Steel factor (SF), also referred to as Kit ligand, stem cell factor, or mast cell growth factor, is essential for the development of hematopoietic stem cells in vivo. It is.
THEJOURNAL OF B I O ~ I C CHEMISTRK AL 0 1994 by The American Society for Biochemistry and Molecular’ Biology, Inc.
Vol. 269,No. 16,Issue ofApril 22,pp. 12084-12091,
1994 Printed in U.S.A.
Product of the Steel Locus Suppresses Apoptosis in Hemopoietic Cells COMPARISON WITH PATHWAYS ACTTVATED BY GRANULOCYTE MACROPHAGE COLONY-STIMULATING FACTOR* (Received for publication, September 27, 1993, and in revised form, January 31, 1994)
Julio Caceres-CortBsSP,Daniel Rajottem, Julie DumouchelS, Pierre Haddadll, and “rang Hoang$ll** From the $Clinical Research Institute of Montreal and the Department of Pharmacology, Universityof Montreal, Montreal, Quebec HZW lR7, Canada
Steel factor (SF), also referred to as Kit ligand, stem stem cell populations: hematopoietic stem cells, melanoblasts, cell factor, or mast cell growth factor, is essential for the and germcells. Dominant white spotting has been shown to be development of hematopoietic stem cells in vivo. It is allelic with the gene encoding the tyrosine kinase receptor cshown here that SF is mainly a survival factor for hemo- Kit, whereas Steellocus encodes the corresponding growth facpoietic cells with little if any proliferative effect. In con- tor (reviewed in Ref. 1).Steel factor (SF),’ also referred to as trast, granulocyte macrophage colony-stimulating fac- stem cell factor, Kit ligand, or mast cell growth factor is protor (GM-CSF) acts both as a survival factor and as a duced as a soluble factor of 18.5-kDa molecular mass in human potent growth factor. We have probed the pathways acor a membrane bound form generated through tivated by SF and GM-CSF in suppression of active cell (33 kDa in rat), death (apoptosis) using two classes of inhibitors: Tyr- alternative splicing (2). In hemopoietic cells, SF has been phostins that arespecific inhibitors of protein tyrosine shown to act synergistically with granulocytemacrophage kinase, and amiloride derivatives (5-(N-ethyl-n-isopro- colony-stimulatingfactor(GM-CSF), interleukin-3 (IL-3), or py1)amiloride and 5-(N,N-hexamethylene)amiloride) erythropoietin to support the growthof primitive multipotent that have been designed as specific inhibitors of the and erythroid progenitors (3, 4). Our previous data indicate Na+/H+antiporter. Both SF-dependent and GM-CSF-de- that soluble SF can replace the requirementfor cell interaction pendent pathways are sensitive to inhibition by Tyrin primary cultures of human myeloid leukemic cells (5). In phostins with, nonetheless, a quantitativedifference.All primordial germ cells, SF appears to alleviate the requirement Tyrphostins tested are more potent inhibitors of c-Kit in feeder cells for the initial survival of primordial germ cells in than of GM-CSF receptor triggered pathways, the most culture ( 6 , 7). striking being Tyrphostin B42 that is10 times more poApoptosis is a process of active cell death (ACD)implicated in tent. In contrast to thediscrepancy in Tyrphostin dose- the regulationof cell numbers inneurogenesis (8), in lymphoid response curves, titration curves for 5-(N-ethyl-n-iso- cell development and tolerance, and inhemopoiesis (reviewed propy1)amiloride and 5-(N,N-hexamethylene)amiloride are comparable in SF- or GM-CSF-stimulated cells. Fur- in Refs. 9 and 10). ACD is characterizedby the appearanceof a thermore, SF induces a rapid and sustainedalkaliniza- typical DNA ladder, due to thecleavage of cellular DNA in the tion of the intracellular pH, as assessed with the pH- internucleosomal space by an activated endonuclease. While sensitive probe 2’,7’-bis(2-carboxyethyl)-5-carboxy-the molecular pathways that control ACD in mammalian cells still remain tobe uncovered, it appears that ACD is regulated fluorescein. Taken together, our data indicate that input from two distinct pathways with discrepancy in imme- by trophic factors in both neurons (11,12) andhemopoietic cells diate early events, that of c-Kit and GM-CSF receptor, (9). Four hemopoietic growth factors, GM-CSF, IL-3, G-CSF, results in a common output, activation of the Na+/€I+ an- and erythropoietin, have been shown to maintain hemopoietic tiporter and suppression of apoptosis by the two li- cell survival through the retardation of DNA breakdown (13gands. 15). Previous datafrom our laboratory (16)and elsewhere (17) indicate that the suppression ofACDby GM-CSF and IL-3 requires protein kinase C and a functional Na+/H+ antiporter. In contrast to these two cytokines that also act as potent miMutations in both the Steel locus (Sl) and the “dominant white spotting“ locus result in developmental defects in three togenic signals for hemopoietic precursors, we report herein that SF is mainlya trophic factor in hemopoiesis. Furthermore, * This work was supported in part by grants from the National Can- we have used inhibitors of tyrosine kinase (Tyrphostins) (18) cer Institute of Canada, with funds fromthe Canadian Cancer Society, and theNa+/H+antiporter (amiloride derivatives) (19)to probe the Medical Research Council of Canada, and the Leukemia Research the pathways activated by SF and GM-CSF in suppression of Fund. The costs of publication of this article were defrayed in part by ACD. Our data indicate that all five Tyrphostins tested are the payment of page charges. This article must therefore be hereby marked “advertisement” in accordancewith 18 U.S.C.Section 1734 more potent inhibitorsof SF-dependent pathways thanof GMCSF dependent pathways. We have specifically identified Tyrsolely to indicate this fact. 5 Supported by a studentshipfrom the National Autonomous Univer- phostin B42 as a preferential inhibitor of SF-induced suppressity of Mexico and the University of Montreal. Current address: Laboratorio de Diferenciacion Celular y Cancer, FES-ZARAGOZA,U N A M , Theabbreviations used are: SF,Steel factor; IL-3, interleukin-3; Apartado 9020, Mexico D.F., CP 15000, Mexico. 1 Supported by a studentship from the Cancer Research Society Inc. ACD, active cell death; GM-CSF, granulocyte macrophage colony-stimu** Scholar of the National Cancer Institute of Canada. To whom cor- lating factor; AML, acute myeloblastic leukemia; EIPA, 54N-ethyl-nrespondence should be addressed: Laboratoryof Hemopoiesis and Leu- isopropy1)amiloride; HMA, 5-(N,N-hexamethylene)amiloride;BCECF, MTT, 3-(4,5-dimethylkemia, Clinical Research Institute of Montreal, 110 Avenue des Pins 2’,7’-bis(2-carboxyethyl)-5-carboxyfluorescein; thiazoyl-2-yl)-2,5-diphenyltetrazolium bromide. Ouest, Montreal, Quebec, Canada H2W 1R7. Fax: 514-987-5688.
12084
Suppression of Apoptosis by SF sion of ACD. In contrast to thedivergence in Tyrphostin doseresponse curves between SF-dependent and GM-CSFdependent pathways, dose-response curves for amiloride derivatives areidentical, indicating that the two pathways converge towards a proteinkinase C-dependentactivation of Na+/H+ exchange.
12085
O D (570)
MATERIALS AND METHODS Cells and Growth Factors-The cell line TF-1 was a kind gift of Dr. Toshio Kitamura (DNAX, Seattle) (20) and the cell line M07-E, of Drs. G. C. Avanzi (21) and S. C. Clark (Genetics Institute, Cambridge, Ma). Both cell lines were seeded three times weekly, TF-1 a t 105/ml,and M07-E at 3 x 106/ml,in IMDM (Life Technologies Inc.) supplemented n with 10%fetal calf serum (Life TechnologiesInc.). TF-1 cells weremain0 24 48 72 96 tained in the presence of 200 PM GM-CSF, and M07-E, 10 unitdm1 of time in culture (h) IL-3. The bcr-ab1 transfected M07-E cell line was kindly provided by FIG.1. Kinetics of cell survival and proliferation in the presDrs. J. Dick and C. Sirard (Hospital for SickChildren, Toronto, Ontario, ence of SF or GM-CSF. TF-1 cells were exposed to SF (550 PM) or Canada). The cells are growth factor-independent and were maintained maintained in Iscove's medium supplemented with 10% fetal calf serum (22). Pri- GM-CSF (400 PM) for the indicated times. Control cells were mary myeloblasts were isolated from the peripheral blood of patients in culture medium in the absence of growth factors. Cell viability was with acute myeloblastic leukemia (AML) as described previously (5). determined by the MTI' assay as described under "Materials and Methods." Cells were kept frozen until use. Purified recombinant GM-CSF was a generous gift of Dr. S. C. Clark and purified recombinant SF, of Dr. K. Szebo (Amgen,Thousand Oaks, through centrifugation, and the cell pellet was resuspended in 100 plof acidified isopropyl alcohol. The solubilized MTT formazan was transCA). RecombinantIL-3 was produced in COS cells, after transient transfection with an IL-3 expression vector, kindly provided by Dr. K. ferred to microtiter plates and the optical density of each well was quantitated in an enzyme-linked immunosorbent assay plate reader Humphries (Terry Fox Laboratory, Vancouver, British Columbia). Chemicals-Tyrphostin 11 was generouslyprovided byDr. Chi (Bio-Rad)at 570 nm wavelength. Kuang Huang (University of Connecticut, Farmington, CT) (23). All other m h o s t i n s are from Calbiochem (La Jolla, CA). Staurosporine RESULTS (Kyowa Hakko USA, New York) (241, genistein (Sigma)(25),and the two SF Preferentially Maintains Cell Survival-TF-1 cells (lo5/ amiloride derivatives, 5-(N-ethyl-n-isopropyl)amiloride(EIPA) and ml) were exposed to GM-CSF alone, SF alone, or SF in combi5-(N,N-hexamethylene)amiloride(HMA)(Dr. E. Cragoe,Nacogdoches, nation with GM-CSF for the indicated times. Data shown in TX), were stored and prepared as described previously (16). Fig. 1indicate that, in theabsence of growth factors, the cells Analysis of DNA Fragmentation by Electrophoresis-TF-1 cells or primary myeloblasts (lo6)were incubated under different culture con- died after the first cells stimulated 24 hof culture. In contrast, ditions, as indicated. After 24 h, equal cell numbers were lysed in 20 pl with GM-CSF proliferated rapidly and continuously throughof Sarcosyl lysis buffer, as described previously(16). Proteins and RNAs out the culture period. Finally, cells that were exposed to SF were removedthrough two consecutivedigestions of 1h each at 50 "C in the presence of 0.5 mg/ml proteinase K and 0.15 mg/ml RNase A, re- doubled or remained at input levels. These data suggest that GM-CSF is a much more potent mitogen than SF. spectively.Samples were loadedon a 1.2% agarose gel (20 x 20 cm) and The viability of cells maintained with SF was further tested resolved by electrophoresis for 15 h at 40V. A positive photograph (Eagle Eye Model,Stratagene, La Jolla, CA) and anegative photograph in secondary cultures. Primary myeloblasts (AML41) or TF-1 (4 x 5 in,Kodak TechnicalPan film, Kodak, Rochester,N Y ) were taken cells were maintained under thedifferent culture conditions, as of the gel. Data were analyzed by scanning densitometry (model 620, above. After 3 days, recovery in viable cells was determined in Bio-Rad). Theareas under the curve were integrated and the percentcultures wasseeded age of DNA degradation was calculated as the area under the DNA one set of cultures, whereas the other ofset a t 105/ml in the presence of the same growth factors as in the ladder uersus the total area under the curve. primary cultures (Fig. 2). In addition, half of the cells exposed Analysis of DNA Integrity through Membrane Filtration of PHldThd Labeled DNA-Cells were incubated for 24 h in the presence of SF or to SF in the primary cultures were further stimulated with GM-CSF as indicated. At the initiation of culture, L3HldThd (20 Cii GM-CSF in thesecondary cultures. The pattern of cell recovery mmol, DuPont NEN) was added at a final concentration 3 pcilml. The was identical in the secondary cultures, ascompared to thatof cells wereretained on fiberglass filters which were washedsequentially the of the same with 12 ml of phosphate-buffered saline, 4 ml of trichloroacetic acid, and primary cultures, when maintained in presence 4 ml of methanol. Filters were air dried prior to liquid scintillation growth factors. Cultures exposed to GM-CSF showed the highcounting. Only high molecular weightDNA is retained, whereas faster est increase in viable cells, whereas cultures maintained with migrating bands are filtered through (26). SF remained at input levels. In contrast,cells exposed to SF in Intracellular p H Monitoring-Variations in pH, were monitored ona the primary cultures and stimulated with GM-CSF in the secSPEX model CMlTllI dual excitation spectrofluorometer as described ondary cultures proliferated to the same extent as cells that previously, using the fluorescent probe bis(carboxyethy1)carboxytluorescein (BCECF)(Molecular Probes, Eugene, OR). Aftera 30-min load- were continuously exposed to GM-CSF in both cultures. These ing period with 10 p~ BCECF, cells were monitored at 37 "C in a ther- data indicated thatcells maintained with SF remainedviable mostated cuvette under constant agitation. The ratio of the pH- and were capableto proliferate when challengedwith GM-CSF. sensitive fluorescence (excitation wavelength, 500 nm) and pH- Hence, SF ismainly a survivalfactor, whereas GM-CSF is both insensitive fluorescence (excitation wavelength 450 nm) allows for a a survival factor and a potent growth factor. calculation of pH, that is independent of cell number and dye concenSF Suppresses ACD in Primary Myeloblasts, M07-E, a n d tration t 16). TF-1 Cells--In order to confirm the role of SF as a survival Determination of Viable Cells Using the MTT Colorimetric AssayThe assay was based on the procedure describedby Mosmann (27) with factor, DNA was extracted from cells maintained in the presminor modifications. Cells were seeded at 105/mlin 35-mm Petri dishes ence or absence of SF. As controls, DNA was also extracted from for varying time points as indicated. Four hours before the end of the cells stimulated with GM-CSF or IL-3, two growth factors preculture, the M?Tdye (3-(4,5-dimethyl-thiazoyl-2-yl)-2,5-diphenyltetraviously shown to suppressACD in hemopoietic cells (13,15,16). zolium bromide)(Sigma)was added at a final concentration of 1mg/ml. absence of a n approDuring this incubation time, viable cells with active mitochondria will Data shown in Fig. 3 indicate that in the reduce the MTT dye to the purple formazan, resulting in a shiR in the priate source of growth factor, primary myeloblasts undergo absorption spectrum of the dye. Becausethe insoluble purple formazan the process of active cell death. Thepresence of SF or GM-CSF stays associated with the cells, the culture medium was removed in the culture medium suppresses apopotic cell death.
Suppression of Apoptosis by SF
12086
I
AML 41
2
0
+
GM
0 SF
I
/?
/ I
SF->OM
72
TF-1
0 SFGM
subculture
144
subculture
A
AML 41
OM ->OM
AML 56
FIG.3. SF suppressesACD in primary myeloblasts. Primary myeloblasts (AML 41 and 54) were maintained in the presence of SF or GM-CSF, a s shown in Fig. 2. After 3 days at 37 "C, DNA was extracted from lo6 cells as described under "Materials and Methods" and resolved by electrophoresis on a 1.2% agarose gel.
TF-1
M07-E
SF -> SF
0
72
144
time in culture (h) FIG.2. Role of SF as a survival factor and GM-CSF as a mitogenic factor. Primarymyeloblasts (AML 41) and TF-1 cells were seeded a t 105/ml, corresponding to an A of 0.05,in the presence of the indicated growth factors. After3 days a t 37 "C, viable cell recovery was determined in one set of cultures using the M'IT assay, as described under "Materials and Methods." The other set was subcultured at 105/ml and stimulated with the indicated growth factors for a n additional 3 days. Growth factor concentrations were: 400 PM for GM-CSF and 550 PM for SF. Data shown are the mean of duplicate cultures.S.D. was less than5% in all cases.
The biologically active concentration of SF in suppression of ACD was determined in both TF-1 cells and M07-E cells, which were exposed to varying concentrations of growth factor (Fig. 4). In the absence of SF, DNA degradation into oligonucleosomal bands wasobserved at 24-36 h. Addition of SF results ina dose-dependent suppressionofACD. The optimal concentration was in the range of 500 PM. In order to estimateEC,, values, data were analyzed through densitometry as described under "Materials and Methods." A typical densitometric profile is shown in Fig. 5. The slowly migrating peak represents intact genomic DNA, whereas the oligonucleosomal bands appear as faster migrating peaks. EC,, values estimated through curve fitting of densitometric data were in the range of 100-200 PM for TF-1 cells. In a separate set of experiments, cells were exposed to ['HIdThd, and DNA integrity was assessed through membrane filtration, as described previously (26). The doseresponse curves shown in Fig. 6 were comparable to those obtained through densitometry. Data analyses throughcomputer modeling (29) indicate that EC,, values were in the range of 50-80 PM,with optimal concentrationsa t 300-500 PM. Because SF is only a weak mitogen, the thymidine assay inSF-stimu-
FIG.4. The suppressionof ACD by SF is dose-dependent. TF-1 and M07-Ecells were exposed to the indicated concentrationsof SF for 24 h. DNA loaded in each lane was extractedfrom loficells. Negative images of the gels are shown.
lated TF-1cells represents cell survival rather thancell proliferation. This technique also has the advantageof allowing for a direct quantitation of results, whereas densitometry relieson an indirect assessment of aphotograph of the gel. Consequently, experiments designed to estimate the EC,, values of specific inhibitors were performed using the membrane filtration approach. Preferential Inhibition of Effects of SF by prphostin B42Tyrphostins are specific protein tyrosine kinase inhibitors (18). It has been shown previously that erbstatin, another protein tyrosine kinase inhibitor, can prevent the induction ofc-fos mRNA by GM-CSF (28). We have, therefore, investigated the specificities of six Tyrphostins (Table I), with regards to the suppression ofACD by GM-CSF or SF. Tyqhostins were added at different concentrations, ranging from 0.1 to 400 w, to cultures containing optimal concentrationsof either GM-CSF (400
Suppression of Apoptosis by SF
I
12087
PM) or SF (550 PM).Typical dose-response curves for three
0
Tyr-
phostins, B42, B46 and 1, are shown in Fig. 7. Quantitative data analysesindicated that allTyrphostins tested, except two, were 2 to 3 times more potent inhibitors of SF as compared to GM-CSF (Table I). Qrphostin 1 was inactive in either case at concentrations up to 50 p~ (Fig. 7 and Table I). There was a , evinarrow window of toxicity between 100 and 200 p ~ as denced by a slope factor of 5. In contrast to the other TyrSF phostins, Tyrphostin B42 was 10 times more efficient in antagonizing SF-induced suppression of ACD, when compared to GM-CSF triggered pathways. The effect of Tyrphostin 11 on DNA fragmentation was also studied through DNA electrophoresis. Data shown in Fig. 5 confirmed the results observed with the membrane filtration technique. The nontoxicity of Tyrphostin 1 on parental cells a t SF + EIPA concentrations as high as 50-75 p~ suggests that theeffects of the other Tyrphostins are specifically due toinhibition of tyrosine kinase activities associated with c-Kit or GM-CSFR activated pathways. Activation of Na+lH+Exchange by SFlc-Kit Interaction-We have previously shown that ligand-induced activation of the GM-CSF receptor results in pH, alkalinization (16). We have, II SF + STAUROSPORINE therefore, addressed the question whether the activation of c-Kit may result in increased Na+/H+exchange. M07-E cells were deprivedof GM-CSF for 5 h priorto labeling withBCECF. The additionof SF results ina rapid and sustainedincrease in pH,, which was in the order of 0.2 units after 2-3 min and lasted for at least 15 min after stimulation(Fig. 8). When cells were pretreated withEIPA, an amiloride derivative that blocks the function of the antiporter, there wasno significant change in pH,. Similarly, pretreatment with genistein, a tyrosine kinase inhibitor (25), and with staurosporine, a protein kinase C inhibitor (24), abrogated the effect of SF on pH,. All inhibitors were used a t concentrations that were below their toxic range, MOBILITY as shown previously (16) andhere on bcr-ab1 transformed FIG.5. Effect of inhibitors of protein kinases and of the Na+M+ antiporter on the suppression of ACD by SF. TF-1 cells wereincu- M07-E cells (Fig. 10). Taken together, our data suggest that ligand-induced activation of c-Kit results in a protein kinase bated in thepresence or absence of SF, as indicated, for 24 h.Inhibitors were added at theinitiation of culture: Tyrphostin 11(0.3-30 m), EIPA C-dependent activation of the Na+/H+antiporter. (1-20 m), staurosporine (5-100 n ~ ) Data . are shown for the highest Amiloride Derivatives Prevent the Biologic Effects of SF-In concentrations for each inhibitor. m e r electrophoresis, a negative imorder to address the role of the Na+/H+antiporter in the biologic age of the gel was taken and analyzed by densitometry, as described under "Materials and Methods." DNA peaks are shown as a function of effects of SF, we have selected two amiloride derivatives, EIPA the distance of migration. and HMA, on the basis of their specificities for the antiporter
&
250
cprn (x 1000)
Dm (x10001
TF- 1
200
-
150
-
100
-
50
-
t
01 1
*
'
""""
10
' '
1'"'''
*
' ' """'
100
SF (DM)
W07-E
*
1000
*
*
' ' """' 10000
1
10
100
1000
10000
SF (DM) FIG.6. Analysis of DNA integrity through membrane filtration of [SHldThdcells. L3H]dThd labeling and membrane filtration were performed as described under "Materials and Methods." EC, values estimated from data analysis are 52 7 PM for TF-1 cells and 83 2 14 PM for M07-E cells.
Suppression of Apoptosis by SF
12088
TABLE I Median effectiveconcentrations of Qrphostins in inhibition of SF-dependent or GM-CSF-dependent cellsurvival TF-1 cells were exposed to the indicated Tyrphostins at concentrations ranging from 0.15to 200 PM,in the presence of either SF(550 PM) or GM-CSF (400PM).Representative dose-responsecurves for Tyrphostin 1,B42,and B48 are shown in Fig. 5 . EC,, values are derived by data analysis of complete dose-response curves using the program ALLFIT
(28). EC., "" (UM) .. . SF GMCSF m TYROSINE HO
TYRPHOSTIN I
C
O
O
300 cprn Ixl O W I 250
+GM-CSF
100 50
0.1
10
100
Tyrphostin 842 ( P M )
H
M NH, '
"
9 4 t180 5 i 9 H,CO
0
lot3
24t1
Tyrphostin 848 (cM)
0 TYRPHOSTIN 842 CN
H
O
~
N
-
34.3 27r0.8 +SF
HO
0 TYRPHOSTIN 846 CN
H
o
W
t
d
CN
TYRPHOSTlN 848 CN
2t0.2
k0.3
HO
0.1
H
o
o
W
d
N
d 34.5
7t0.4
N
o
3t0.4 5e0.7
HO
Tyrphostin 1 Tyrphostin 11 Tyrphostin B42 Tyrphostin B46 Tyrphostin B48 Tyrphostin B56
10
100
1000
~
HO
H
1
Tyrphostin 1 (pM)
0 TYRPHoSTIN 856
T
(4-Methoxybenzy1idene)malononitrile 3,4-Dihydroxybenzylidenecyanoacetamide N-Benzyl-3.4-dihydroxybenzylidenecyanoacetamide N-(3-Phenylpropyl)-3,4-dihydroxybenzylidenecyanoacetamide
N-Phenyl-3,4-dihydroxybenzylidenecyanoacetamide N-(4-Phenylbutyl)-3,4-dihydroxybenzylidenecyanoacetamide
FIG.7. Comparative effect of Qrphostins on TF-1cells incubated with SF or GM-CSF.TF-1 cells (2x 105/ml) were incubated with SF (550 PM) or GM-CSF (400PM) for 24 h. At the initiation of culture, Tyrphostin B42, B48, or 1 were added at the indicated concentrations. C3H1dThd labeling was performed as described under "Materials and Methods." DNA integrity was assessed through membrane filtration. Data shown are the mean of duplicate determinations. The curves passing through the data were obtained by nonlinear regression analysis with the program ALLFIT. EC,, values estimated from curvefitting are shown in Table I.
ceeding 20 p ~ as, observed through a 50% reduction in cell count (data not shown). The effect of EIPA and HMA on DNA fragmentation in SF-stimulatedTF-1 cells was confirmed through DNA extraction andelectrophoresis (Fig. 5). Our data, therefore, indicate that the suppression of ACD by SF requires the presence of a functional Na+/H+antiporter. DISCUSSION
The present study provides direct evidence for a role of S F a s a trophic factor for hemopoietic cells. In contrast to GM-CSF (or IL-3) that acts both as a mitogen and a survival factor, SF (19). EIPA or HMA were addedat the indicated concentrations appears tobe mainly a survival factorfor hemopoietic cells. The to cultures stimulated with GM-CSF or with SF (Fig. 9). In suppression ofACD by SF is dose-dependent, with anEC,, of 50 striking contrast with Tyrphostins, the dose-response curves PM. Through theuse of Tyrphostins as specific protein tyrosine were exactly comparable using either growth factor. Hence, kinase inhibitors, and amiloride derivatives as blockers of the EC,, values, estimated through computer modeling of titration Na+/H+antiporter, we were able to show that the triggering of discrepancy in the immediate early curves, were7-8 PM for EIPA, and 12-13 p~ for HMA, either in two distinct pathways with the presence of SF or GM-CSF (Table 11). Similarly, the con- events, namely that of c-Kit and of GM-CSFR, converges tocentrations requiredfor maximum inhibitionof the response to wards the activationof the Na+/H+antiporter. It is shown here SF or GM-CSF were the same and were in 10-20 the PM range. that SF induces a rapid and persistent alkalinization of pH,, Comparable results were observed for both TF-1 cells (Fig. 9) which is required for its biologic effects. The severely decreased hemopoiesis and lethality at the hoand M07-E (Fig. 10). Again, the concentrations of EIPA and HMA used in these experimentsdid not induceDNA fragmen- mozygous state in Steel mice indicate the importanceof SF in tation in bcr-ab1 transformed M07-E (Fig. 10) or other growth hemopoietic cell development. Its biologic function was, howfactor-independent cell lines (HL-60, IRCMS, andOCI-AML-1) ever, notwell defined. In the present study, we provide evidence (16). There was, however, some toxicity at concentrations ex- for a main role of SF in suppression of apoptosis in primary
Suppression of Apoptosis by SF
-0.11
0
300
600
900
1200
I
time (sec)
.-
I Q
Q -0.1‘ 0
J 300
600
900
1200
time (sed
0
300
600
900
1200
time (sec) FIG.8. Effect of SF on intracellular pH of M07-E cells. M07-E cells (lo6)were labeled with the pH-sensitive probe BCECF, as described under “Materials and Methods.” At the indicated times, SF was added (550 PM) and the fluorescent profile was monitored for a n additional 20 min. Where indicated, cells were preincubated for either 10 min with EIPA (10 m),or 30 min with one of the following: genistein (20 mM) or staurosporine (100 nM). Initial pH, were 6.9 to 7.1 in all experiments.
cnm ( x 1000)
12089
myeloblasts and two model cell lines, with little if any effect on cell proliferation. After this paper was submitted, SF was reported to also prevent apoptosisinIL-3-dependent murine mast cells (30). In the hemopoietic system, SF is produced by stromal cells whereas IL-3 and GM-CSF are produced by activated T lymphocytes. Hence, our data suggest that in steady state hemopoiesis, when the production of GM-CSF and IL-3 is minimal, the main factor that regulatesapoptosis may be Steel factor. Both GM-CSF and SF can suppress ACD in the two model cell lines studied. While c-Kit is a tyrosine kinase receptor, GM-CSFR belongs to a family of cytokine receptors that show homology instructural motifs intheextracellular domain, whereas the cytoplasmic tail of either chain of the receptor lacks a typical kinase domain (31, 32). A tyrosine kinase referred to as Jak-2 has been shown to associate with GM-CSFR (for review see Ref. 33). Our data indicate that both SF-induced pathways and GM-CSF-induced pathways are sensitive to inhibition by Tyrphostins, albeit with a quantitative difference. All Tyrphostins tested in the present study are more potent inhibitors of SF-dependent pathways than of GM-CSF-induced pathways. Tyrphostins were initially designed with regards to inhibition of epidermal growth factor receptorkinase activities (34). Whether Tyrphostins are more potent inhibitors of c-Kit kinase activities than of Jak-2 kinase activitiesbecause of structural design remains to be documented. Another possibility is the difference between the multicomponent structure of the GM-CSFR complex, Q and p chains (Ref. 35, and reviewed in Ref. 361, associated with Jak-2 (33), and the simpler structure of activated c-Kit, which is a homodimer with an intrinsic kinase activity (37). We have, nonetheless, identified Tyrphostin B42 as a preferential inhibitor of SF-dependent pathways. Tyrphostin 1 does not bind the protein tyrosine kinase domain, hence it is inactive and serves as a negative control. Substitution of one of the two cyano groups by an acetamide group in Tyrphostin 11, together with two hydroxy groups at positions 3,4 of the benzyl nucleus, has rendered the compound biologically active. Tyrphostin B42 differs from Tyrphostin 11through the addition of the N-benzyl side chain, which results in in-
corn (x1000)
TF- 1
0.1
1
10
100
0.1
1
10
100
EIPA (uM) HMA (uM) FIG.9. Comparative effect of EIPAand HMA on cell survival maintained in the presence of SF or GM-CSF. TF-1 cells were incubated for 24 h with SF (550 PM) or GM-CSF (400 PM).HMA or EIPA were added at the initiation of culture. DNA integrity was determined through membrane filtration as described. The curves passing through the data were calculated using nonlinear regression EC,, analysis. values are shown in Table 11.
Suppression of Apoptosis by SF
12090
TABLEI1 HMA Median effectiveconcentration of EIPA and HMA in inhibition of SF-dependent or GM-CSF-dependent cellsurvival TF-1 cells were exposedto HMAor EIPA(concentration range 0.3-30 p ~ in ) the presence of SF (550 PM) or GM-CSF (400 PM)for 24 h. EC, values are derived by regression analysis of the data shown in Fig. 7 with the ALLFIT program.
HMA
022
SF:
0
HMA 10
ON10
N
- + + + + +
IL-3:
~
++++
EIPA "
OlVfiZ N10Z
EC, (pur)
SF GM-CSF
HMA
EIPA
12.3 t 0.52
7.04 2 0.37 7.95 f 0.24
12.7 f 0.42
creased potency in inhibition of c-Kit pathway without any apparent change in affinity for Jak-2 or other putative GMCSFR-associated kinases. However, when the phenyl group was furtherseparated from the benzilidenecyanoacetamide nucleus through a propyl or butyl group, as in Tyrphostin B46 and B56, respectively, the potency for GM-CSF-dependent pathway increases, i.e. the selectivity for c-Kit was significantly decreased. Taken together, our data suggest that theproximity of the phenyl side chain maybe important for increased affinity for either c-Kit- or GM-CSFR-associated kinases, either through direct N-attachment of the phenyl group (Tyrphostin B48), or through refolding of a flexible propyl- and butyl-phenyl side chain(Tyrphostin B46 and B56). Finally, when the phenyl group is separated by 1 carbon residue (Tyrphostin B42), the side chain may be less flexible and does not allow refolding, hence the selective inhibition of c-Kit. Our data suggest that new Tyrphostins can be modeled on B42 for increased potency and selectivity towards c-Kit. Conversely, GM-CSFR-directed Tyrphostins can be designed on the basis of the molecular structure of B46 and B56. In contrast to the divergence in the concentrations of Tyrphostins required for inhibition of c-Kit- and GM-CSFR-activated pathways,the inhibition curves for EIPA in thepresence of GM-CSF or SF could be exactly superimposed. Hence, our data suggest that both pathways converge towards the activation of Na+/H+exchange. Whether the Na+/H+antiporter itself or a proximal upstream event mightbe the point of convergence remains to be documented. It has been shown previously that the antiporter isphosphorylated on serine residues, following growth factor stimulation of resting cells (38). It is, therefore, possible that input from different pathways may activate a common switch kinase, resulting in the activation of specific serine kinases (reviewed in Ref. 39). One such candidate could be the mitogen-activated protein kinase, which has been shown to be tyrosine phosphorylated following the activation of GMCSFR and c-Kit (40), or the mitogen-activated protein kinase kinase (39). A role for the Na+/H+antiporter in cell proliferation was previously suggested through several lines of investigation that include the use of amiloride derivatives, and the study of Na+/H+exchange mutants. The use of inhibitors was limited because of their possible interference with other exchangers, and the possibility that the inhibition might not be related to Na+/H+exchange. We have addressed these questions in two ways. First, we have selected two amiloride derivatives developed by Cragoe, that are highly specific blockers of Na+/H+ exchange, as opposed to epithelial Na+ channels. Second, we have performed control experiments on M07-E cells that were rendered growth factor-independent through constitutive expression of bcr-abl. Growth in these cells was not sensitive to EIPA or HMA at the concentrations used against parental M07-E or TF-1 cells, indicating that theirinhibitory effects in SF-stimulated cells may be attributed to the specific blockingof Na+/H+exchange. Previous data suggested a role for the Na+/H+ antiporter in suppression of ACD by GM-CSF and IL-3. The
M07E
MEA
FIG.10. Effect of EIPA and HMA on bcr-ab1transformed MO7-E cells. M07-E cells maintained with IL3 or bcr-ab1 transformed cells (MBA) were exposed to the indicated concentrations of HMA or EIPA (PM)in culture in low NaCl medium (16). DNA equivalent of lo6 cells was loaded per lane.
interpretation of these resultswas, nonetheless, complicated by the fact that thesecytokines are potent mitogens for hemopoietic cells. Because SF is mainly a survival factor in our model, data shown here confirm that the Na+/H+antiporter plays an important role in the suppression of ACD. Acknowledgments-We thank Dr. Marielle Gascon-Barn6(Centre de recherches cliniques Andre Wallet, Saint-Luc Hospital, Montreal) for kindly providing access tothe dual excitation spectrofluorometer,to Dr. Marie Trudel (IRCM) for critical reading of the manuscript, and Francine De Coste and Christine Lemire for secretarial assistance. REFERENCES 1. Fleishman, R.A. (1993) IIknds Genet. 9.259-292 2. Anderson, D. M., Lyman, S.D., Baird, A,, Wignall, J. M., Eisenman. J., Rauch, C., March, C. J., Boswell, H.S., Gimpel, S.D., Cosman, D., and Williams, D. E. (1990) Cell 63,235-243 3. McNiece, I. K.,Langley, K.E., and Zsebo, K M. (1991) Exp. Hemafol. 19, 226-23 1 4 6239, 4. Metcalf, D., and Nicola, N.A. (1991) Proc. Natl. Acad. Sci. U.S. A. € 6243 5. Caceres-Corks, J. R.,and Hoang, T.(1992) Cancer Res. 62,5208-5212 6. Godin, I., Deed, R., Cooke, J., Zsebo, K., Dexter, M., and Wylie, C. C. (1991) Nature 352,807-809 7. Dolei, S., Williams, D. E., Ernst, M. K., Fksnick, J. L., Brannan, C. J., Lock, L. F.,Lyman, S. D., Boswell, H. S., and Donovan, P.J. (1991) Nature 352, 809-811 8. Martin, D. P., Schmidt, R. E., DiStefano, P.S., Lowry, 0.H., Carter, J. G., and Johnson, E. M., Jr. (1988) J. Cell. Biol. 106,829-844 9. Williams, G. T. (1991) Cell 66, 1097-1098 10. McConkey, D. J., Orrenius, S., and Jondal, M. (1990) Zmmunol. lb&y 11, 120-121 11. Barres, B. A., Hart, I. K., Coles, H. S., Burne, J. F., Voyvodic, J. T., Richardson, W. D., and Raff, M. C. (1992) Cell 7 0 , 3 1 4 6 12. Masu, J., Wolf, E., Holtmann, B., Sendtner, M., Brem, G., and Thoenen, H. (1993) Nature 366,2742 13. Williams, G. T.,Smith, C. A., Spooncer, E., Dexter, T.M., and Taylor, D.R. (1990) Nature 343,76-79 14. Koury, M. J., and Bondurant, M. C. (1990) Science 248,376481 15. Rodriguez-Tarduchy, G.,Collins, M., and Lopez-Rivas, A. (1990) EMBO J. 9, 29974002 16. Rajotte, D., Haddad, P., Haman, A,, Cragoe, E. J., Jr., andHoang, T. (1992) J. Biol. Chem. 267,9980-9987 17. Lotem, J., Cragoe, E. J., Jr.. and Sachs, L. (1991) Blood 78,953-960 18. Gazit, A., Yaisch, P., Gilon,C., and Levitzki, A. (1989) J. Med. Chem. 32, 2344-2352 19. Kleyman, T. R., and Cragoe, E. J., Jr. (1988)J. Mernbr. Biol. 106,l-21 20. Kitamura, T.,Tange, T., Terasawa, T., Chiba, S., Kuwaki, T.,Miyagawa, IC, Piao, Y.-F., Miyazono, K., Urabe, A,, and Takaku, F. (1989) J. Cell. Physiol. 140,323434 21. Avanzi, G. C.,Brizzi, M. F., Giannotti, J., Ciarletta, A., Yang, Y.-C., Pegoraro, L., and Clark, S. C. (1990) J. Cell. Physiol. 146, 458-464 22. Sirard, C., Laneuville, P., and Dick, J. E. (1994) Blood, in press 23. Stanley, J. B., Gorczynski, R., Huang, C. K, Love, J., and Mills, G. B. (1990)J. Immunol. 147,2189-2198 24. Tamaoki, T.,Nomoto, H., Takahashi, I., &to, Y., Morimoto, M.,and 'bmatin, I. (1986) Biochem. Biophys. Res. Commun. 136,397402 25. Akiyama, T., Ishida, J., Nakagawa, S., Ogawara, H., Watanabe, S.,Itoh, N., Shibuya, M., and Fukami, Y. (1987) J. Biol. Chem. 262,5592-5595 26. Matzinger, P.(1991) J. Zmmunol. Methods 146, 185-192
Suppression of Apoptosis by SF 27. Mosmann, T.(1983)J . Immunol. Methods 65,55-58 28. McColl, S. R., DiPersio, J. F., Caon, A. C., Ho, P., and Naccache, P. H. (19911 Blood 78, 1842-1852 29. DeUan, A,, Munson, P. J., and Rodbard, D. (19781Am. J. Physiol. 236, E97E102 30. Mekori, Y.A,, Oh, C . K., and Metcalfe, D.D. (1993)J. Immunol. 151, 37753785 31. Gearing, D. P., King, J.A,, Gough, N. M., and Nicola, N. A. (1989)EMEO J. 8, 3667-3676 32. Hayashida, K., Kitamura, T.,Gorman, D.M., Arai, K. J., Yokota, T., and Miyajima, A. (1990)Proc. Natl. Acad. Sci. U.S. A. 87, 9655-9659
12091
33. Stahl, N.,and Yancopoulos, G . D. (1993)Cell 74,587-589 34. Yaisch, P., Gazit, A,, Gilon, C., and Levitzki, A. (1988)Science 242,933-935 35. Hoang, T.,De U a n , A., Haman, A., Beauchemin, V., Kitamura, T., and Clark, S. C. (1993)J. Biol. Chem. 268, 11881-11887 36. Nicola, N. A,, and Metcalf, D. (1991)Cell 67, 1 4 37. Lev, S., Yarden, Y.,and Givol, D. (1992)J . Biol. Chem. 267, 15970-15977 38. Sardet, C . , Counillon, L., Franchi, A., and Pouyssegur, J. (1990)Science 247, 723-726 39. Crews, C. M., and Erikson, R. L. (1993)Cell 74,215-218 40. Hallek, M., Druker, B., Lepisto, E. M., Wood, K. W., Ernst, T. J., and Griffin,J. D. (1992)J. Cell. Physiol. 153, 176-186
