Journal of Cell Science 101, 395-401 (1992)Printed in Great Britain © The Company of Biologists Limited 1992

395

Vimentin expression as a late event in the in vitro differentiation of human

promonocytic cells

CARLOS RIUS and PATRICK) ALLER*

Centro de Investigaciones Bioldgicas, CS1C Veldzquez 144, 28006 Madrid, Spain

•Author for correspondence

Summary

The administration of either 12-0-tetradecanoyl phor-bol-13-acetate (TPA, 3xl(T8M), sodium butyrate (SB,10~3M), Af6,2'-O-dibutyryladenosine-3':5'-cyclic mono-phosphate (dbcAMP, 10~3M), cytosine arabinoside(ara-C, 10~7M), amsacrine (mAMSA, 10~7M) orretinoic acid (RA, 10~6M) inhibits the growth activity ofhuman promonocytic U-937 cells, by arresting them atd or at the G,/S border (SB, RA, ara-C), at G2

(mAMSA) or at d and G2 (dbcAMP). All these agentstrigger cell differentiation, as proved by the increasedexpression of the maturation-associated CDllb andCD l ie surface antigens, and induce the expression of thevimentin gene at both the protein and the mRNA levels.TPA, SB and dbcAMP behave as "early" inducers, inthe sense that vimentin mRNA levels are rapidlyincreased (hour 6) upon drug administration. In con-trast, mAMSA and RA behave as "late" inducers, since

they do not increase vimentin mRNA levels until 48 to 72hours, following the stimulation of surface antigenexpression. The action of RA is characterized by aninitial inhibition period, in which the basal level ofvimentin mRNA is abolished (hour 24). Nevertheless,this RNA is later re-induced, to reach at 72 hours higherlevels than in untreated cells. Moreover, RA is capableof delaying the early induction of vimentin expressioncaused by TPA and SB, without affecting the normalexpression of differentiation markers. Taken together,these results strongly suggest that vimentin expression isnot required at the initial stages of promonocytic celldifferentiation, although it could play a role at anadvanced stage.

Key words: vimentin expression, differentiation,promonocytic cells.

Introduction

The intermediate-size filament protein vimentin isexpressed in cells of mesenchymal origin, as well as inmost cell types in culture. Vimentin behaves as a cellgrowth-dependent gene, in the sense that its expressionis rapidly stimulated when quiescent, Go-arrested cellsare induced to proliferate by serum or specific mitogens(Hirschhorn et al., 1984; Ferrari et al., 1986). Also,vimentin is a differentiation-dependent gene, since itsexpression is modulated during cell maturation, accord-ing to the cell lineage. For instance, in vivo studies usingnormal hemopoietic precursors showed that vimentin isconserved in mature granulocytes and monocytes,whereas it disappears in megakaryocytes and erythroidcells (Dellagi et al., 1983). In addition, in vitro studiesusing mammalian leukemia cell lines indicated thatvimentin expression is inhibited in murine (MEL) cellsdifferentiating along the erythroid pathway (Ngai et al.,1984), while it is increased when murine (Ml) or human(HL-60 and U-937) cells differentiate along the mono-cyte pathway (Rius and Aller, 1989; Hass et al., 1989;Tsuru et al., 1990). In the latter case, the observations

that vimentin is increased by diverse differentiationinducers such as the phorbol ester TPA (Hass et al.,1989), sodium butyrate (Rius et al., 1990) and con-ditioned medium from rat embryo fibroblasts (Tsuru etal., 1990), and that it reverts to basal levels when cells"retro-differentiate" upon withdrawal of the inducer(Hass et al., 1990), seem to indicate that vimentinexpression is functionally linked to the maturation ofpromonocytic cells. Moreover, vimentin expression israpidly stimulated upon treatment with those inducers,suggesting that this protein could play a role at theinitial stages of the differentiation process (Bernal andChen, 1982). Nevertheless, Taimi et al. (1990) haverecently reported that 1,25-dihydroxyvitamin D3 andretinoic acid induce monocyte-like properties in U-937cells in the absence of vimentin expression. Hence,vimentin stimulation could be a particular effect ofsome inducers, instead of being associated with differ-entiation itself.

The relationship between vimentin expression andpromonocyte cell differentiation is examined in thiswork by comparatively analyzing the action of severalmaturation inducers on U-937 human promonocytic

396 C. Rius and P. Alter

cells. The inducers used are: the phorbol ester TPA,sodium butyrate (SB), dibutyryl cyclic AMP(dbcAMP), cytosine arabinoside (ara-C), amsacrine(mAMSA) and retinoic acid (RA). The results indicatethat although these agents have different action mech-anisms at the molecular level and different effects onthe growth cycle, all of them stimulate vimentinexpression both at the protein and the mRNA levels.Nevertheless, the kinetics of vimentin induction varyvery much with the agent used, suggesting that thisprotein is not required at the early stages of thedifferentiation process.

Materials and methods

Cell growth and drug treatmentU-937 cells were grown in RPMI-1640 medium supplementedwith 10% heat-inactivated fetal calf serum, 0.2% sodiumbicarbonate and antibiotics, in a humidified 5% CO2atmosphere at 37°C. Cells were seeded in 100mm plasticdishes at approximately 2xlO~5 cells per millilitre andmaintained in continuous logarithmic growth by passing themevery 2 or 3 days. TPA, dbcAMP, ara-C and RA wereobtained from Sigma Chemical Co. (St. Louis, MO, USA).SB was obtained from Merck (Darmstadt, FRG). m-AMSA(Lamasine) and a--difiuoromethylornithine (DFMO) weregenerous gifts from Bristol Myers, SAE (Spain) and theMerrel Dow Research Institute (Strasbourg, France), respect-ively. Stock solutions of TPA (1.5xlO"3M), mAMSA(10"3M) and RA (10"2M) were prepared in DMSO, RPMIand absolute ethanol, respectively, and kept at —20°C. At thefinal concentrations used, DMSO and ethanol were withoutsignificant effects on cell growth and differentiation. Stocksolutions of dbcAMP (KT?M), SB (HT'M), ara-C (10~5M)and DFMO (10-1M) were freshly prepared in RPMI justbefore their application. Cells were seeded for experiments,in a mixture of conditioned medium and fresh medium(approximately 1/3, v/v). Cells from control cultures wereharvested at the logarithmic growth phase. Cells from treatedcultures were harvested at the times indicated in eachexperiment. Cell growth and viability were checked by using ahemocytometer and by trypan blue exclusion, respectively.

Flow cytometry determinationsTo measure DNA content, cells were incubated for 5 to 15minin RPMI medium containing 0.05% (w/v) NP40 and 50/ig/mlpropidium iodide. DNA profiles were then generated with anEPICS-CS flow cytometer (Coulter Cientffica, M6stoles,Spain), with an argon laser operated at 200mV and excitationwavelength of 488nm. Five thousand cells were scored foreach profile.

The surface expression of the leukocyte integrins CR3(CDllb/CD18) and pl50,95 (CDllc/CD18) was analyzed bymeasuring reactivity with the mAbs Bear 1 (anti-CDllb) andHCl/l (anti-CDllc). These antibodies were obtained byimmunization of Balb/c mice with human monocytes in thecase of Bear 1 (Keizer et al., 1985), and with TPA-stimulatedcells, in the case of HCl/l (Cabanas et al., 1988). Cells werefirst incubated for 30min at 4°C with the mAb, then washedwith RPMI medium and incubated again under the sameconditions with FTTC-labelled sheep anti-mouse IgG (Amer-sham, Buckinghamshire, UK). Some cell samples wereincubated only with this second antibody to determine thebackground of non-specific fluorescence. After washing the

cells with RPMI, their fluorescence was estimated by flowcytometry, as above.

To measure the cytoskeleton-associated vimentin, cellswere fixed for 5min at room temperature with 3.7% (v/v)formaldehyde, in a buffer containing 2mM MgCl2, lmMEGTA and lOOmM PIPES, pH 6.8, and then permeabilizedfor 5min with 0.2% (v/v) Triton X-100, 3.7% (v/v) formal-dehyde, in the same buffer. After washing with RPMImedium, the cells were incubated for 45min at 37°C with amouse anti-vimentin mAb (Amersham), washed again, andincubated under the same conditions with FITC-labeled sheepanti-mouse IgG. The cells were washed with RPMI mediumand either examined by flow cytometry, as described above,or mounted on glass slides and examined by fluorescencemicroscopy.

RNA blot assaysTotal cytoplasmic RNA was prepared as described previously(Aller and Baserga, 1986). RNA samples (15/ig) weredenaturated, electrophoresed in 1.1% agarose-formaldehydegels (Lehrach et al., 1977) and blotted onto nylon membranes(Hybond-N, Amersham). RNA blots were prehybridized,hybridized with excess 32P-labeled probes, washed underhighly stringent conditions (Hirschhorn et al., 1984) andfinally autoradiographed. The probes used were: the 1.1kbhuman vimentin-specific Xhol fragment of p4Fl plasmid(Ferrari et al., 1986) and the 0.66kb mouse ^-actin-specificKpnl/Bgtll fragment of pAL41 plasmid (Alonso et al., 1986).The fragments were labeled to approximately 109cts/min perjug of DNA with [a-32P]dCTP (3000Ci/mmol, New EnglandNuclear, Boston, MA, USA) by random hexanucleotidepriming (Feinberg and Vogelstein, 1984).

Results

Cell growth and differentiationWe first determined the action of SB (1(T3M),dbcAMP (KT3M), ara-C (1(T7M), mAMSA (1(T7M),RA (1(T6M) and DFMO (5xl(T3M) on the growth ofU-937 cells. Some of the results obtained are rep-resented in Fig. 1. All treatments inhibited theproliferation activity to a similar extent (Fig. la), butthey had different effects on cell cycle distribution (Fig.lb). In fact, while SB-, RA- and DFMO-treated cellswere arrested at Gi, mAMSA-treated cells werearrested at G2, and dbcAMP blocked cells at both Giand G2. ara-C is believed to block cells at the Gi/Sborder or at very early S-phase (Tobey and Crismann,1972), but this could not be distinguished from the d -blockade by the technique employed here (Fig. lb).

We afterwards determined the capacity of thesegrowth inhibitor agents to induce cell differentiation.This was carried out by measuring the surface ex-pression of CDllb and CDllc antigens. CDllb andCDllc are the a- subunits of the leukocyte adhesionmolecules CR3 (CDllb/CD18) and pl50,95(CDllc/CD18), respectively (Arnaout, 1990), the ex-pression of which has been found to be increased duringthe in vitro differentiation of human myeloid cells(Keizer et al., 1985; Cabaflas et al., 1988; Dudley et al.,1989; Rosmarin et al., 1989). We also examined theaction of the phorbol ester TPA, a potent maturationinducer of these cells. With the exception of DFMO, all

Vimentin expression in promonocytic cells 397

3, -•-Control -o-Treated

1 2 3 4 1 2 3 4Days of treatment

—. ioo-

4)

gQ.

60

un I• CD11b

I I I

-Treated

dbcAMP

mAMSA

Fluorescence(DNA content)

CD11c

ITPA SB dbcAMP araC m-MSA RA DFMO

Fig. 1. Growth anddifferentiation of U-937 cells.(a) Increase in cell number inuntreated (control) cultures andin cultures treated with theindicated agents, as determinedby hemocytometer counting ofcells which exclude trypan blue.(b) Cycle distribution ofuntreated cells and cells treatedfor 3 days with the indicatedagents, as determined by flowcytometry analyses of propidiumiodide-stained nuclei.(c) Frequency of cells expressingthe differentiation-specificCDllb and CDllc surfaceantigens in untreated cultures(C) and in cells treated for 2days with TPA, for 4 days withDFMO, and for 3 days with theother indicated agents. Thereactivity with the mAbs Bear 1(anti-CDllb) and HCl/l (anti-CDllc) was determined byindirect immunofluorescencecombined with flow cytometry.The values represent netpercentage of positive cells,obtained by subtracting thebackground of non-specificfluorescence. The concentrationsused were: 3xlO~8M TPA;10"3M SB and dbcAMP; 1(T7Mara-C and mAMSA; 10"6M RA;5xlO"3M DFMO.

agents increased the number of cells expressing CDllband CDllc (Fig. lc), suggesting that they trigger thefunctional maturation of U-937 cells. This conclusionwas cytochemically confirmed by observing an increasein the capacity of the treated cells to reduce nitrobluetetrazolium (results not shown). The inability of DFMOto induce the expression of differentiation markers inU-937 cells is consistent with results obtained earlierusing the human myelomonocytic HL-60 cell line(Sugiuraet al., 1984).

Vimentin expressionThe effect of the assayed agents on vimentin expressionwas first determined at the protein level by using aspecific anti-vimentin antibody and examining the cellsby fluorescence microscopy. Some of the resultsobtained are represented in Fig. 2. The fraction of cellswith detectable amounts of vimentin, which was below5% in untreated cells, increased upon treatment withTPA (2 days), SB, dbcAMP, ara-C, mAMSA (3 days)and RA (4 days) to values ranging from 40% (ara-C) to80% (TPA). DFMO, which did not induce differen-tiation, also failed to stimulate vimentin expression.

Then, we attempted to determine vimentin ex-pression at the mRNA level. For this purpose,Northern blot assays were carried out using RNA

extracted from either untreated cells or cells treated fordifferent periods of time with the indicated agents. Theresults in Fig. 3 show that all differentiation inducersaugmented the vimentin mRNA content. This up-regulation was not due to non-specific stimulation oftranscription, since the level of /3-actin mRNA was littlemodified, or even decreased at the latest times oftreatment (Fig. 3, and results not shown). Nevertheless,the kinetics of vimentin mRNA increase varied greatlywith the agent used. In fact, while a significantstimulation of this mRNA level was already observedupon 6 hours of treatment with TPA, SB or dbcAMP,stimulation was not observed until after 24 or 48 hoursof treatment with ara-C and mAMSA, respectively.The action of RA was characterized by an initialinhibition period, lasting for at least 24 hours, in whichthe basal vimentin mRNA level was completely abol-ished. Nevertheless, this RNA was later re-induced, toreach after 72 hours higher levels than in untreatedcells. This re-induction cannot be attributed to cell de-differentiation, caused by drug exhaustion or inacti-vation. In fact, over the 72 hours treatment periodstudied, RA increased progressively the expression ofdifferentiation-associated antigens (see Fig. 4), anddecreased progressively the growth activity.

A trivial explanation for the differences in thekinetics of vimentin expression is that they could be

398 C. Rius and P. Alter

Fig. 2. Changes in vimentin content. Followingformaldehyde fixation and Triton X-100 permeabilization,the reactivity of untreated cells (a) and of cells treated for2 days with TPA (b), for 3 days with dbc AMP (c) and for4 days with RA (d) with an anti-vimentin mAb wasexamined by fluorescence microscopy. Bar, 15/an.

reflecting differences in the capacity of the agents usedto trigger cell maturation. For this reason, we measuredthe time-course expression of the CD lib differen-tiation-associated antigen upon treatment with SB,mAMSA and RA. A significant increase in antigenexpression was first observed in all cases after approxi-mately 24 hours of treatment. This increase followedthe induction of vimentin mRNA by SB, but precededits induction by mAMSA and RA (Fig. 4). Based onthese observations, we can interpret vimentin mRNAinduction as an early event in the case of SB, and as alate event in the case of the mAMSA or RA-produceddifferentiation process.

Since RA transiently decreased vimentin RNA level,we wondered whether this agent could prevent therapid activation of the vimentin gene caused by otherinducers. With this purpose in mind, cells were treatedfor 12 and 24 hours with TPA or with SB, either in theabsence or in the presence of RA. We found that RAsuppressed or greatly reduced the stimulation ofvimentin expression produced by SB and TPA, as

coO

SB ( h ) a RA(h)?! 2 O

OCO (M

Vimentjn

B-Actin HIM" <•••

dbcAMP,o

mAMSA

P

6 12 24Hours of treatment

Fig. 3. Time course of changes in vimentin mRNA levels.Samples of total cytoplasmic RNA extracted fromuntreated cells (Control) and cells treated with theindicated agents were electrophoresed, blotted tomembranes, and sequentially hybridized with 32P-labelledvimentin and /S-actin probes. The autoradiogramcorresponding to vimentin in the RA experiment wasoverexposed, to show more clearly both the down-regulation and the up-regulation of this RNA. The valuesin the lower panel were obtained by densitometric readingsof RNA blots. Controls were arbitrarily given the value 1.

measured at the RNA level by Northern blot assays(Fig. 5a). No treatment significantly altered /J-actinmRNA level, which was measured as a control. Thisexcluded possible experimental artifacts such as differ-ences in sample loading in the gel (Fig. 5a). Suppressionby RA of most TPA- and SB-produced vimentininduction was also observed at the protein level byimmunofiuorescence combined with flow cytometry(Fig. 5b). However, RA did not prevent the inductionof the CDllb and CDllc antigen expression, observedupon 24 hours treatment with TPA or SB (Fig. 6). Also,RA did not prevent cell to cell adhesion nor cellattachment to the plate surface, which normally occursupon treatment of myeloid cells with TPA (Fig. 7).After prolonged periods of treatment, RA no longerinhibited, but instead potentiated the stimulation ofvimentin expression caused by the other inducers(results not shown).

60-

»40""5oo

Ino

Q.

SB

i

£

RA,mAMSA ^ 8

r^^ / — • — 5>B' / -o-mAMSA

( / -O-RA

6 12 24 48Hours of treatment

72

Fig. 4. Time course of changes in CDllb expression. Thefrequency of cells reacting with the mAb Bear 1 (anti-CDllb) was determined by immunofluorescence combinedwith flow cytometry. The arrows indicate the time at whichvimentin mRNA levels start to increase upon treatmentwith the indicated agents (see Fig. 3). The values representnet percentage of positive cells, obtained by subtracting thebackground of non-specific fluorescence.

Discussion

In vitro maturation of myeloid leukemia cells may betriggered by agents with different primary actionmechanisms. Among the agents used here, TPA andcyclic AMP derivatives seem to act through theactivation of protein kinases C (Ca2+-dependent) andA (cyclic AMP-dependent), respectively (Kikkawa andNishizuka, 1986; Krebs and Beavo, 1979). SB causeshistone hyperacetylation, which may result in theweakening of histone/DNA interactions and in theenhanced expression of some genes, including thosecritical for differentiation induction (Kruh, 1982). TheDNA replication inhibitor ara-C causes early termin-ation of nascent DNA chains (Kufe et al., 1980), whilethe topoisomerase II inhibitor mAMSA provokes DNAbreakage (Nelson et al., 1984). Thus, induction ofdifferentiation by ara-C and mAMSA may represent acellular response to alterations in the integrity and thenormal functioning of the genome (Bloch, 1989). Inaddition to the diversity in their primary actionmechanisms, these agents display different effects onthe cell cycle, since the cells are arrested at Gt by TPAand SB, at G2 by mAMSA, at Gi and G2 by dbcAMP,and probably at the G-JS border by ara-C (Tobey andCrismann, 1972; Yen et al., 1987; and results in thiswork). The observation that, in spite of all thesedifferences, the assayed inducers stimulate withoutexception the expression of the vimentin gene, stronglyindicates that vimentin expression is somewhat linkedto differentiation itself. Moreover, vimentin is notincreased by DFMO, which inhibits cell growth withoutprovoking differentiation, indicating that its inductionis not a mere consequence of proliferation arrest.

Differentiation is a multistep process, in which eachstage is associated with the expression of specific genes(Davis et al., 1987). The observation that TPA rapidlystimulated vimentin expression (Bernal and Chen,

Vimentin expression in promonocytic cells 399

a 2 TPA SB12h 24 h 12h

o - + -24 h

- +

Vimentin

B-Actin

I

SB. 24h

Fluorescence ( log )

Control Treatment

Fig. 5. The effect of RA on the TPA and SB-producedactivation of vimentin expression. The upper panel (a)shows the relative vimentin and /3-actin mRNA levels inuntreated cells (Control) and in cells treated with TPA orwith SB either in the absence (—) or in the presence (+)of RA for 12 or 24 hours, as determined by Northern blotassays using total cytoplasmic RNA samples. The lowerpanel (b) shows changes in vimentin protein levels as aconsequence of cell treatment with TPA and with SB,either alone or in combination with RA. Afterformaldehyde fixation and Triton X-100 permeabilization,cell reactivities with an anti-vimentin mAb weredetermined by immunofluorescence combined with flowcytometry. The broken vertical lines delimit the regionscorresponding to non-specific fluorescence (—) which weregiven by the distribution of cells incubated only with thesecond antibody (FTTC-labelled anti-IgG).

1982), a fact later corroborated with other inducers(Rius et al., 1990; Tsuru et al., 1990), led to theproposition that vimentin could play a role at the earlystages of myeloid cell differentiation (Bernal and Chen,1982). Although "early" and "late" are rather impre-cise terms, since they depend much on the criteria usedto examine differentiation, the results in the presentwork seem to argue against this proposition. In fact,vimentin induction was not detected until after 48 to 72hours of treatment with mAMSA and RA, followingthe activation of differentiation-specific antigens.Moreover, the activation of vimentin expression by theearly inducers TPA and SB was delayed when RA wasadded together with these agents, without affectingthe timing of expression of differentiation markers.Hence, it appears that whatever the time of induction,

400 C. Rius and P. Aller

100-

w 20HoQ. n

CD11b CD11c

Fig. 6. The effect of RA on the TPA and SB-producedactivation of CDllb and CDllc expression. The frequencyof cells reacting with the mAbs Bear 1 (anti-CDlib) andHCl/l (anti-CDllc) in untreated cultures (C) and incultures treated for 24 hours with TPA or with SB, eitheralone (—) or in the presence (+) of RA, was determinedby immunofluorescence combined with flow cytometry. Thevalues represent net percentage of positive cells, obtainedby subtracting, the background of non-specificfluorescence.

vimentin expression is not required at the initial stagesof promonocytic cell maturation.

RA was the only assayed inducer capable ofdecreasing vimentin mRNA levels in U-937 cells.Down-regulation of this RNA by RA was earlierreported by Ferrari et al. (1986) and by us (Rius andAller, 1989) in human myelomonocytic HL-60 cells. RAseems to exert its biological action through interactionwith specific receptors which, upon internalization,behave as nuclear transcription factors (Collins et al.,1990). Since transcription of the vimentin gene is underthe control of both positive and negative "cw-acting"regulatory DNA sequences (Rittling and Baserga,1987), inhibition of vimentin expression could be theresult of specific binding of internalized RA receptorsto the negative regulatory sequences. Nevertheless,following the initial decrease, vimentin mRNA wasreinduced in RA-treated U-937 cells, reaching higherlevels at 72 hours than in untreated cells. This contrastswith the result in RA-treated HL-60 cells, where thistranscript remained permanently inhibited (Ferrari etal., 1986; Rius and Aller, 1989). RA induces monocyte-like maturation of U-937 cells (Taimi et al., 1990),whereas it causes granulocyte-like maturation of thebipotent HL-60 cells (Breitman et al., 1980). Hence, thedifferent behaviour of vimentin expression in these celllines might indicate that this protein is somewhatinvolved in cell maturation along the monocyte path-way but not along the granulocyte one. Nevertheless,Taimi et al. (1990) reported that RA confers monocyte-like properties to U-937 cells in the absence of vimentin

Fig. 7. Intercellular adhesion of U-937 cells. Phase contrastphotomicrographs of untreated cells (a) and cells treatedfor 24 hours with RA (b), with TPA (c) and with RA plusTPA (d). Bar, lOO^m.

expression, measured at the protein level. This obser-vation contrasts with the results obtained by us (Fig.2d). We do not know the reasons for such a discrepancy- whether it is due to differences in the appliedmethodology or in properties inherent to the cell clonesused - but it imposes obvious caution when proposing arole for vimentin on cell differentiation.

In summary, the results in this work indicate thatvimentin expression is not required to initiate thedifferentiation of human promonocytic cells, althoughthis protein might have a function at a later stage of thematuration process. Although the physiological signifi-cance of the intermediate filaments is not well under-stood, and different functions have been attributed tothese proteins (for reviews see Traub, 1985; Klym-kowsky et al., 1989), one of their major roles seems tobe to participate in the mechanical integration of theintracellular space. The acquisition of the differentiatedphenotype often involves changes in cell morphology,such as cell attachment and spreading (TPA), cell to celladhesion (TPA and, to some extent, RA), growth insize (ara-C, mAMSA) and adoption of elongated orother irregular shapes (SB, dbcAMP, ara-C, mAMSA).Increased amounts of vimentin could then be needed,as a part of the cytoskeleton reorganization, in order tosustain these changes.

This work was supported by DGICYT (Spain) Grant PB87-0351. The authors are indebted to Drs. C. Bernabeu and C.Cabanas for gifts of Bear 1 and HCl/l mAbs, to Mr. P.Lastres for technical assistance, and to Mr. E. Wiltshire andMs. O. Partearroyo for helping with the preparation of themanuscript.

Vimentin expression in promonocytic cells 401

References

AUer, P. and Baserga, R. (1986). Selective increase of c-myc mRNAlevels by methylglyoxal bis (guanylhydrazone) and novobiocin inserum-stimulated fibroblasts. / . Cell. Physiol. 128, 362-366.

Alonso, S., Minty, A., Borlet, Y. and Buckingham, M. (1986).Comparison of three actin-coding sequences in the mouse:evolutionary relationships between actin genes of warmbloodedvertebrates. / . Mol. Evol. 23, 1-22.

Arnaout, M. A. (1990). Structure and function of the leukocyteadhesion molecules CD11/CD18. Blood 75, 1037-1050.

Bernal, S. D. and Chen, L. B. (1982). Induction of cytoskeleton-associated proteins during differentiation of human myeloidleukemia cell lines. Cancer Res. 42, 5106-5116.

Bloch, A. (1989). Induction of tumor cell differentiation as a commonmechanism of action of DNA-specific antitumor agents. InAnticancer Drugs (ed. H. Tapiero, J. Robert and T. J. Lampidis),pp. 145-151. London, Paris: John Libbey Eurotext.

Breitman, T., Selonick, S. and Collins, S. (1980). Induction ofdifferentiation of the human promyelocytic leukemia cell line (HL-60) by retinoic acid. Proc. Nat. Acad. Sci. U.S.A. 77, 2936-2940.

Cabanas, C , Sanchez-Madrid, F., Acevedo, A., BeUon, T.,Fernandez, J. M., Larraga, V. and Bernabeu, C. (1988).Characterization of a CDllc-reactive monoclonal antiboy (HC1/1)obtained by immunizing with phorbol ester differentiated U-937cells. Hybridoma 7, 167-176.

Collins, S. J., Robertson, K. A. and Mueller, L. (1990). Retinoic acid-induced granulocytic differentiation of HL-60 myeloid leukemiacells is mediated directly through the retinoic acid receptor (RAR-a). Mol. Cell. Biol. 10, 2154-2163.

Davis, R. C , Thomason, A. R., Fuller, M. L., Sloving, J. P., Chou, C.C , Chada, S., Gatti, R. A. and Salser, W. A. (1987). mRNAspecies regulated during the differentiation of HL-60 cells tomacrophages and neutrophiles. Develop. Biol. 119, 164-174.

Dellagi, K., Vainchenker, W., Vinci, G., Paulin, D. and Brouet, J. C.(1983). Alteration of vimentin intermediate filament expressionduring differentiation of human hemopoietic cells. EMBO J. 2,1509-1514.

Dudley, D., Baker D. M., Hlckey, M. J. and Hkkstein, D. D. (1989).Expression of surface antigen and mRNA for the CDl lc (ox, pl50)subunit of the human leukocyte adherence receptor family inhematopoietic cells. Biochem. Biophys. Res. Commun. 160, 346-353.

Feinberg, B. P. and Vog«lstein, B. (1984). A technique forradiolabelling DNA restriction endonuclease fragments to highspecific activity. Anal. Biochem. 137, 266-267.

Ferrari, S., Battinl, R., Kazmarek, L., Rittling, S., Calabretta, B., DeRiel, J. K., Phlliponis, V., Wei, J. F. and Baserga, R. (1986).Coding sequence and growth regulation of the human vimentingene. Mol. Cell. Biol. 6, 3614-3620.

Hass, R., Bartels, H., Topley, N., Hadam, M., Kohler, L., Goppelt-Strflbe, M. and Resch, K. (1989). TPA-induced differentiation andadhesion of U-937 cells: changes in ultrastructure, cytoskeletalorganization and expression of cell surface antigens. Eur. J. CellBiol. 48, 282-293.

Hass, R., Giese, G., Meyer, G., Hartmann, A., Dork, T., Kohler, L.,Resch, K., Traub, P. and Goppelt-Strflbe, M. (1990).Differentiation and retrodifferentiation of U937 cells: reversibleinduction and suppression of intermediate filament proteinsynthesis. Eur. J. Cell Biol. 51, 265-271.

Hirschhorn, R. R., AUer, P., Yuan, Z. A., Gibson, C. W. andBaserga, R. (1984). Cell-cycle-specific cDNAs from mammaliancells temperature-sensitive for growth. Proc. Nat. Acad. Sci.U.S.A. 81, 6004-6008.

Keizer, G. D., Borst, J., Figdor, C. G., Spits, H., Mledema, F.,Terhorst, C. and De Vries, J. E. (1985). Biochemical and functionalcharacteristics of the human leukocyte membrane antigen familyLFA-1, Mo-1 and pl50,95. Eur. J. Immunol. 15, 1142-1148.

Kikkawa, U. and Nishiznka, Y. (1986). The role of protein kinase Con transmembrane signalling. Annu. Rev. Cell Biol. 2, 149-178.

Klymkowsky, M. W., Bachant, J. B. and Domingo, A. (1989).Functions of intermediate filaments. Cell Modi. Cytoskel. 14, 309-331.

Krebs, E. C. and Beavo, J. A. (1979). Phosphorylation-dephosphorylation of enzymes. Annu. Rev. Biochem. 48, 923-959.

Kruh, J. (1982). Effects of sodium butyrate, a new pharmacologicalagent, on cells in culture. Mol. Celt. Biochem. 42, 65-82.

Kufe, D., Major, P., Egan, E. and Beardsley, G. (1980). Correlationof cytotoxicity with incorporation of ara-C into DNA. J. Biol.Chem. 255, 8997-9000.

Lehrach, H., Diamond, D., Wozney, M. and Boedtke, H. (1977).RNA molecular weight determination by gel electrophoresis underdenaturing conditions: a critical re-examination. Biochemistry 16,4743-4749.

Nelson, E. M., Tewey, K. M. and Liu, L. (1984). Mechanism ofantitumour drugs: poisoning of mammalian DNA topoisomerase IIon DNA by 4'-(9-acridinylamino)-methanesulfon-m-anisidido.Proc. Nat. Acad. Sci. USA 75, 4838-4842.

Ngai, J., Capetakani, Y. G. and Lazarides, E. (1984). Differentiationof murine erythroleukemia cells results in the rapid repression ofvimentin gene expression. /. Cell Biol. 99, 306-314.

Rittling, S. R. and Baserga, R. (1987). Functional analysis and growthfactor regulation of vimentin promoter. Mol. Cell Biol. 7, 3908-3915.

Rius, C. and Aller, P. (1989). Modulation of ornithine decarboxylasegene transcript levels by differentiation inducers in humanpromyelocytic leukemia HL-60 cells. Cell Differ. Dev. 28, 39-46.

Rius, C , Cabanas, C. and Aller, P. (1990). The induction of vimentingene expression by sodium butyrate in human promonocyticleukemia U937 cells. Exp. Cell Res. 188, 129-134.

Rosmarin, A. G., Weil, S. C , Rosner, G. L., Griffin, J. D., Arnaout,M. A. and Tenen, D. G. (1989). Differential expression ofCDllb/CD18 (Mol) and myeloperoxidase genes during myeloidcell differentiation. Blood 73, 131-136.

Sugiura, M., Shafman, T., Mitchell, T., Grifln, J. and Kufe, D.(1984). Involvement of spermidine in proliferation anddifferentiation of human promyelocytic leukemia cells. Blood 63,1153-1158.

Taimi, M., Chateau, M. T., Marti, J. and Pacaud, M. (1990).Induction of differentiation of the human histiocytic lymphoma cellline U937 in the absence of vimentin expression. Differentiation 45,55-60.

Tobey, R. A. and Crismann, H. A. (1972). Use of flowmicrofluorometry in detailed analysis of effects of chemical agentson cell cycle progression. Cancer Res. 32, 2726-2732.

Traub, P. (1985). Intermediate Filaments: A Review. Berlin,Heidelberg, New York, Tokyo: Springer-Verlag.

Tsuru, A., Nakamura, N., Takayama, E., Suzuki, Y., Hlrayoshl, K.and Negata, K. (1990). Regulation of the expression of vimentingene during the differentiation of mouse myeloid leukemia cells. / .Cell Biol. 110, 1655-1664.

Yen, A., Brown, D. and Fishbaugh, J. (1987). Control of HL-60monocytic differentiation. Different pathways and uncoupledexpression of differentiation markers. Exp. Cell Res. 168, 243-254.

(Received 12 July 1991 - Accepted 30 October 1991)