of mechanisms responsible for cell differentiation. In spite of their tumoral character, these cells may be induced to differentiate along specific lineages by several physiological and nonphysiological inducers [ 12,131, and this process has been correlated with changes in the expression of specific genes [13,14]. (b) It has been suggested that vimentin, the major intermediate-size filament protein in cells of mesenchimal origin, might be implicated in the differentiation of hemopoietic cells. This idea is supported by two types of evidences. First, the amount and the organization of vimentin may vary, in a different manner according to the cell lineage, during the maturation of normal hemopoietic precursors [ 151. Second, the expression of vimentin is stimulated by some differentiation inducers in stablished myeloid leukemia cell lines [16]. The results presented here indicate that sodium butyrate promotes the phenotypic differentiation of U937 cells. The drug greatly induces the expression at the mRNA level of the vimentin gene, and this induction starts early during the differentiation process. Nevertheless, although it also promotes their differentiation, butyrate fails to stimulate the expression of vimentin in HL60 cells, another human myeloid leukemia cell line [17]. MATERIALS
AND METHODS
Cell culture conditions. U937 cells and HL60 cells were grown in RPMI-1640 medium supplemented with 10% heat-inactivated fetal calf serum, 5 m&f Hepes buffer, 0.2% (w/v) sodium bicarbonate and antibiotics, in a humidified 5% COz atmosphere at 37’C. Cells were seeded in lOO-mm plastic dishes at 2.5-3.0 X 1O’cells per milliliter and
129
0014~4827/90
Copyright All rights
0 1990 by Academic
of reproduction
$3.00
Press, Inc.
in any form reserved.
130
RIUS, CABANAS, z
52P fz b Zl= t s! % 6
0 Control o +SB
AND ALLER
l
/
Control /O
4
O/O
I 1 1 2 Days after
1 3 plating
+SB
I 4
I 5
3 c = 6 Fluorescence intensity (DNA content )
FIG. 1. Effect of butyrate (1 mM) on U937 cell proliferation. (Left panel) Increase in cell number in untreated (control) and treated (+SB) cultures. The values (mean of three determinations) were adjusted to account for dilution of cells (both control and treated) on the 2nd day of incubation (see Materials and Methods). The approximate doubling times of the cells were 23 h in untreated cultures, and 25,47, and 59 h in cultures treated with butyrate for 1, 3, and 5 days, respectively. The frequencies of viable cells were 90 and 70% at Days 3 and 5 of butyrate treatment, respectively, and higher than 95% under all other conditions. (Right panel) Cell cycle distribution of untreated cells and cells treated for 5 days with SB. The histograms were generated by flow cytometry analysis of propidium iodide-stained nuclei. maintained in continuous logarithmic growth by passing them every 2 or 3 days. Drug treatment. Sodium butyrate (Merck) was dissolved at 0.1 M in RPM1 medium. The solution was freshly prepared, just before its application. Cell cultures were initiated at a density of 2.5 X lo5 cells per milliliter in a mixture of conditioned medium and fresh medium (approximately l/3, v/v) in the absence (control cultures) or in the presence (treated cultures) of 1 mA4 sodium butyrate. Cells from control cultures were harvested in the logarithmic growth phase. Cells from treated cultures were harvested at the times indicated in each experiment. To prevent long-term cultures from reaching plateau densities or nutrient exhaustion, which per se might affect the expression of the studied genes, on the 2nd day of treatment they were supplemented with an equal volume of fresh medium containing the inducer. In addition, in some experiments partof the medium plus butyrate was renewed daily. This was aimed at preventing possible effects derived from rapid degradation of the inducer. Both cell growth and viability were daily checked by using an hemocytometer and by trypan blue exclusion, respectively. [3H]!I’hymidine incorporation. To estimate the growth activity, cells were pulse-labeled for 2 h with 1 &i/ml of (mcthyl-3H)thymidine (25 Ci/mmol, Amersham, UK). The cells were collected by centrifugation and washed twice with cold PBS after which the radioactivity incorporated into trichloroacetic acid precipitable material was determined. Cetl cycle distribution. To determine the relative number of cells with different DNA contents, cells were incubated for 30 min at 4°C with RPM1 medium containing 0.1% NP40 and 50 @g/ml of propidium iodide. DNA histograms were then generated by flow cytometry with an EPICS-CS flow cytometer (Coulter Cientifica, Mostoles, Spain) with 200 mW excitation at 488 nm. To detect the expression of cell Detection of cell surface antigens. surface differentiation-specific antigens, indirect immunofluorescence determinations were carried out using the MAbs Bear 1 and HCl/l. The MAb Bear 1 (kindly provided by Dr. Jan de Vries, Unicet, France) recognizes the alpha subunit of the CDllb (Mol) antigen [18]. The MAb HCl/l (generated in our laboratories) recognizes the CDllc (p150/95) antigen [19]. The cells were incubated with the antibodies for 30 min at 4°C. After two washes with RPM1 medium, FITC-labeled sheep anti-mouse IgG (Amersham, UK) was added and the incubation followed for an additional period of 30 min at 4°C. After washing the cells twice with RPM1 medium, their fluorescence was estimated by flow cytometry with an EPIC-CS flow cytometer. Probes. The probes usedwere: the l.l-kb human vimentin-specific XhoI fragment of L&A plasmid [20]; the 1.5-kb ClaI-EcoRI fragment of pMC413rC plasmid, which contains the 3rd exon of human c-myc
[21]; the 1.5-kb human ODC-specific XhoI fragment of the OB-821 plasmid [22] and the 0.66-kb chicken fl-actin-specific KpnI-BglII fragment of pAL41 plasmid [23]. The fragments were isolated by elution after binding to NA45-DEAE membrane (Schleider and Schuell, FRG) in agarose gels and labeled to 1.0-1.5 X 10’ cpm/pg of DNA with (a-32P)dCTP (3000 Ci/mmol, Amersham, UK) by random hexanucleotide priming (241. RNA blot assays. Total cytoplasmic RNA was prepared as described in a previous work [25]. RNA samples (15 pg) were denatured and then electrophoresed in 1.1% (w/v) agarose-formaldehyde gels [26] and blotted onto nylon membrane (Hybond-N, Amersham). RNA blots were prehybridized, hybridized with excess 32P-labeled probes, washed under highly stringent conditions [27] and finally autoradiographed. The probes were routinely eluted with boiling water, to permit the same filters to be reused with different clones. Nuclear “run-on” transcription assay. Nuclei preparation and transcriptional activity determination were carried out as described by Simpson et al. [28]. pBR322 plaamid, L3A7A (vimentin) plasmid, and pMC413rC (c-myc) plasmid were linearized, denatured by adding l/10 vol of 3 M NaOH and the mixture was neutralized by adding l/ 10 vol of 10 M NH,Ac. Aliquots of 5 gg DNA were blotted onto nylon membrane. The filters were prehybridiid for 24 h with a mixture containing 4~ SSC, 5~ Denhardt’s solution, 50% deionized formamide, 0.5 n&f cold UTP, and 200 pg/ml salmon sperm DNA. Hybridization was carried out for 3 days at 37”C, using 2 X lo6 cpm per filter in the same mixture, except for the cold UTP, as described above. RESULTS
Cell Shape, Cell Growth, and Diff@rentiation Incubation with 1 mM sodium butyrate (SB) caused some change. in the shape of U937 cells. This consisted in the loss of the round form and the progressive adoption of more elongated, irregular forms. The change was first observed after 24 h of treatment and reached its maximum on the 2nd day. After 3 ddys of treatment, the formation of some cell clusters, which remained in suspension, was usually observed. Over the whole period studied (5 days) the cells did not attach to the plate surface. Incubation of U937 cell cultures with SB resulted in the progressive reduction of the growth activity, as mea-
VIMENTIN
GENE
EXPRESSION
IN
BUTYRATE-TREATED
TABLE
U937
131
CELLS
1
Effect of Butyrate (1 m&f) on [ 3H]Thymidine
Incorporation
in U937 Cells
Hours of treatment 0 cpm/106 %ofOh Note.
cells
41,241
12
+ 2,755 -
Results
are means
37,040
24
k 2,659 89.8
+ SD of two independent
32,246
f 2,255 78.2
26,005
72
f 1,251 63.1
20,084
f 1,778 48.7
120 16,123
f 1,625 39.1
determinations.
sured both by cell number increase (Fig. 1) and by [3H]thymidine incorporation (Table 1). Nevertheless, over the whole period studied a complete suppression of cell proliferation was not achieved. Examination of cell cycle distribution showed that butyrate-treated cells were arresting at the prereplicative stage, as revealed by the increase in the number of cells with Go/G1 DNA content and the decrease in the number of those having S and G, + M DNA content (Fig. 1). The capacity of butyrate to induce U937 cell differentiation was determined by measuring the reactivity with MAbs Bear 1 (anti-CDllb) and HCl/l (anti-CDllc) at different times of treatment. Previous studies had determined that CDllb and CDllc are differentiation-specific antigens [X3, 191. No increase in reactivity was observed when the cells were treated with SB for a short period of time (6 h). At 24 h of treatment we observed a slight reactivity with Bear 1, which increased thereafter. A great reactivity with HCl/l was already observed at 24 h of treatment (Fig. 2). Changes in the Expression
48
of Specific Genes
To analyze possible changes in the cellular content of vimentin RNA, Northern blot analyses were carried out using cytoplasmic RNA extracted from either untreated U937 cells or cells treated for increasing periods of time with 1 mMSB. Three different experiments were carried out which, with minor quantitative differences, gave similar results. One representative autoradiogram is
shown in Fig. 3. The levels of vimentin RNA were barely detectable in uninduced cells, but they increased greatly after treatment with SB. A slight increase (two- to threefold) was first observed at 6 h of treatment, and reached the maximum at Hour 24, after which vimentin RNA maintained roughly constant levels. To exclude the possibility of a general, nonspecific stimulatory action of butyrate on transcription, we measured the levels of other RNAs, namely @actin, ornithine decarboxylase (ODC), and c-myc. As shown by Fig. 3, none of these transcripts was increased by butyrate. The levels of @actin and ODC decreased to different extents after 6 h of treatment with the drug. The levels of c-myc RNA declined rapidly (l-3 h), but they were later reinduced. This reinduction was observed in all performed experiments, although both its intensity and the time when it occurred varied greatly. The reinduction of c-myc RNA was also observed in experiments in which part of the medium plus SB was daily renewed (see Materials and Methods), which seems to exclude the possibility of a recovery of c-myc expression caused by drug exhaustion. It must be noted that, although such a recovery was not observed with other inducers in human myeloid cells, a similar biphasic expression of c-myc has been reported to occur during the differentiation of murine erytroleukemia cells [ 29,301. To determine whether butyrate increases the transcription rate of the vimentin gene, “run-on” transcription assays were carried out using nuclei from either un-
24h
6h
24 h
1
72 h
I /’;“‘.“.fi
Bear 1
72 h
HCl/l
Fluorescence
FIG. (Dotted
2. Reactivity line) Control
of MAbs Bear cells. (Continuous
intensity
1 (anti-CDllb) and HCl/l (anti-CDllc) with untreated line) Cells treated with 1 n-&f SB for the indicated times.
(control)
and butyrate-treated
U937
cells.
132
RIUS, a
bcdefgh
i j
k
a
CABANAS, b
c
AND
ALLER 3oa
cl
Vimentin
&j
IS-actin
i:;:;:>:. .:.:.:.:.: ::::::::;:_ _ 3;:;::::; ::::g: ;;z;z .:.:.:.:.: :::::::::: .:::::::::: .,.,i. :::::::: j ij ::::::::.: it3 +CHM
c- myc
pBR322 +CHM FIG. 3. Expression of specific genes in butyrate-treated U937 cells. (Left panel) Time-course accumulation of vimentin, @-actin, ODC, and c-myc RNAs. Total cytoplasmic RNA was extracted from either untreated cells (a) or cells treated for 30 min (b), 1 h (c), 3 h (d), 6 h (e), 12 h (f), 24 h (g), 48 h (h), 72 h (i), 120 h (i), and 168 h (k) with 1 mM SB. RNA blots (15 pg per lane) were prepared and hybridized sequentially with the indicated clones. (Right panel) Transcriptional activation of the vimentin gene. Nuclei were isolated from either untreated cells (a) or cells treated for 24 h (b) and 72 h (c) with 1 n& SB, and run-on transcription assays were carried out. The 32P-labeled transcripts were hybridized against nylon membrane-immobilized pBR322, vimentin, and c-myc clones.
treated cells or cells treated with SB for 24 and 72 h, times at which the vimentin RNA levels have already reached the maximum. We found a great increase in the hybridization signal with the vimentin probe as a consequence of the butyrate treatment (Fig. 3), suggesting stimulation of the transcription activity of the gene. This observation does not necessarily rule out other effects, such as possible changes in vimentin RNA stability. This caution seems to be stressed by the fact that the hybridization signal with the vimentin probe was significantly higher at 24 h than at 72 h of treatment in the nuclear transcription assays, but not in the Northern blot assays (Fig. 3).
+S8
FIG. 4. Effect of cycloheximide (CHM, 70 j&f) on butyrate-produced stimulation of vimentin RNA levels in U937 cells. All treatments lasted for 6 h. When applied together, CHM was added 10 min before SB. Under the conditions used here, 1 h treatment with CHM inhibited in more than 97% the incorporation of [3H]leucine. The values, representing the average of two experiments, were obtained by densitometric readings of autoradiograms from Northern blots, and normalized in relation to the value in untreated (control) cells.
translation inhibitor (Fig. 4). Although other interpretations cannot be ruled out, these results might indicate that vimentin induction is a direct response to the action of butyrate, not mediated by the prior induction of other gene products. The Reversibility of the Butyrate Effects
To analyze whether the induction of vimentin RNA by butyrate is a reversible effect, U937 cells were treated for 24 h with SB and then allowed to recover in the absence of the drug. Both the cell number and the vimentin RNA level were measured at the end of butyrate treatment and at different times of recovery. The results are summarized in Fig. 5. We observed that both the growth kinetics and the vimentin RNA levels reassumed the same patterns as in untreated cells, but not immediately
The Effect of Cycloheximide
It has been reported that the induction of vimentin during the growth activation of resting cells is a primary response to mitogen action, since it is not dependent on newly synthesized proteins [31]. We were interested in finding out whether this is also the case for the butyrateproduced induction of vimentin RNA in differentiating U937 cells. For this purpose we measured the levels of this RNA in untreated cells, in cells treated for 6 h with 70 PM cycloheximide alone, and in cells treated for 6 h with SB either in the absence or in the presence of the translation inhibitor. Longer periods of treatment or higher concentrations of cycloheximide had to be avoided, since the drug then caused significant cell damage. We observed, first, that cycloheximide alone did not alter significantly the basal levels of vimentin RNA and, second, that butyrate increased this RNA to roughly the same levels in the absence and in the presence of the
:::::::::: ::::::::i: g$$ :::::::::: .:.:.:.:.:. y:::::::: ::i’ .:.:.:.:.:. $$$$ il:::::::::::
2-i
-l# =u) ‘i0” -24
abcde
/
.-. b
+SB-
i4 Hours
of
4.8 recovery
75
FIG. 5. Reversibility of butyrate effects. U937 cells were treated for 24 h with 1 mM SB, after which they were washed twice with RPM1 medium and allowed to recover in the absence of the drug. (Graph) Changes in cell number at different times of recovery. The values represent the average of two determinations. (Insert) Vimentin RNA levels in untreated cells (a), in cells treated for 24 h with SB (b), and in cells allowed to recover for 6 h (c), 24 h (d), and 46 h (e).
VIMENTIN
GENE
abcdefgh
EXPRESSION
./‘...,.-
IN 24h
I
BUTYRATE-TREATED
U937
CELLS
133
bol-13-acetate (TPA) [22], as was also shown to occur in other laboratories [ 201.
vim. DISCUSSION
Ftuorescence
Intensity
FIG, 6. Effect of butyrate on the expression of specific genes and on the expression of surface antigens in HL60 cells. (Left panel) Timecourse accumulation of specific RNAs. Total cytoplasmic RNA was extracted from either untreated cells (a) or cells treated for 1 h (b), 3 h (c), 6 h (d), 24 h (e), 48 h (f), 72 h (g), and 120 h (h) with 1 mM SB. RNA blots (15 pg per lane) were prepared and hybridized sequentially with the indicated clones. (Right panel) Reactivity of MAbs Bear 1 and HCI/l. (Dotted line) Untreated (control) cells. (Continuous line) Cells treated with 1 mM SB for the indicated times.
after butyrate withdrawal. In fact, a 24-h lag period had to occur before the cells reassumed a rapid growth kinetics, and a 48-h recovery period was necessary for vimentin RNA to drop to levels similar to those in noninduced cells. The Effect of Butyrate in HL60 Cells
Finally, we were interested in finding out whether butyrate also induces vimentin expression in other human myeloid cell types. For this purpose we measured the levels of vimentin RNA in untreated and in SB-treated HL60 cells, a human promyelocytic leukemia cell line which has been reported to undergo differentiation upon butyrate treatment [32]. Three different experiments were carried out, with essentially the same results. One representative autoradiogram is shown in Fig. 6. We observed that noninduced HL60 cells had relatively high levels of vimentin RNA, when compared to the barely detectable levels of this transcript in noninduced U937 cells (see Fig. 3). In addition, it could be observed that butyrate failed to stimulate vimentin expression in HL60 cells. In fact, the levels of this RNA remained constant until Hour 24 and then decreased (Fig. 6). The failure of butyrate to stimulate vimentin expression in HL60 cells cannot be attributed to a general lack of response of the cell clone used by us to the action of this inducer. In fact, butyrate provoked the expression of differentiation-specific antigens, as proved by the increase in reactivity with MAbs Bear 1 and HCl/l (Fig. 6). Moreover, the drug produced a rapid decline of c-myc RNA levels, a response which has been commonly observed upon treatment with other inducers [22, 33, 341. Finally, it may be of interest to point out that in our experience the expression of vimentin RNA in HL60 cells was rapidly stimulated by 12-O-tetradecanoylphor-
The results presented here indicate that: (a) Sodium butyrate induces the phenotypic differentiation of U937 cells, as proved by the induction of two surface differentiation-specific antigens. (b) In addition, butyrate greatly induces the expression of vimentin, at the mRNA levels, in this cell line. The term expression is currently used here to indicate steady-state levels of specific RNAs, without prejudicing whether they are the result of transcriptional or post-transcriptional regulatory mechanisms. Nevertheless, it appears that the stimulation of vimentin RNA content is the result, at least in part, of an increased rate of gene transcription, as suggested by transcriptional assays in isolated nuclei. (c) The induction of vimentin seems to be an early phenomenon in the differentiation process, in the sense that it precedes other measured alterations in cell morphology and phenotype, and in that it is not probably mediated by the prior induction of other gene products. (d) While butyrate is also effective in promoting the differentiation of HL60 cells, the drug does not stimulate the expression of vimentin in this cell line. One of the goals in the studies of cell differentiation consists of the characterization of the genes implicated in this process. The possibility that vimentin could be one of these genes has been often inferred from the fact that TPA, a potent maturation-inducing agent, stimulates its expression in cultured blood cells [16, 20, 221. Nevertheless, this observation must be considered with caution, since TPA, a protein kinase C activator, is capable per se of increasing the expression of many genes, irrespective of the cell species and of the final consequence for the cell phenotype (see Ref. [35] for review). For instance, it has been shown that TPA increases vimentin mRNA [36] and protein [37] levels in fibroblast cell lines which do not undergo differentiation. Also, we have observed that the phorbol ester stimulates the expression of the ODC gene in human myeloid cells [22], whereas ODC does not appear to be directly implicated in the differentiation of these cells [22, 381. The use of butyrate as a differentiation-inducing agent allowed us to overcome these difficulties. First, butyrate is not per se a general activator of vimentin expression. Earlier results have indicated that this agent does not alter [7] or even decrease [39] vimentin expression in fibroblast cells. Second, butyrate is not a general inducer of gene expression, since only vimentin, and no one of the other genes examined by us, was stimulated. Hence, it appears that the induction of vimentin in butyrate-treated U937 cells is an event directly linked to the differentiation of these cells. We are at present confirming this conclusion by using other inducers. Nevertheless, the exact role of
RIUS,
CABANAS,
vimentin in the differentiation process remains to be determined (see Ref. [40] for a detailed discussion on this subject). It may be somewhat surprising that, while butyrate activates greatly vimentin expression in U937 cells, it fails to produce the same effect in the related HL60 cell line. Nevertheless, it must be noted that noninduced HL60 cells-or, at least, the cell clone used by us-seem to have a significantly high expression of vimentin. This was here measured at the RNA level, but was also confirmed at the protein level by immunofluorescence observations (results not shown). Therefore, an explanation could be that the vimentin content in noninduced HL60 cells suffices to fulfill the requirements for this protein during differentiation, thus making unnecessary its further increase. This work was supported by DGICYT (Spain) Grant PB87-0351 (to P.A.). C.R. is the recipient of a fellowship from the Fundacion Cientifica de la Asociacion Espaiiola Contra el Cancer. The authors are indebted to Ms. M. C. Granados and to Mr. P. Lastres for technical assistance, to Mr. E. Wiltshire for checking the English version of this manuscript, and to Mrs. M. 0. Partearroyo for secretarial work.
REFERENCES 1.
Kimech, CellRes.
N., Hill,
A-M.,
Seif, R., and Pantaloni,
D. (1986)
162,326-334.
2. Kruh, J. (1982) Mol. Cell. Biochem. 42,65-82. 3. Xue, S., and Rao, P. N. (1981) J. Cell Sci. 51,163-171. 4. Chabanas, A., Khoury, E., Goeltz, P., Froussard, P., Gjerset, Dod, B., Eisen, 141-151.
H., and Lawrence,
5. Lanzagorta, Plant
6. Yuan, Chem.
J. M. A., De la Torre, 72,775-781. Z. A., Hirschhorn, 260,3778-3783.
J. J. (1985) C., and Aller,
R. R., and Baserga,
7. Aller,
P., Hirschhorn, l&119-127.
R. R., and Rius,
J. Mol.
Samstrom,
12.
Hozumi,
13.
Collins,
S. J. (1987)
14.
Muller,
R. (1986)
Received Revised
D. J., Chick, 7,560-563.
C., andNilsson, M. (1983)
August 15,1989 version received
16.
Bernal, S. D., and Chen, L. B. (1982) Cancer Collins, S. J., Gallo, R. C., and Gallagher, (London) 270,347-349.
17. 18.
19.
Mol.
deh, B. N. (1987) J. Biol. Chem. 262,4104-4108. H., and Skoultchi, A. (1984) Nature
29. Lachman, 592-594. 30. Watanabe,
Ada Blood TZBS
K. (1976) Cancer
70,1233-1244.
11,129-132.
December
Znt. J. Cancer
Res. 38,121-169.
27,1989
17,565-577.
J.
T., Sherman, J. Cell. Physiol.
M.,
Shafman,
fLondon,J
T., Iwata,
310,
T., and Kufe,
127,480-484.
31. Rittling,
Znt.
J. F.
M. (1986)
Euol. 23,11-22.
24. Feinberg, B. P., and Vogelstein, B. (1984) Anal. Biochem. 137, 266-267. 25. Aller, P., and Baserga, R. (1986) J. Cell Physiol. 128,362-366. 26. Lehrach, H., Diamond, D., Wozney, M., and Boedtke, H. (1977) Biochemistry 16,4743-4749. 27. Hirschhorn, R. R., Aller, P., Yuan, Z. A., Gibson, C. W., andBaserga, R. (1984) Proc. Natl. Acad. Sci. USA 81,6004-6008. 28. Simpson, R. U., Hsu, T., Begley, D. A., Mitchell, B. S., and Aliza-
34.
Habener,
S., Franchini, G., F. (1982) Proc. Natl.
22. Rius, C., and Aller, P. (1989) Cell Differ. 28,39-46. 23. Alonso, S., Minty, A., Bourlet, Y ., and Buckingham,
J. Biol.
S. (1980)
F., Ter-
l&1142-
Sanchez-Madrid,F., Acevedo, A.,Bellon,T., Fern&nLarraga, V., and Bernabeu, C. (1988) Hybridoma 7,
R., Gelmann, E. P., Martinotti, Papas, T. S., Gallo, R. C., and Wong-Staal, Acad. Sci. USA 79,6497-6501.
R. (1985)
W. L., and
Cabaiias,C., dez, J. M., 167-175.
21. Dalla Favera,
32. 33.
Biochem.
Res. 42,5106-5116. R. E. (1977) Nature
S., Battini, R., Kaczmarek, L., Rittling, S., Calabretta, B., De Riel, J. K., Philiponis, V., Wei, J. F., and Baserga, R. (1986) Mol. Cell. Biol. 6,3614-3620.
Physiol.
C. (1989)
D., and Bronet,
Keizer, G. D., Birst, J., Figdor, C. G., Spits, H., Midema, host, G., and De Vries, J. E. (1985) Eur. J. Zmmunol. 1148.
D. (1986) R.,
G., Paulin,
20. Ferrari,
35.
Philippe, J., Drucker, (1987) Mol. Cell Biol.
11.
Dellagi, K., Vaincheuker, W., Vinci, J. C. (1983) EMBO J. 2,1509-1514.
183,
Biol.
ALLER
15.
P. (1988)
8. Leder, A., and Leder, P. (1975) Cell 5,319-322. 9. Koshihara, Y., Senshu, T., Kawamura, M., and Murota, Biochim. Biophys. Acta 617,536-539. 10.
Exp.
AND
36.
S. R., Gibson, G. W., Ferrari, S., and Baserga, R. (1985) Biochem. Biophys. Res. Commun. 132,327-335. Boyd, A., and Metcalf, D. (1984) Leuk. Res. 8,27. Wattanabe, T., Sariban, E., Mitchell, T., and Kufe, D. (1985) Biochem. Biophys. Res. Commun. 126,999-1005. Trepel, J. B., Calamoni, 0. R., Kelly, K., Schwab, G., Watt, R. A., Sansville, E. A., Jaffle, E. S., and Neckers, L. M. (1987) Mol. Cell. Biol. 7,2644-2648. Kikkawa, U., and Nishizuka, Y. (1986) Annu. Rev. Cell Biol. 2, 149-178. Rittling, S. R., Coutinho, L., Amram, T., and Kolbe, M. (1989) Nucleic Acids Res. 17,1619-1632. Laszlo, A., and Bissell, M. J. (1983) Exp. Cell Res. 148,221-234.
37. 38. Luk,
(1982)
G. D., Civin, C. I., Weismann, Science 216,75-77.
R. M.,
and
Bayleu,
39. Ryan, M. P., and Higgins, P. J. (1988) J. Cell Physiol. 34. 40. Traub, P. (1985) Intermediate Filaments: A Review, Verlag,
Berlin.
S. B.
137, 25Springer
CELL
RESEARCH
188,129-134
(1990)
The Induction of Vimentin Gene Expression by Sodium Butyrate in Human Promonocytic Leukemia U937 Cells CARLOS RIUS,* CARLOS CABARAS,*~~ AND PATRICIO ALLERGY’ *Centro de Investigaciones Biob’gicas, CSZC Vel&qw.z, 144, E-28006-Madrid, Spain; and j’Departamento de Bioquimica y Biologih Mokcu.!nr II, Facultaa’ de Medicina, Universidad Complutense, E-28040-Madrid, Spain
The administration of 1 n&sodium hutyrate induced the phenotypic differentiation of human promonocytic leukemia U937 cells, as judged by the expression of cD1 lb and cD1 lc antigens, two differentiation-specific surface markers. At the same time, butyrate greatly induced the expression at the mRNA level of the vimentin gene. The increase in the level of this RNA started at 6 h of treatment and reached the maximum at Hour 24. Such an increase was caused at least in part by a stimulation in the rate of gene transcription, as suggested by transcription assays in isolated nuclei. Experiments in the presence of cycloheximide suggested that vimentin induction is probably a direct response to the action of butyrate, not mediated by the prior induction of other gene products. Unlike the case of vimentin, the levels of other RNAs, namely @-actin, ornithine decarboxylase, and c-myc, were not enhanced, but they decreased at different times of treatment with butyrate. Finally, we observed that bntyrate induced also the differentiation of HL60 cells, another human myeloid cell type. Nevertheless, the drug failed to stimulate the expression of o 1990 Academic PIWSS, 1~. vimentin in this cell line.
INTRODUCTION The biological action of sodium butyrate has been a matter of considerable interest in recent years. This agent causes a variety of morphological and functional effects in plant and animal cells, such as increase in cell size [ 11, changes in cell shape and in cytoskeleton structure (see Ref. [2] for review), cell blockade at specific stages of the growth cycle [3-51, and stimulation or inhibition of the expression of specific genes [6,7]. In many cell types butyrate is also a powerful differentiation-inducing agent. This effect is often manifested by the induction of gene products characteristic of the mature phenotype, such as globin in erythroleukemia cells [8], prostaglandin synthetase in mastocytoma P-815 cells [9], and insulin and glucagon in pancreatic islet tumor 1 To whom
reprint
requests
should
be addressed.
cells [lo]. Therefore, butyrate may represent a useful tool with which to study the mechanisms controlling cell differentiation, as well as the regulation of genes the expression of which is associated with this process. In this work we analyze the capacity of sodium butyrate to promote the differentiation of U937 cells, a human promonocytic leukemia cell line [ 111, and to modulate the expression at the mRNA levels of the vimentin gene in this cell type. The work was undertaken for the following reasons: (a) Some established myeloid leukemia cell lines represent useful models for the study of mechanisms responsible for cell differentiation. In spite of their tumoral character, these cells may be induced to differentiate along specific lineages by several physiological and nonphysiological inducers [ 12,131, and this process has been correlated with changes in the expression of specific genes [13,14]. (b) It has been suggested that vimentin, the major intermediate-size filament protein in cells of mesenchimal origin, might be implicated in the differentiation of hemopoietic cells. This idea is supported by two types of evidences. First, the amount and the organization of vimentin may vary, in a different manner according to the cell lineage, during the maturation of normal hemopoietic precursors [ 151. Second, the expression of vimentin is stimulated by some differentiation inducers in stablished myeloid leukemia cell lines [16]. The results presented here indicate that sodium butyrate promotes the phenotypic differentiation of U937 cells. The drug greatly induces the expression at the mRNA level of the vimentin gene, and this induction starts early during the differentiation process. Nevertheless, although it also promotes their differentiation, butyrate fails to stimulate the expression of vimentin in HL60 cells, another human myeloid leukemia cell line [17]. MATERIALS
AND METHODS
Cell culture conditions. U937 cells and HL60 cells were grown in RPMI-1640 medium supplemented with 10% heat-inactivated fetal calf serum, 5 m&f Hepes buffer, 0.2% (w/v) sodium bicarbonate and antibiotics, in a humidified 5% COz atmosphere at 37’C. Cells were seeded in lOO-mm plastic dishes at 2.5-3.0 X 1O’cells per milliliter and
129
0014~4827/90
Copyright All rights
0 1990 by Academic
of reproduction
$3.00
Press, Inc.
in any form reserved.
130
RIUS, CABANAS, z
52P fz b Zl= t s! % 6
0 Control o +SB
AND ALLER
l
/
Control /O
4
O/O
I 1 1 2 Days after
1 3 plating
+SB
I 4
I 5
3 c = 6 Fluorescence intensity (DNA content )
FIG. 1. Effect of butyrate (1 mM) on U937 cell proliferation. (Left panel) Increase in cell number in untreated (control) and treated (+SB) cultures. The values (mean of three determinations) were adjusted to account for dilution of cells (both control and treated) on the 2nd day of incubation (see Materials and Methods). The approximate doubling times of the cells were 23 h in untreated cultures, and 25,47, and 59 h in cultures treated with butyrate for 1, 3, and 5 days, respectively. The frequencies of viable cells were 90 and 70% at Days 3 and 5 of butyrate treatment, respectively, and higher than 95% under all other conditions. (Right panel) Cell cycle distribution of untreated cells and cells treated for 5 days with SB. The histograms were generated by flow cytometry analysis of propidium iodide-stained nuclei. maintained in continuous logarithmic growth by passing them every 2 or 3 days. Drug treatment. Sodium butyrate (Merck) was dissolved at 0.1 M in RPM1 medium. The solution was freshly prepared, just before its application. Cell cultures were initiated at a density of 2.5 X lo5 cells per milliliter in a mixture of conditioned medium and fresh medium (approximately l/3, v/v) in the absence (control cultures) or in the presence (treated cultures) of 1 mA4 sodium butyrate. Cells from control cultures were harvested in the logarithmic growth phase. Cells from treated cultures were harvested at the times indicated in each experiment. To prevent long-term cultures from reaching plateau densities or nutrient exhaustion, which per se might affect the expression of the studied genes, on the 2nd day of treatment they were supplemented with an equal volume of fresh medium containing the inducer. In addition, in some experiments partof the medium plus butyrate was renewed daily. This was aimed at preventing possible effects derived from rapid degradation of the inducer. Both cell growth and viability were daily checked by using an hemocytometer and by trypan blue exclusion, respectively. [3H]!I’hymidine incorporation. To estimate the growth activity, cells were pulse-labeled for 2 h with 1 &i/ml of (mcthyl-3H)thymidine (25 Ci/mmol, Amersham, UK). The cells were collected by centrifugation and washed twice with cold PBS after which the radioactivity incorporated into trichloroacetic acid precipitable material was determined. Cetl cycle distribution. To determine the relative number of cells with different DNA contents, cells were incubated for 30 min at 4°C with RPM1 medium containing 0.1% NP40 and 50 @g/ml of propidium iodide. DNA histograms were then generated by flow cytometry with an EPICS-CS flow cytometer (Coulter Cientifica, Mostoles, Spain) with 200 mW excitation at 488 nm. To detect the expression of cell Detection of cell surface antigens. surface differentiation-specific antigens, indirect immunofluorescence determinations were carried out using the MAbs Bear 1 and HCl/l. The MAb Bear 1 (kindly provided by Dr. Jan de Vries, Unicet, France) recognizes the alpha subunit of the CDllb (Mol) antigen [18]. The MAb HCl/l (generated in our laboratories) recognizes the CDllc (p150/95) antigen [19]. The cells were incubated with the antibodies for 30 min at 4°C. After two washes with RPM1 medium, FITC-labeled sheep anti-mouse IgG (Amersham, UK) was added and the incubation followed for an additional period of 30 min at 4°C. After washing the cells twice with RPM1 medium, their fluorescence was estimated by flow cytometry with an EPIC-CS flow cytometer. Probes. The probes usedwere: the l.l-kb human vimentin-specific XhoI fragment of L&A plasmid [20]; the 1.5-kb ClaI-EcoRI fragment of pMC413rC plasmid, which contains the 3rd exon of human c-myc
[21]; the 1.5-kb human ODC-specific XhoI fragment of the OB-821 plasmid [22] and the 0.66-kb chicken fl-actin-specific KpnI-BglII fragment of pAL41 plasmid [23]. The fragments were isolated by elution after binding to NA45-DEAE membrane (Schleider and Schuell, FRG) in agarose gels and labeled to 1.0-1.5 X 10’ cpm/pg of DNA with (a-32P)dCTP (3000 Ci/mmol, Amersham, UK) by random hexanucleotide priming (241. RNA blot assays. Total cytoplasmic RNA was prepared as described in a previous work [25]. RNA samples (15 pg) were denatured and then electrophoresed in 1.1% (w/v) agarose-formaldehyde gels [26] and blotted onto nylon membrane (Hybond-N, Amersham). RNA blots were prehybridized, hybridized with excess 32P-labeled probes, washed under highly stringent conditions [27] and finally autoradiographed. The probes were routinely eluted with boiling water, to permit the same filters to be reused with different clones. Nuclear “run-on” transcription assay. Nuclei preparation and transcriptional activity determination were carried out as described by Simpson et al. [28]. pBR322 plaamid, L3A7A (vimentin) plasmid, and pMC413rC (c-myc) plasmid were linearized, denatured by adding l/10 vol of 3 M NaOH and the mixture was neutralized by adding l/ 10 vol of 10 M NH,Ac. Aliquots of 5 gg DNA were blotted onto nylon membrane. The filters were prehybridiid for 24 h with a mixture containing 4~ SSC, 5~ Denhardt’s solution, 50% deionized formamide, 0.5 n&f cold UTP, and 200 pg/ml salmon sperm DNA. Hybridization was carried out for 3 days at 37”C, using 2 X lo6 cpm per filter in the same mixture, except for the cold UTP, as described above. RESULTS
Cell Shape, Cell Growth, and Diff@rentiation Incubation with 1 mM sodium butyrate (SB) caused some change. in the shape of U937 cells. This consisted in the loss of the round form and the progressive adoption of more elongated, irregular forms. The change was first observed after 24 h of treatment and reached its maximum on the 2nd day. After 3 ddys of treatment, the formation of some cell clusters, which remained in suspension, was usually observed. Over the whole period studied (5 days) the cells did not attach to the plate surface. Incubation of U937 cell cultures with SB resulted in the progressive reduction of the growth activity, as mea-
VIMENTIN
GENE
EXPRESSION
IN
BUTYRATE-TREATED
TABLE
U937
131
CELLS
1
Effect of Butyrate (1 m&f) on [ 3H]Thymidine
Incorporation
in U937 Cells
Hours of treatment 0 cpm/106 %ofOh Note.
cells
41,241
12
+ 2,755 -
Results
are means
37,040
24
k 2,659 89.8
+ SD of two independent
32,246
f 2,255 78.2
26,005
72
f 1,251 63.1
20,084
f 1,778 48.7
120 16,123
f 1,625 39.1
determinations.
sured both by cell number increase (Fig. 1) and by [3H]thymidine incorporation (Table 1). Nevertheless, over the whole period studied a complete suppression of cell proliferation was not achieved. Examination of cell cycle distribution showed that butyrate-treated cells were arresting at the prereplicative stage, as revealed by the increase in the number of cells with Go/G1 DNA content and the decrease in the number of those having S and G, + M DNA content (Fig. 1). The capacity of butyrate to induce U937 cell differentiation was determined by measuring the reactivity with MAbs Bear 1 (anti-CDllb) and HCl/l (anti-CDllc) at different times of treatment. Previous studies had determined that CDllb and CDllc are differentiation-specific antigens [X3, 191. No increase in reactivity was observed when the cells were treated with SB for a short period of time (6 h). At 24 h of treatment we observed a slight reactivity with Bear 1, which increased thereafter. A great reactivity with HCl/l was already observed at 24 h of treatment (Fig. 2). Changes in the Expression
48
of Specific Genes
To analyze possible changes in the cellular content of vimentin RNA, Northern blot analyses were carried out using cytoplasmic RNA extracted from either untreated U937 cells or cells treated for increasing periods of time with 1 mMSB. Three different experiments were carried out which, with minor quantitative differences, gave similar results. One representative autoradiogram is
shown in Fig. 3. The levels of vimentin RNA were barely detectable in uninduced cells, but they increased greatly after treatment with SB. A slight increase (two- to threefold) was first observed at 6 h of treatment, and reached the maximum at Hour 24, after which vimentin RNA maintained roughly constant levels. To exclude the possibility of a general, nonspecific stimulatory action of butyrate on transcription, we measured the levels of other RNAs, namely @actin, ornithine decarboxylase (ODC), and c-myc. As shown by Fig. 3, none of these transcripts was increased by butyrate. The levels of @actin and ODC decreased to different extents after 6 h of treatment with the drug. The levels of c-myc RNA declined rapidly (l-3 h), but they were later reinduced. This reinduction was observed in all performed experiments, although both its intensity and the time when it occurred varied greatly. The reinduction of c-myc RNA was also observed in experiments in which part of the medium plus SB was daily renewed (see Materials and Methods), which seems to exclude the possibility of a recovery of c-myc expression caused by drug exhaustion. It must be noted that, although such a recovery was not observed with other inducers in human myeloid cells, a similar biphasic expression of c-myc has been reported to occur during the differentiation of murine erytroleukemia cells [ 29,301. To determine whether butyrate increases the transcription rate of the vimentin gene, “run-on” transcription assays were carried out using nuclei from either un-
24h
6h
24 h
1
72 h
I /’;“‘.“.fi
Bear 1
72 h
HCl/l
Fluorescence
FIG. (Dotted
2. Reactivity line) Control
of MAbs Bear cells. (Continuous
intensity
1 (anti-CDllb) and HCl/l (anti-CDllc) with untreated line) Cells treated with 1 n-&f SB for the indicated times.
(control)
and butyrate-treated
U937
cells.
132
RIUS, a
bcdefgh
i j
k
a
CABANAS, b
c
AND
ALLER 3oa
cl
Vimentin
&j
IS-actin
i:;:;:>:. .:.:.:.:.: ::::::::;:_ _ 3;:;::::; ::::g: ;;z;z .:.:.:.:.: :::::::::: .:::::::::: .,.,i. :::::::: j ij ::::::::.: it3 +CHM
c- myc
pBR322 +CHM FIG. 3. Expression of specific genes in butyrate-treated U937 cells. (Left panel) Time-course accumulation of vimentin, @-actin, ODC, and c-myc RNAs. Total cytoplasmic RNA was extracted from either untreated cells (a) or cells treated for 30 min (b), 1 h (c), 3 h (d), 6 h (e), 12 h (f), 24 h (g), 48 h (h), 72 h (i), 120 h (i), and 168 h (k) with 1 mM SB. RNA blots (15 pg per lane) were prepared and hybridized sequentially with the indicated clones. (Right panel) Transcriptional activation of the vimentin gene. Nuclei were isolated from either untreated cells (a) or cells treated for 24 h (b) and 72 h (c) with 1 n& SB, and run-on transcription assays were carried out. The 32P-labeled transcripts were hybridized against nylon membrane-immobilized pBR322, vimentin, and c-myc clones.
treated cells or cells treated with SB for 24 and 72 h, times at which the vimentin RNA levels have already reached the maximum. We found a great increase in the hybridization signal with the vimentin probe as a consequence of the butyrate treatment (Fig. 3), suggesting stimulation of the transcription activity of the gene. This observation does not necessarily rule out other effects, such as possible changes in vimentin RNA stability. This caution seems to be stressed by the fact that the hybridization signal with the vimentin probe was significantly higher at 24 h than at 72 h of treatment in the nuclear transcription assays, but not in the Northern blot assays (Fig. 3).
+S8
FIG. 4. Effect of cycloheximide (CHM, 70 j&f) on butyrate-produced stimulation of vimentin RNA levels in U937 cells. All treatments lasted for 6 h. When applied together, CHM was added 10 min before SB. Under the conditions used here, 1 h treatment with CHM inhibited in more than 97% the incorporation of [3H]leucine. The values, representing the average of two experiments, were obtained by densitometric readings of autoradiograms from Northern blots, and normalized in relation to the value in untreated (control) cells.
translation inhibitor (Fig. 4). Although other interpretations cannot be ruled out, these results might indicate that vimentin induction is a direct response to the action of butyrate, not mediated by the prior induction of other gene products. The Reversibility of the Butyrate Effects
To analyze whether the induction of vimentin RNA by butyrate is a reversible effect, U937 cells were treated for 24 h with SB and then allowed to recover in the absence of the drug. Both the cell number and the vimentin RNA level were measured at the end of butyrate treatment and at different times of recovery. The results are summarized in Fig. 5. We observed that both the growth kinetics and the vimentin RNA levels reassumed the same patterns as in untreated cells, but not immediately
The Effect of Cycloheximide
It has been reported that the induction of vimentin during the growth activation of resting cells is a primary response to mitogen action, since it is not dependent on newly synthesized proteins [31]. We were interested in finding out whether this is also the case for the butyrateproduced induction of vimentin RNA in differentiating U937 cells. For this purpose we measured the levels of this RNA in untreated cells, in cells treated for 6 h with 70 PM cycloheximide alone, and in cells treated for 6 h with SB either in the absence or in the presence of the translation inhibitor. Longer periods of treatment or higher concentrations of cycloheximide had to be avoided, since the drug then caused significant cell damage. We observed, first, that cycloheximide alone did not alter significantly the basal levels of vimentin RNA and, second, that butyrate increased this RNA to roughly the same levels in the absence and in the presence of the
:::::::::: ::::::::i: g$$ :::::::::: .:.:.:.:.:. y:::::::: ::i’ .:.:.:.:.:. $$$$ il:::::::::::
2-i
-l# =u) ‘i0” -24
abcde
/
.-. b
+SB-
i4 Hours
of
4.8 recovery
75
FIG. 5. Reversibility of butyrate effects. U937 cells were treated for 24 h with 1 mM SB, after which they were washed twice with RPM1 medium and allowed to recover in the absence of the drug. (Graph) Changes in cell number at different times of recovery. The values represent the average of two determinations. (Insert) Vimentin RNA levels in untreated cells (a), in cells treated for 24 h with SB (b), and in cells allowed to recover for 6 h (c), 24 h (d), and 46 h (e).
VIMENTIN
GENE
abcdefgh
EXPRESSION
./‘...,.-
IN 24h
I
BUTYRATE-TREATED
U937
CELLS
133
bol-13-acetate (TPA) [22], as was also shown to occur in other laboratories [ 201.
vim. DISCUSSION
Ftuorescence
Intensity
FIG, 6. Effect of butyrate on the expression of specific genes and on the expression of surface antigens in HL60 cells. (Left panel) Timecourse accumulation of specific RNAs. Total cytoplasmic RNA was extracted from either untreated cells (a) or cells treated for 1 h (b), 3 h (c), 6 h (d), 24 h (e), 48 h (f), 72 h (g), and 120 h (h) with 1 mM SB. RNA blots (15 pg per lane) were prepared and hybridized sequentially with the indicated clones. (Right panel) Reactivity of MAbs Bear 1 and HCI/l. (Dotted line) Untreated (control) cells. (Continuous line) Cells treated with 1 mM SB for the indicated times.
after butyrate withdrawal. In fact, a 24-h lag period had to occur before the cells reassumed a rapid growth kinetics, and a 48-h recovery period was necessary for vimentin RNA to drop to levels similar to those in noninduced cells. The Effect of Butyrate in HL60 Cells
Finally, we were interested in finding out whether butyrate also induces vimentin expression in other human myeloid cell types. For this purpose we measured the levels of vimentin RNA in untreated and in SB-treated HL60 cells, a human promyelocytic leukemia cell line which has been reported to undergo differentiation upon butyrate treatment [32]. Three different experiments were carried out, with essentially the same results. One representative autoradiogram is shown in Fig. 6. We observed that noninduced HL60 cells had relatively high levels of vimentin RNA, when compared to the barely detectable levels of this transcript in noninduced U937 cells (see Fig. 3). In addition, it could be observed that butyrate failed to stimulate vimentin expression in HL60 cells. In fact, the levels of this RNA remained constant until Hour 24 and then decreased (Fig. 6). The failure of butyrate to stimulate vimentin expression in HL60 cells cannot be attributed to a general lack of response of the cell clone used by us to the action of this inducer. In fact, butyrate provoked the expression of differentiation-specific antigens, as proved by the increase in reactivity with MAbs Bear 1 and HCl/l (Fig. 6). Moreover, the drug produced a rapid decline of c-myc RNA levels, a response which has been commonly observed upon treatment with other inducers [22, 33, 341. Finally, it may be of interest to point out that in our experience the expression of vimentin RNA in HL60 cells was rapidly stimulated by 12-O-tetradecanoylphor-
The results presented here indicate that: (a) Sodium butyrate induces the phenotypic differentiation of U937 cells, as proved by the induction of two surface differentiation-specific antigens. (b) In addition, butyrate greatly induces the expression of vimentin, at the mRNA levels, in this cell line. The term expression is currently used here to indicate steady-state levels of specific RNAs, without prejudicing whether they are the result of transcriptional or post-transcriptional regulatory mechanisms. Nevertheless, it appears that the stimulation of vimentin RNA content is the result, at least in part, of an increased rate of gene transcription, as suggested by transcriptional assays in isolated nuclei. (c) The induction of vimentin seems to be an early phenomenon in the differentiation process, in the sense that it precedes other measured alterations in cell morphology and phenotype, and in that it is not probably mediated by the prior induction of other gene products. (d) While butyrate is also effective in promoting the differentiation of HL60 cells, the drug does not stimulate the expression of vimentin in this cell line. One of the goals in the studies of cell differentiation consists of the characterization of the genes implicated in this process. The possibility that vimentin could be one of these genes has been often inferred from the fact that TPA, a potent maturation-inducing agent, stimulates its expression in cultured blood cells [16, 20, 221. Nevertheless, this observation must be considered with caution, since TPA, a protein kinase C activator, is capable per se of increasing the expression of many genes, irrespective of the cell species and of the final consequence for the cell phenotype (see Ref. [35] for review). For instance, it has been shown that TPA increases vimentin mRNA [36] and protein [37] levels in fibroblast cell lines which do not undergo differentiation. Also, we have observed that the phorbol ester stimulates the expression of the ODC gene in human myeloid cells [22], whereas ODC does not appear to be directly implicated in the differentiation of these cells [22, 381. The use of butyrate as a differentiation-inducing agent allowed us to overcome these difficulties. First, butyrate is not per se a general activator of vimentin expression. Earlier results have indicated that this agent does not alter [7] or even decrease [39] vimentin expression in fibroblast cells. Second, butyrate is not a general inducer of gene expression, since only vimentin, and no one of the other genes examined by us, was stimulated. Hence, it appears that the induction of vimentin in butyrate-treated U937 cells is an event directly linked to the differentiation of these cells. We are at present confirming this conclusion by using other inducers. Nevertheless, the exact role of
RIUS,
CABANAS,
vimentin in the differentiation process remains to be determined (see Ref. [40] for a detailed discussion on this subject). It may be somewhat surprising that, while butyrate activates greatly vimentin expression in U937 cells, it fails to produce the same effect in the related HL60 cell line. Nevertheless, it must be noted that noninduced HL60 cells-or, at least, the cell clone used by us-seem to have a significantly high expression of vimentin. This was here measured at the RNA level, but was also confirmed at the protein level by immunofluorescence observations (results not shown). Therefore, an explanation could be that the vimentin content in noninduced HL60 cells suffices to fulfill the requirements for this protein during differentiation, thus making unnecessary its further increase. This work was supported by DGICYT (Spain) Grant PB87-0351 (to P.A.). C.R. is the recipient of a fellowship from the Fundacion Cientifica de la Asociacion Espaiiola Contra el Cancer. The authors are indebted to Ms. M. C. Granados and to Mr. P. Lastres for technical assistance, to Mr. E. Wiltshire for checking the English version of this manuscript, and to Mrs. M. 0. Partearroyo for secretarial work.
REFERENCES 1.
Kimech, CellRes.
N., Hill,
A-M.,
Seif, R., and Pantaloni,
D. (1986)
162,326-334.
2. Kruh, J. (1982) Mol. Cell. Biochem. 42,65-82. 3. Xue, S., and Rao, P. N. (1981) J. Cell Sci. 51,163-171. 4. Chabanas, A., Khoury, E., Goeltz, P., Froussard, P., Gjerset, Dod, B., Eisen, 141-151.
H., and Lawrence,
5. Lanzagorta, Plant
6. Yuan, Chem.
J. M. A., De la Torre, 72,775-781. Z. A., Hirschhorn, 260,3778-3783.
J. J. (1985) C., and Aller,
R. R., and Baserga,
7. Aller,
P., Hirschhorn, l&119-127.
R. R., and Rius,
J. Mol.
Samstrom,
12.
Hozumi,
13.
Collins,
S. J. (1987)
14.
Muller,
R. (1986)
Received Revised
D. J., Chick, 7,560-563.
C., andNilsson, M. (1983)
August 15,1989 version received
16.
Bernal, S. D., and Chen, L. B. (1982) Cancer Collins, S. J., Gallo, R. C., and Gallagher, (London) 270,347-349.
17. 18.
19.
Mol.
deh, B. N. (1987) J. Biol. Chem. 262,4104-4108. H., and Skoultchi, A. (1984) Nature
29. Lachman, 592-594. 30. Watanabe,
Ada Blood TZBS
K. (1976) Cancer
70,1233-1244.
11,129-132.
December
Znt. J. Cancer
Res. 38,121-169.
27,1989
17,565-577.
J.
T., Sherman, J. Cell. Physiol.
M.,
Shafman,
fLondon,J
T., Iwata,
310,
T., and Kufe,
127,480-484.
31. Rittling,
Znt.
J. F.
M. (1986)
Euol. 23,11-22.
24. Feinberg, B. P., and Vogelstein, B. (1984) Anal. Biochem. 137, 266-267. 25. Aller, P., and Baserga, R. (1986) J. Cell Physiol. 128,362-366. 26. Lehrach, H., Diamond, D., Wozney, M., and Boedtke, H. (1977) Biochemistry 16,4743-4749. 27. Hirschhorn, R. R., Aller, P., Yuan, Z. A., Gibson, C. W., andBaserga, R. (1984) Proc. Natl. Acad. Sci. USA 81,6004-6008. 28. Simpson, R. U., Hsu, T., Begley, D. A., Mitchell, B. S., and Aliza-
34.
Habener,
S., Franchini, G., F. (1982) Proc. Natl.
22. Rius, C., and Aller, P. (1989) Cell Differ. 28,39-46. 23. Alonso, S., Minty, A., Bourlet, Y ., and Buckingham,
J. Biol.
S. (1980)
F., Ter-
l&1142-
Sanchez-Madrid,F., Acevedo, A.,Bellon,T., Fern&nLarraga, V., and Bernabeu, C. (1988) Hybridoma 7,
R., Gelmann, E. P., Martinotti, Papas, T. S., Gallo, R. C., and Wong-Staal, Acad. Sci. USA 79,6497-6501.
R. (1985)
W. L., and
Cabaiias,C., dez, J. M., 167-175.
21. Dalla Favera,
32. 33.
Biochem.
Res. 42,5106-5116. R. E. (1977) Nature
S., Battini, R., Kaczmarek, L., Rittling, S., Calabretta, B., De Riel, J. K., Philiponis, V., Wei, J. F., and Baserga, R. (1986) Mol. Cell. Biol. 6,3614-3620.
Physiol.
C. (1989)
D., and Bronet,
Keizer, G. D., Birst, J., Figdor, C. G., Spits, H., Midema, host, G., and De Vries, J. E. (1985) Eur. J. Zmmunol. 1148.
D. (1986) R.,
G., Paulin,
20. Ferrari,
35.
Philippe, J., Drucker, (1987) Mol. Cell Biol.
11.
Dellagi, K., Vaincheuker, W., Vinci, J. C. (1983) EMBO J. 2,1509-1514.
183,
Biol.
ALLER
15.
P. (1988)
8. Leder, A., and Leder, P. (1975) Cell 5,319-322. 9. Koshihara, Y., Senshu, T., Kawamura, M., and Murota, Biochim. Biophys. Acta 617,536-539. 10.
Exp.
AND
36.
S. R., Gibson, G. W., Ferrari, S., and Baserga, R. (1985) Biochem. Biophys. Res. Commun. 132,327-335. Boyd, A., and Metcalf, D. (1984) Leuk. Res. 8,27. Wattanabe, T., Sariban, E., Mitchell, T., and Kufe, D. (1985) Biochem. Biophys. Res. Commun. 126,999-1005. Trepel, J. B., Calamoni, 0. R., Kelly, K., Schwab, G., Watt, R. A., Sansville, E. A., Jaffle, E. S., and Neckers, L. M. (1987) Mol. Cell. Biol. 7,2644-2648. Kikkawa, U., and Nishizuka, Y. (1986) Annu. Rev. Cell Biol. 2, 149-178. Rittling, S. R., Coutinho, L., Amram, T., and Kolbe, M. (1989) Nucleic Acids Res. 17,1619-1632. Laszlo, A., and Bissell, M. J. (1983) Exp. Cell Res. 148,221-234.
37. 38. Luk,
(1982)
G. D., Civin, C. I., Weismann, Science 216,75-77.
R. M.,
and
Bayleu,
39. Ryan, M. P., and Higgins, P. J. (1988) J. Cell Physiol. 34. 40. Traub, P. (1985) Intermediate Filaments: A Review, Verlag,
Berlin.
S. B.
137, 25Springer
