the role of granulocyte colony stimulating factor receptor signalling in neytrophil dev

8/12/2019 The Role of Granulocyte Colony Stimulating Factor Receptor Signalling in Neytrophil Dev

1/136

THE ROLE OF GR NULOCYTE COLONY-STIMUL TING F CTOR RECEPTOR SIGN LINGIN NEUTROPHIL DEVELOPMENT

JOHN DE KONING


2/136


3/136

THE ROLE OF GR NULOCYTE COLONY-STIMUL TING F CTOR RECEPTOR SIGN LINGIN NEUTROPHIL DEVELOPMENT

E ROL V AN G CSF RECEPTOR SIGNAAL TRANSDUCTIE INE ONTWIKKELING VAN NEUTROFIELE GRANULOCYTEN

PROEFSCHRIFT

Ter verkrijging van de graad van doctor aan de Erasmus Universiteit Rotterdamop gezag van de Rector Magnificus Prof. Dr. P.W.C. Akkermans M.A.

en volgens besluit van het College voor PromotiesDe open bare verdediging zal plaatsvindenop woensdag 10 maart 1999 om 15:45 uur

door

JOHN P UL E KONINGgeboren te Rotterdam


4/136

rornotiecommissie

romotoropromotor

Overige leden

Prof. Dr. B LowenbergDr. I.P. TouwProf. Dr. R. BernardsProf. Dr. J.A. GrootegoedProf. Dr. J.H.J. Hoeijmakers

The work described in this thesis was performed at the Institute of Hematology,Erasmus University and the Department of Hematology, Dr. Daniel den Hoed CancerCenter, Rollerdam, The Netherlands. This work was supported by grants from theDutch Cancer Society Koningin Wilhelmina Fonds and the NetherlandsOrganization for Scientific Research (NWO). Financial support from the DutchCancer Society for printing of this thesis is gratefully acknowledged.Cover designed by Karola van RooyenPrinted by Haveka B.Y., Alblasserdam, The Netherlands


5/136


6/136


7/136

Contents

Chapter1 11 21 31 41 51 61 71 8

Chapter 2

Chapter 3

Chapter 4

Chapter 5

Chapter 6

AbbreviationsIntroductionHematopoiesisHematopoietic growth factorsG-CSF and granulopoiesisG-CSF receptor: structure and subdomainsG-CSF receptor: isoformsG-CSF receptor: diseaseG-CSF receptor: signal transductionScope of this thesisSpecific involvement of tyrosine 764 of human granulocytecolony.stimulating factor receptor in signal transductionmediated by p145/Shc/GRB2 or p90 GRB2 complexesBlood 87: /32 996The membranedistal cytoplasmic region of humangranulocyte colony.stimulating factor receptor is requiredfor STAT3 but not STAT1 homodimer formationlood 87: 1335 1996

Tryptophan 650 of human granulocyte colony.stimulatingfactor GCSF) receptor implicated in the activation ofJak2, is also required for GCSFmediated activation ofsignaling complexes of the p21 ~ routeBlood 87:2148 1996Proliferation signaling and activation of Shc p21 ~ andMyc via tyrosine 764 of human granulocyte colonystimulating factor receptorBlood 91:1924 1998STAT3controlled expression of the cyclin.dependentkinase inhibitor p27 K1pl during GCSFinduced growtharrest of myeloid cellsSubmitted

121213141516172027

45

61

73

91


8/136

Chapter 77 17.27 37 47.5

General discussion I 13Mechanisms of p21 R and STAT activation by G CSF R 114Control of proliferation by G CSF R 116Differentiation signaling by G CSF R 119Function of mutant G CSF R in severe congenital neutropenia 122Future directions 123Summary Samenvatting Summary ill Dutch) 129Dallkwoord Acknowledgments) 135Curriculum vitae 136


9/136


10/136


11/136

CHAPTER

ntroduction

partly published in

Advances in understanding the biology and function of the G-CSF receptorJohn P de Koning and Ivo P TOllw

ClIrrellt Opinioll ill Hematology 3:180 1996


12/136

1.1 HematopoiesisHematopoiesis or blood cell formation is a strictly regulated process that, in

adult individuals, takes place mainly in the bone marrow. All blood cells are derivedfrom a small population of pluripotent stem cells that are capable of self-renewal anddifferentiation towards distinct lineage-committed progenitor cells. These committedprogenitor cells can undergo proliferation followed by terminal differentiation intomature blood cells that include granulocytes, monocytes/macrophages, lymphocytes,erythrocytes, and platelets. Most mature blood cells have a limited life span and needto be replenished constantly. This continuous production is tightly balanced and isregulated essentially by two mechanisms. Stromal cells in the bone marrow affecthematopoiesis by direct cell-to-cell contact and provide a suitable microenvironmentrequired for hematopoietic cell development. In addition, a network of cytokines andhematopoietic growth factors HGF) specifically control the proliferation,differentiation, survival, and function of different hematopoietic cells. This network isparticularly important under stress conditions, such s infection or bleeding, when arapid rise in specific blood cell types is needed.

1.2 Hematopoietic growth factors

HGFs are glycoproteins that are produced by stromal cells, fibroblasts,endothelial cells, lymphocytes, monocytes and macrophages, and by specialized cellsin various organs, such as kidney and liver. The levels of HGFs are normally low, butcan be enhanced substantially in response to extracellular stimuli. HGFs can beroughly subdivided into early-acting lineage-nonspecific factors and late-actinglineage-specific factors. Some early-acting factors are interleukin IL)-I, IL-3, IL-4,IL-6, IL-II stem cell factor SCF), and granulocyte-macrophage colony-stimulatingfactor GM-CSF). For instance, IL-3 stimulates the growth of multipotentialprogenitor cells that can develop into granulocytes, macrophages, erythrocytes, andmegakaryocytes. However, IL-3 does not support the terminal stages ofhematopoiesis, as IL-3 responsiveness of progenitor cells declines with differentiationI). In contrast, late-acting HGFs, including IL-5, erythropoietin EPO),

thrombopoietin TPO), macrophage colony-stimulating factor M-CSF), andgranulocyte colony-stimulating factor G-CSF), are mostly lineage specific. They notonly stimulate proliferation, but also induce terminal differentiation of hematopoieticcells of particular lineages. Further characteristic features of HGFs are their functionalpleiotropy and redundancy. Many HGFs, in particular those acting at early stages ofhematopoiesis, display overlapping biological activities on a variety of cell types.Conversely, different HGFs can act on the same cell type to mediate similar effects.The functional interplay between HGFs, either synergistically or antagonistically, willdetelmine the cellular response.

2


13/136

1.3 G-CSF and granulopoiesisNeutrophilic granulocytes are n essential component of the host defense system

against infections. Mature neutrophils arise from bone marrow stem cells following aprocess involving proliferation, commitment to the granulocytic lineage, and terminaldifferentiation. Several HGFs, including IL-3, IL-6, SCF, GM-CSF, and G-CSF, havebeen shown to be positive regulators of granulopoiesis 2,3). Most of these HGFssupport the proliferation of early myeloid progenitor cells and have only a limitedability to induce differentiation. G-CSF, however, not only stimulates theproliferation but also strongly induces terminal differentiation of granulocyticprogenitor cells. G-CSF also activates certain functions of mature neutrophils andpromotes their survival 4). G-CSF enhances the production of superoxide, alkalinephosphatase, and myeloperoxidase, the chemotactic activity and the antibodydependent cellular cytotoxicity of neutrophils 5-9). Using serum-free culturesystems, an absolute requirement for G-CSF for neutrophil colony formation has beenshown 10). Neither IL-3 nor GM-CSF alone is capable of effectively supportingneutrophil colony formation. However, G-CSF can synergize with both IL-3 andGM-CSF to support neutrophil progenitor proliferation in culture. Thus, IL-3 andGM-CSF induce proliferation of progenitor cells in the early stages of development,whereas G-CSF is a neutrophil lineage-specific regplator that stimulates the terminalstages of neutrophil development.

vivo studies in mice have indicated that G-CSF induces a much higherperipheral blood granulocyte count than does IL-3 or GM-CSF 3). Thegranulopoietic effects of G-CSF have been reproduced in humans and have led to itswidespread clinical application in the setting of chemotherapy-induced neutropeniaand bone marrow transplantation II). Several observations have indicated thatG-CSF is indispensable for normal granulopoiesis ill vivo. In humans, low neutrophilcounts coexist with high serum G-CSF levels, suggesting a regulatory role for G-CSFin maintaining steady state neutrophil production 12). Concurrent elevation inneutrophil numbers and serum G-CSF levels observed during infections implies a rolefor G-CSF as n emergency factor for granulopoiesis 13). Furthermore, dogs injectedwith human G-CSF that produced neutralizing antibodies cross-reactive againstendogenous canine G-CSF developed prolonged neutropenia 14). Infusion of plasmafrom a neutropenic dog with anti-G-CSF antibodies into a normal dog also causedneutropenia. More recently, it was shown that mice lacking G-CSF due to targeteddisruption of the G CSF gene developed chronic neutropenia 15). Additionally, thegranulopoietic response to infection was severely impaired in G-CSF-deficient mice.Collectively, these results indicate that G-CSF plays a key role in regulatinggranulopoiesis in both steady state and stress conditions.

3


14/136

1 4 G-CSF receptor: structure and subdomainsThe biological effects of G-CSF are mediated through a specific receptor on the

surface of responsive cells. The G-CSF receptor (G-CSF-R) is a member of thehematopoietin receptor superfamily to which the receptors for IL-2 to IL-7, IL-9,IL-II, IL-12, EPO, TPO, GM-CSF, leukemia inhibiting factor (LIF), growthhormone, prolactin, ciliary neurotrophic factor, and the gp 130 signal transducingprotein also belong (16). Characteristic structural features of members of this familyare the presence of four highly conserved cysteine residues and a motif of tryptophanserine-X-tryptophan-serine (WSXWS), within an approximately 200 amino acidregion in the extracellular domain. This region is referred to as the cytokine receptorhomology (CRH) region and is crucial for ligand binding. Molecular cloning showedthat the murine and human G-CSF-R are single transmembrane proteins of 812 and813 amino acid residues, respectively, with 62.5 homology at the amino acid level(17-19). The extracellular domain of the human G-CSF-R comprises 603 amino acidsand is composed of an immunoglobulin-like sequence, a CRH region, and three

CCCCwsxws

t1'::F;yry-{ ,

704

o~ 1 . u7 _.. STAT3

blo\:a-HA

. . . . Ig

bIO ;n-STAT3 E- - STAT3Figure 2. Expression of STATI variants in 32DfG CSF R cells. In STATIF, tyrosine residueat position 705 essential for STAT dimer fomlation was substituted for phenylalanine. InSTATID, glutamic acids 434 and 435 important for DNA binding were replaced by alanines.STAT3 immunoprecipitation on Iysates of 32D transfectants cultured in IL-3-containingmedium ITllTl1unoprecipitates were subjected to \Vestem blot analysis using anti-HAantibodies. Blots were reprobed with anti-STAT3 antibodies against amino acids 750-769).Ig immunoglobulins.

97


98/136


99/136


100/136

TABLE I. Neutrophilic differentiation of 32 cells expressing STAT3 variantsCells

vector controlSTAT3WTSTAT3FSTAT3

Neutrophilic differentiationG-CSF G-CSF Ara-C

63 7 53 27 4 38 30 43 II

0 34 8J Percentages o temlinally differentiated neutrophils Cells were cultured for 9 or 3 days

in G - C S F ~ - c Q n t ? i n i n g medhml in t h ~ cont.inuous presence or absence o Ara-C respectivelyAt least 200 cells were scored The means SE o 4 independent clones are shown

Ara-C, G-CSF induced terminal neutrophilic differentiation in STAT3WT and vectorcontrol cells as well as in STAT3 and STAT3F cells with comparable efficiencies(Table I). Thus, activation of STAT3 appears to be essential for G-CSF-mediatedgrowth arrest, but not for induction of differentiation itself.

The p27Kip

promoter contains a functional STAT3 binding site To obtain anindication as to how STAT3 mediates a growth atTest of 32 cells, we analyzed theavailable promoter sequences of cdk inhibitors. Previously, STAT responsiveelements in the p2 WAFI/CIPI promoter have been found that showed binding activityfor STATI and STAT5 (29,32). We identified a potential STAT-binding site in thepromoter of p27 at position -1585 (28). EMSAs with oligonucleotides derived fromeither the p2 or the p27 promoter were performed to verify whether G-CSF-activatedSTAT proteins indeed bind to these sequences. G-CSF induced the formation of threedistinct nuclear complexes that bound to p27 oligonucleotides. Supershift analysisshowed that the slowest migrating complex consists of STAT3 homodimers, thecomplex with intermediate mobility of STAT I-STAT3 heterodimers, and the fastestmigrating complex of STAT I homodimers (Figure SA). EMSAs with p2oligonucleotides on the same Iysates indicated that G-CSF induced the formation oftwo distinct nuclear complexes with binding activity for the p2 promoter sequence(Figure 5B). Supershift analysis revealed that these complexes consist of STAT5 andSTATI homodimers (data not shown). Because STAT I homodimers are not activatedand STAT5 homodimer formation is not increased upon G-CSF treatment comparedwith IL-3 stimulation (Figure I), we do not consider this of functional importance forinducing a cell cycle arrest in these cells (see below). Overexpression of DN-STAT3reduced the levels of STAT3 binding to p27 oligonucleotides, whereas binding ofSTAT5 to p2 oligonucleotides was unaffected (Figure 5), further supporting thatSTAT3 is involved in the regulation of p27, but not of p21, expression. Comparison100


101/136

A p27 promoter GAAATTTAAT TTCCTGTAA CATGnTG,STAT3F 6 1' 15 30,--' -,',,-0_-,,-0 15' 30' 60- ~ STAT3: - STAT113-, . STAT1/1, -B p2 promoter GATCeT TTCTGAGAA ATGG

,'=Clor control STAT3FO' 15' 30' 60' Il3 O' 15' 30 60' Il3

.1,_- STATS S1- STAT'Figure 5. Identification of STAT proteins that bind to spedn p27 or p2 promoter sequencesby EMSA. Serum aild growth factor deprived 32D transfectants were incubated at 37C withG-CSF 100 nglml) for the times indicated or with IL-3 100 nglml) for 15 minutes. Nuclearextracts were prepared and incubated with 31P-Iabeled double-stranded p27 (A) or p21 (6)oligonucleotides. For supershift analysis. nuclear extracts from vector control cells stimulatedfor 15 minutes with G-CSF 100 nglml) were preincubated without -) or with the indicatedantibodies before the addition of 32P labeled p27 oligonucleotides.

o Figures 3 and 5 reveals a striking overlap in binding activity between p27 and m67and between p21 and 13 casein oligonucleotides. This was further confirmed bycompetition analyses. Binding o STATs to radioactive p27 oligonucleotides wasefficiently competed by an excess o unlabeled m67, but not or hardly by p21 or13 casein probe. Conversely, an excess o 13,casein, but not o p27 or m67 probe,prevented STAT binding to p21 oligonucleotides data not shown). Theseobservations are in agreement with the similarities in palindromic sequences betweenp27 TTCCTGTAA) and m67 TTCCCGTAA) and between p21 TTCTGAGAA)and 13 casein TTCTAGGAA) oligonucleotides.To investigate whether STAT3 directly induces p 7 gene expression, weperfomled reporter assays in HeLa cells transiently expressing a p27-promoterluciferase construct, a G-CSF,R expression plasmid, and different ratios o WT- andDN-STAT3. HeLa cells were used for these experiments because the differentiationcompetent 32D clone used in this study appeared to be unsuitable for Iransienttransfections. In the presence o WT-STAT3, G-CSF induced a 7,fold increase in p27promoter activity Figure 6). The p27 promoter also contains a cAMP responsiveelement CRE) at position ,286 28) and cAMP induces p27 expression in M-CSF-

101


102/136

treated macrophages 33) and in G-CSF-treated NFS-60 cells 34). As a control forthe specific action of ON-STAT3 in this assay, we also treated cells with choleratoxin to elevate intracellular cAMP levels. p27 promoter activity was augmentedapproximately 2A fold by cholera toxin and IS-fold by G-CSF plus cholera toxin. Asexpected, increasing amounts of ON-STAT3 progressively inhibited transactivationinduced by G-CSF, whereas CRE-driven luciferase activity was not affected Figure6).

STAT3 s esselltial for G CSF illdllced expressioll o p27KiPI To confirm theeffect of STAT3 on p27 promoter activity, we investigated the ability of G-CSF toinduce expression of p27 mRNA in 320 transfectants. Northern blot analyses showedthat in both nontransfected and vector control cells p27 mRNA levels increased after2 to 3 days of stimulation with G-CSF Figure 7 A). In contrast, p27 mRNA was notinduced by G-CSF in 320 cells overexpressing ON-STAT3 Figure 7A). We nextexamined whether increase in p27 mRNA results in induction of p27 proteinexpression during G-CSF-induced growth arrest and neutrophilic differentiation of320 cells. In vector control cells, p27 protein appeared upon 3 days of stimulationwith G-CSF, and reached a maximum after 5 to 7 days Figure 7B). This timing ofp27 protein expression is preceded by the increase in p27 mRNA and coincides withthe observed growth arrest in these cells Figure 4A). As expected, G-CSF did notinduce p27 protein in 320 cells overexpressing ON-STAT3 Figure 7B). Ouring thisprolonged G-CSF stimulation, STAT3 activation remained significantly diminished in

2000

1500~Q

1000o::

500

ono f ctor GCSF CT

WT F250 ng

IIlI 200 ng 50 ngIIlI 125 ng 125 ngE 50 ng 200 ng13 250 ng

G CSF CT

Figure 6. Effect o STAT3 variants on transactivation o p27 promoter in HeLa cells. HeLacells were transfected with a p27-promoter luciferase reporter constmct a human G-CSF-Rexpression plasmid and different amounts o constructs encoding STAT3WT or STAT3F asindicated. After transfection. cells were cultured for 48 hours without factor with cholera toxin200 ng/ml; CT), and/or with G-CSF 100 ng/ml). Cell Iysates were prepared and assayed forluciferase activity. Luciferase light units o extracts from untransfected cells were used as basal

level. Data represent the means o 4 independent experiments. Incubation with forskolin ordibutyryl-cAMP resulted in responses similar to those obtained with cholera toxin.102


103/136

A __ ~ G ~ ~ C ~ S F ~ __d1 d2 d3 d4 d5L 3

32.5 T

2 T 1 T1 T.L1.5 .LT T T T 1 T.L .L 1 ~B GCSF 0.5IL 3 61:'- l 62 03 d.t tiS rl6 d7

I- :J-'3F ~ l-p

0 I I I I IIL 3 d1 d2 d3 d, d5

days n cu lure0 vector control --- -- STAT3F

Figure 7. Induction of p27 expression in 32D transfectants. Cells were switched from lL-3- toG-CSF-containing medium after extensive washing to remove residual IL-3 and incubated at37C for the times indicated. A) Total RNA (10 JIg) of 32DIG CSF R cells was analyzed byNorthern blot hybridization using a 32P-labeled p27 probe. The blot was reprobed with GAPDHto contiml equal loading. The graph shows quantitative analyses of p 7 rnRNA induction,expressed as fold induction by G-CSF compared to by IL-3. The means SE of 3 independentclones are shown. B) Lysates were prepared and analyzed by Westem blotting with anti-p27antibodies.STAT3F and STAT3D cells as compared with vector control cells (data not shown).Reprobing of the blots with anti-p21 antibodies did not reveal any alterations in the(low) levels of expression of p21 protein (data not shown).

p 1- mice show enhallced G CSF illduced colollY formatioll. To furtherconfirm the involvement of p27 in the control of G-CSF-induced myeloid cellproliferation, we performed ill vitro colony assays with bone malTOW and spleenmononuclear cells of p27 knockout mice. The generation of p27-deficient mice hasbeen previously described (31). All tissues of these mice are enlarged and containmore cells than of their wild-type littermates. Furthermore, the absence of p27 resultsin enhanced proliferation of hematopoietic progenitor cells in response to severalgrowth factors. However, the effect of p27 deficiency on G-CSF-induced colonyformation was not addressed. We therefore performed ill vitro colony assays withcells of p2T1. and p 7 1 mice in the presence of G-CSF, GM-CSF, or IL-3. Inagreement with previous observations (31), the total number of GM-CSF-responsiveprogenitor cells per organ was increased in p2T mice compared with controllittermates (femur: I.S-fold; spleen: 2-fold; Figure SA). Equivalent results wereobtained for IL-3-responsive colony forming cells (femur: lA-fold; spleen: 2-fold;Figure SA). Interestingly, the difference in G-CSF-responsive progenitors between

103


104/136


105/136

p 2 T ~ mice compared to control Iittermates (Figure 8B). G ~ C S F concentrationsrequired to reach maximal colony formation were similar. The increase in colonynumbers might be the result of both a selective expansion of the G ~ C S F ~ r e s p o n s i v eprogenitor compartment in p 2 7 ~ d e f i c i e n t mice as well as of an enhanced proliferativecapacity of the progenitor cells in response to G ~ C S F

If p27 plays a key role in G C S F ~ i n d u c e d growth arrest that is a prerequisite formyeloid cells to proceed with differentiation, G ~ C S F ~ s t i m u l a t e d colonies derivedfrom p 2 T ~ mice are also expected to be larger in size and to contain lessdifferentiated cells. Although the size of the colonies was similar on day 7 of culture,the average number of cells per colony on day 14 was twice as high in coloniesderived from either bone malTOW or spleen cells of p 2 T ~ mice compared with thoseof p 7 1 mice. Consistent with these findings, cytological examination of G ~ C S F -stimulated colonies on day 7 of culture demonstrated lower percentages of terminallydifferentiated neutrophils in colonies derived from p 2 T ~ mice compared with p27+i+mice (i.e. 44 6 vs. 67 7 for bone malTOW colonies, and 38 4 vs. 52 1for spleen colonies). These results provide additional evidence supporting thehypothesis that p27 is involved in G-CSF-mediated neutrophilic differentiation bydown modulation of proliferation.STAT3f3 does not affect G CSF induced proliferation and neutrophilicdifferentiation. The short form of STAT3, STAT3/3, that differs from STAT3 by thereplacement of the C-terminal 55 amino acid residues by 7 residues specific forSTAT3/3 as a result of altemative splicing, is predominantly expressed in cellscapable of differentiation (35). To examine a possible role of STAT3/3 in G-CSFmediated responses, we first investigated whether the expression level of STAT3f3 isinfluenced by G-CSF during proliferation and neutrophilic differentiation of 32Dcells. Although 32D cells express some STAT3/3 protein, G ~ S F did not modulate theexpression level, suggesting no direct involvement of STAT3/3 in the control ofdifferentiation (Figure 9). To further substantiate this, we ectopically overexpressedSTAT3/3 in G-CSF-R expressing 32D cells (Figure IDA). Overexpression of STAT3/3did not interfere with G-CSF-induced STAT3 and STATI homo- and heterodimer

G ~ S FI L ~ 3 60 d1 d2 d d4 d5 d6 d7C : j = ~ ~ ~ ~

Figure 9. Expression of STATI isofomlS in 3 2 D / G ~ C S F ~ R cells. Cells were switched fromIL 3 to G eSP containing medium after extensive washing to remove residual IL 3 andincubated at 37C for the times indicated. Lysates were prepared and analyzed by Westernblotting with anti-STATI antibodies (against amino acids 1 ~ 1 7 5 .

105


106/136

formation data not shown). STAT3P overexpressing cells did not show alteredG-CSF-induced p27 protein expression and displayed growth and differentiationcharacteristics similar to vector control cells Figures lOB and C). Furthermore,G-CSF-induced p27 promoter activity was not affected by transfection of STAT3P inHeLa cells, whereas the inhibition of transactivation by STAT3F could be fullyrestored by cotransfection of STAT3P data not shown). Taken together, these dataindicate that STAT3P does not act as a dominant-negative regulator of STAT3-mediated p27 induction.

A0 '""5 c0'"ic ""g '"c ;(0 I-c (f)[- -IIiIIIII

C

STAT B

N'"0""'";(I-( f)

1 S T A T 3- STAT B-B 107 -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _-

2E

o 1 2 3 4 5 6 7 8 9 10 11days in culture} vector control

--+----- STAT3jl clone 1)STAT3B clone 2)

Figure 10. Overexpression of STAT3[3 in 32D/G-CSF-R cells. A) Western blot analysis ofIysates from 32D transfectants. The blot was hybridized with anti-STAT3 antibodies againstamino acids 1-175). B) Proliferation of 3 D transfectants in response to G-CSF. The numberso viable cells were detennined on the basis o trypan blue exclusion at the indicated times.C) Neutrophilic differentiation of 32D transfectants overexpressing STAT3[3. Morphology of

cells cultured for 9 days in the presence o G-CSF May-Grtinwald-Giemsa staining; originalmagnification: x 630).

106


107/136

iscussion

STAT3 has recently been implicated in IL-6-induced monocytic and G-CSFinduced granulocytic differentiation 7,9), but which mechanisms are controlled bySTAT3 remained unclear. Here, we have shown that the inhibitory effects ofdominant-negative forms of STAT3 on G-CSF-induced neutrophilic differentiation of32D cells could be overridden by the cell cycle inhibitor Ara-C, suggesting thatSTAT3 does not control differentiation itself, but induces a growth arrest that isessential for the cells to undergo differentiation.

Previously, STATI has been implicated in interferon-induced GI atTest via thecdk inhibitor p21 WAFlielPI 29). More recently, thrombopoietin-inducedmegakaryocytic differentiation and growth anest was also functionally .linked toexpression of p21 32). In both studies, potential STAT responsive elements in thep21 promoter were recognized with binding activity for STAT I and STATS,suggesting that p21 is a direct target for these STATs. In contrast, STAT3 did notbind to these elements, predicting that STAT3-mediated cell cycle arrest occurs viaother mechanisms

We have shown here that STAT3 controls cell cycle arrest by enhancing theexpression of the p21 family member p27 K;PI via direct binding to STAT3-bindingelements in the p27 promoter region. Although thus far regulation of p27 expressionhas been mainly attributed to translational mechanisms 19-21), our data indicate thatp27 is under the direct transcriptional control of STAT3. t is of note however thatSTAT3-DNA binding was already maximal I day after G-CSF induction, whereasp27 mRNA levels increased after 2 to 3 days and only after S to 7 days of G-CSFstimulation p27 protein was maximally expressed. There may be several explanationsfor this lag in p27 induction. Firstly, it is possible that additional mechanismscontrolling STAT3-mediated transeription are involved, e.g. conformational changesof the STAT-containing complexes or interaction of the STAT complexes with othertranscription regulating proteins. We have attempted to directly determinetranscriptional activity of STAT3 during progressive differentiation of G-CSFstimulated 32D cells, using a reporter construct containing four copies of the acutephase response element in front of the minimal junB promoter linked to the Illciferasegene. Unfortunately, we did not succeed in rcproducible transfections ofdifferentiating 32D cells, and therefore no conclusive data were obtained. Secondly,induction of the p27 promoter might require transcriptional synergy between STAT3and other transcriptional regulators that are only induced after several days of G-CSFstimulation of 32D cells. The observation that neither dibutyryl-cAMP nor cAMPinducing agents cholera toxin, forskolin) significantly altered the kinetics of G-CSFinduced p27 expression in 32D cells De Koning et al. unpublished data) arguesagainst a direct contribution of the cAMP responsive element CRE) in the p27promoter. In agreement with this, elevation of intracellular cAMP levels did not affectthe proliferation and differentiation characteristics of 32D cells De Koning et al.unpublished data). A putative contribution of other candidate proteins with p27

1 7


108/136

promoter binding activity, e,g, Spl NFkB, and Myb, remains to be investigated 28).In view of this, it is of interest that Sp I physically interacts and synergizes withSTATI in IFN-y-induced activation of the intercellular adhesion molecule-I ICAMI) gene 36). Furthermore, induction of the CCAAT/enhancer binding protein /promoter by IL-6 requires transcriptional synergy between Spl and STAT3 37).Thirdly, p27 protein levels might remain low during the first 3 to 5 days of G-CSFstimulation due to ubiquitin-mediated degradation 38). Recent data indicate thatmitogen-stimulated Ras activity induces signaling via two independent pathways thatare both essential for cell cycle progression, namely a RaslERK pathway that inducescyclin 0 protein expression and a RaslRho pathway that results in ubiquitin-mediatedp27 degradation 39). Importantly, Ras activation plays a crucial role in G-CSFinduced cell cycle progression of 320 cells 24,40). Thus, it is possible that duringday 3 to 5 of G-CSF stimulation, when levels of activated Ras are high, low p27protein levels are due to protein degradation despite active STAT3-mediatedtranscription.

The splice variant STAT3P is specifically expressed in hematopoietic cellscapable of differentiation 35) and has some functional properties that are distinctfrom full-length STAT3. For instance, STAT3p, but. not STAT3, acts synergisticallywith c-Jun in activating a promoter containing the IL-6-responsive element of thearlll crogiobulill gene 41). t was therefore of interest to determine whetherSTAT3P is upregulated by G-CSF and could account for the delayed p27 expression.Expression levels of STAT3P did not increase during prolonged G-CSF stimulationof 320 cells. Furthermore, ectopic overexpression of STAT3P did not affect G-CSFresponses of 320 cells. These data suggest that STAT3P is not involved in G-CSFmediated p27 expression and also argue against a role of STAT3P as a dominantnegative regulator of STAT3-mediated transcription on the p27 promoter, as has beenshown for the ICAM-I promoter 25).

Previously, we detected point mutations in the G CSF R gene of patients withsevere congenital neutropenia who showed disease progression to acute myeloidleukemia 42). These mutations introduce premature stop codons between aminoacids 715 and 731 and result in the deletion of the C-terminal cytoplasmic region ofthe receptor that is essential for the induction of neutrophilic differentiation. Contraryto the transient proliferative signals provided by wild-type G-CSF-R, activation ofG-CSF-R mutants lacking the C-terminal differentiation domain results in sustainedand enhanced proliferation in myeloid cells. Such truncated G-CSF-R activate STAT3in serum-and growth factor-deprived cells at lower levels 16,43). Tyrosine 704 ofG-CSF-R mediates STAT3 activation in these G-CSF-R mutants, whereas wild-typeG-CSF-R also activates STAT3 via the C-terminal region that is deleted in thetruncated G-CSF-R. Whether the reduced levels of STAT3 activation induced by theG-CSF-R mutants are sufficient for p27 expression during prolonged G-CSFstimulation remains to be detennined. This knowledge may provide important cluesas to how G-CSF-mediated cell cycle control is affected in severe congenital

1 8


109/136

neutropenia by G-CSF-R mutations and how this could contribute to progression toacute myeloid leukemia.

After serum and growth factor deprivation of cells, G-CSF activates STAT ISTAT3, and STAT5 14-16), but levels of activation return to base line 60 minutesafter stimulation. The physiological significance of this early and transient peak ofST T activation is still unclear. Our present resuits show that, during progressiveG-CSF-induced differentiation, only the induction of STAT3-containing complexeswas robust and sustained. Importantly, besides STAT3 homodimers, also STAT3-STAT5 heteromers were formed, which were not detected during the early peak ofSTAT activation. The mobility of these STAT3-STAT5 complexes decreased after 3days of culture, suggesting that prolonged G-CSF treatment resulted in additionalconformational changes of these heteromeric complexes. Intriguingly, the appearanceof these slower migrating STAT3-STAT5 heteromers correlated more tightly with thekinetics of p27 induction than the formation of STAT3 homodimers. Further study ofthe composition and function of the STAT3-STAT5 heteromeric complexes shouldprovide insight into how these different STAT3-containing complexes contribute tocontrol the balance between proliferation and differentiation of myeloid cells viaupregulation of p27 protein.

References1. Demetri GO Griffin JD: Granulocyte colony stimulating factor and its receptor. Blood

78:2791, 19912. Isfort RJ, Ihle IN: Multiple haematopoietic growth factors signal through tyrosine

phosphorylation. Growth Factors 2:213, 19903. Dong F Van Paassen M, Van Buitenen C, Hoefsloot LH, Lowenberg B, Touw IP: A

point mut tion in the gr nulocyte colony-stimulating f ctor receptor G-CSF-R) gene in acase o acute myeloid leukemia results in the overexpression o a novel G CSF Risoform. Blood 85:902 1995

4. Quelle FW, Sato N, Witthuhn BA Inhom RC, Eder M, Miyajima A, Griffin JD, Ihle IN:Jak2 associates with the chain o the receptor for granulocyte macrophage colony-stimulating factor and its activation requires the membrane-proximal region. Mol CellBioI 14:4335 1994

5. Shuai K Horvath CM, Tsai Huang LH, Qureshi SA, Cowbum D, Damell JE: Interferonactivation o the transcription factor Stat91 involves dimerization through SH2phosphotyrosyl peptide interactions. Cell 76:821, 1994

6. Mui AL-F \Vakao H Kinoshita T Kitamura T Miyajima A: Suppression o interleukin-3 induced gene expression by a C-terminal truncated Stat5: role o Stat5 in proliferation.EMBO J 15:2425 19967. Nakajima K Yamanaka Y, Nakae K Kojima H Jehiba M, Killchi N, Kitaoka T, FukadaT Hibi M Hirano T: A central role for Stat3 in IL-6-induced regulation o growth anddifferentiation in M I leukemia cells. EMBO J 15:3651 1996

109


110/136

8. O'Farrell A-M, Uu Y, Moore K\V Mui AL-F: IL-JO inhibits macrophage activation andproliferation by distinct signaling mechanisms: evidence for Stat3-dependent and-independent pathways. EMBO 1 17: 1006, 1998

9. Shimozaki K Nakajima K Hirano T, Nagata S: Involvement o STAT3 in thegranulocyte colony-stimulating factor-induced differentiation of myeloid cells. J BioiChem 272:25184, 1997

10. Nicola NA: Granulocyte colony-stimulating factor and differentiation-induction inmyeloid leukemia cells. Int 1 Cell Cloning 5: I, 1987

II. Lieschke GJ Grail D, Hodgson G Metcalf D. Stanley E. Cheers C, Fowler KJ Basu S,Zhan YF, Dunn AR: Mice lacking granulocyte colony-stimulating factor have chronicneutropenia, granulocyte and macrophage progenitor cell deficiency, and impairedneutrophil mobilization. Blood 84: 1737, 1994

12. Bazan JF: Structural design and molecular evolution of a cytokine receptor superfamily.Proc Nat Acad Sci USA 87:6934, 199013. Fukunaga R Seto Y, Mizushima S, Nagata S: Three different mRNAs encoding human

granulocyte colony-stimulating factor receptor. Proc Natl Acad Sci USA 87:8702, 199014. Tian S-S, Lamb P Seidel HM, Stein RB, Rosen J: Rapid activation of the STATI

transcription factor by granulocyte colony-stimulating factor. Blood 84: 1760, 199415. Tian S-S, Tapley P Sincich C, Stein RB, Rosen J Lamb P: Multiple signaling pathways

induced by granulocyte colony-stimulating factor involving activation of JAKs, STAT5,and/or ST T} are required for regulation of three distinct classes of immediate earlygenes. Blood 88:4435, 199616. De Koning 1P, Dong F Smith L Schelen AM, Barge RMY, Van der PI as DC, HoefslootLH, Lowenberg B, Touw IP: The membrane-distal cytoplasmic region of humangranulocyte colony-stimulating factor receptor is required for STAT3 but not STAThomodimer formation. Blood 87:1335,1996

17. Dong F, Van Buitenen C, Pouwels K Hoefsloot LH, Lowenberg B, Touw IP: Distinctcytoplasmic regions of the human granulocyte colony-stimulating factor receptorinvolved in induction o proliferation and maturation. Mol Cell BioI 13:7774, 1993

18. Fukunaga R, Ishizaka-Ikeda E, Nagata S: Growth and differentiation signals mediated bydifferent regions in the cytoplasmic domain of granulocyte colony-stimulating factorreceptor. Cell 74: I 079, 1993

19. Alessandrini A, Chiaur DS Pagano M: Regulation of the cyclin-dependent kinaseinhibitor p27 by degradation and phosphorylation. Leukemia 11:342, 1997

20. Polyak K Lee M-H, Erdjument-Bromage H Koff A, Roberts JM Tempst P Massaguc JCloning of p27K P , a cyclin-dependent kinase inhibitor and a potential mediator ofextracellular antimitogenic signals. Cell 78:59, 1994

21. Toyoshima H Hunter T: p27, a novel inhibitor of GI cyclin-cdk protein kinase activity,is related to p21. Cell 78:67, 1994

22. Greenberger 1S, Sakakeeny MA, Humphries RK, Eaves C1 Eckner R1: Demonstration opemanent factor-dependent multipotential (erythroid/neutrophillbasophil) hematopoieticprogenitor cell lines. Proc Natl Acad Sci USA 80:2931, 1983

110


111/136

23. Morgenstern JP, Land H: Advanced mammalian gene transfer: high titre retroviralvectors with multiple drug selection markers and a complementary helper-free packagingcell line. Nucleic Acids Res 18:3587, 1990

24. De Koning JP, Soede-Bobok AA, Schelen AM, Smith L Van Leeuwen D, Santini VBurgering BMT, Bas JL, Lowenberg B, Touw IP: Proliferation signaling and activationo She, p21 Ras and Myc via tyrosine 764 o human granulocyte colony-stimulating factorreceptor. Blood 91:1924,1998

25. Caldenhoven E, Van Dijk TB, Solari R Annstrong J Raaijmakers JANI, Lammers J-WJ,Koenderman L, De Groot RP: STAT3f3 a splice variant of transcription factor STAT3, isa dominant negative regulator of transcription. J BioI Chern 271: 13221, 1996

26. Wagner BJ, Hayes TE, Hoban CJ, Cochran BH: The SLF binding element conferssis/PDGF inducibility onto the c-fos promotor. EMBO J 9:4477, 1990

27. Wakao H, Gouilleux F, Groner B: Mammary gland factor MGF) is a novel member othe cytokine regulated transcription factor gene family and confers to the prolactinresponse. EMBO J 13:2182, 1994

28. Kwon TK, Nagel JE, Buchholz MA, Nordin AA Characterization o the murine cyclindependent kinase inhibitor gene p27Kpl. Gene 180: 113, 1996

29. Chin YE, Kitagawa M, Su W-CS, You Z-H, Iwamoto Y Fu X-Yo Cell growth arrest andinduction o cyclin-dependent kinase inhibitor p2 WAFlfCIPI mediated by STATI. Science272:719, 1996

30. De Koning JP, Schelen AM Dong F Van Buitenen C Burgering BMT, Bas JL,Lowenberg B, Touw IP: Specific involvement o tyrosine 764 o human granulocytecolony-stimulating factor receptor in signal transduction mediated by p145/Shc/GRB2 orp90lGRB2 complexes. Blood 87: 132, 1996

31. Fero ML, Rivkin M Tasch M Porter P, Carow CE Firpo E, Polyak K Tsai LH, BroudyV Perlmutter RM, Kaushansky K Roberts JM: A syndrome of multiorgan hyperplasiawith features o gigantism, tumorigenesis, and female sterility in p27 Kip1 _deftcient mice.Cell 85:733, 1996

32. Matsumura I, Ishikawa J Nakajima K, Oritani K, Tomiyama Y, Miyagawa J-L, Kato T,Miyazaki H. Matsuzawa Y, Kanakura Y: Thrombopoietin-induced differentiation o ahuman megakaryoblastie leukemia cell line, CMK, involves transcriptional activation op21 WAFllelPI by STAT5. Mol Cell BioI 17:2933, 1997

33. Kato J-Y Matsuoka M, Polyak K Massague J Sherr CJ: Cyclic AMP-induced G I phasearrest mediated by an inhibitor p27 KiPi o cyclin-dependent kinase 4 activation. Cel179:487, 1994

34. Ward AC, Csar XF, Hoffmann BW, Hamilton JA: Cyclic AMP inhibits expression of Dtype cyclins and cdk4 and induces p27KPI in G-CSF-treated NFS-60 cells. BiochemBiophys Res Commun 224:10,1996

35. Chakraborty A White SM, Schaefer TS Ball ED, Dyer KF Tweardy OJ: Granulocytecolony-stimulating factor activation of STADa and STADf3 in immature normal andleukemic human myeloid cells. Blood 88:2442, 1996

36. Look DC, Pelletier MR, Tidwell RM, Roswit WT, Holtzman MJ: Stat l depends ontranscriptional synergy with Spl. J BioI Chem 270:30264,1995


112/136

37. Cantwell CA Sterneck E Johnson PF: Interleukin-6-specitic activation of the ClEBPgene in hepatocytes is mediated by Stat3 and Spl. Mol Cell Bioi 18:2108 1998

38. Pagano M, Tam SW Theodoras AM Romero-Beer P Del Sal G, Chau Y, Yew R,Draetta G Rolfe M: Role o the ubiquitin-proteasome pathway in regulating abundanceof the cyclin-dependent kinase inhibitor p27. Science 269:682 1995

39. Weber JD Hu W, Jefcoat SC Raben riM Baldassare JJ: Ras-stirnulated extracellularsignal-related kinase I and RllOA activities coordinate platelet-derived growth factor-induced G progression through the independent regulation of cyclin D and p27Kipl JBioi Chern 272:32966 1997

40. Okuda K, Ernst TJ, Griffin Jb: Inhibition ofp21' activation blocks proliferation but notdifferentiation of interleukin-3-dependent myeloid cells. J Bioi Chern 269:24602 1994

41. Schaefer TS Sanders LK. Nathans D: Cooperative transcriptional activity o Jun andStat3p a short form ofStat3. Proc Nat Acad Sci USA 92:9097 199542. Dong F, Brynes RK Tidow N, Welte K, Lowenberg B Touw IP: Mutations in the genefor the granulocyte colony-stimulating factor receptor in patients with acute myeloidleukemia preceded by severe congenital neutropenia. New Engl J Med 333:487 1995

43. \Vard AC Van Aesch YM Schelen AM Touw IP: Defective internalization andsustained activation o truncated granulocyte colony-stimulating factor receptor found insevere congenital neutropenia/acute myeloid leukemia. Blood 93:447 1999

112


113/136

CHAPTER

eneral discussion


114/136

G-CSF-R transduces signals that control proliferation and survival of myeloidprogenitor cells, and their differentiation towards neutrophilic granulocytes. Studiesdescribed in this thesis were performed to identify the signal transduction pathwaysactivated by G-CSF-R that are coupled to proliferation and differentiation induction.In particular, the involvement of signals mediated by the cytoplasmic tyrosineresidues of G-CSF-R was investigated, such as activation of the p2 R pathway andSTAT proteins. Tyrosine kinase activity induced by G-CSF resuits in thephosphorylation of the four conserved cytoplasmic tyrosines located in themembrane-distal region of G-CSF-R, providing binding sites for SH2 domaincontaining signaling proteins. To determine the role of these tyrosines in G-CSFsignaling, C-terminal deletion mutants and tyrosine-to-phenylalanine substitutionmutants of G-CSF-R were constructed. The signal transduction abilities of theG-CSF-R mutants were tested in transfectants of the mouse pro-B cell line BAF3, andthe myeloid cell line 32D. Involvement of the described signaling pathways inG-CSF-mediated cellular responses was investigated in a differentiation competentsubclone of 32D cells that lacks endogenous G-CSF-R. In 32D cells transfected withwild-type G-CSF-R cDNA, G-CSF induces transient proliferation followed byterminal neutrophilic differentiation (Chaptcr 5). This is cun-ently the most relevantcell line model available to study the effects of G-CSF-R mutants or signaling proteinmutants on G-CSF-driven neutrophil development.

7 1 Mechanisms ofp21 ' and STAT ctiv tion by G-CSF-RIn Chapter 2, it is shown that G-CSF stimulation of BAF3 cells expressing

G-CSF-R proteins induces the formation of several signaling complexes of the p21pathway, namely p145/Shc/GRB2 p901GRB2 and SHP-2/GRB2 complexes (FigureI). The C-terminal region of G-CSF-R containing the four tyrosine residues isessential for activation of these complexes. G-CSF-induced SHP-2/GRB2 associationappeared to be partly mediated via Y704 and partly via other domain(s) of theC-terminal region of G-CSF-R. In contrast, Y764 of G-CSF-R is indispensable for theformation of pl45/Shc/GRB2 and p90lGRB2 complexes. These observations wererecently further substantiated in our laboratory by binding experiments withrecombinant purified SH2 domains of several signaling proteins and with, on singletyrosincs, phosphorylated mutant G-CSF-R proteins. Far Westem blotting techniqueswere used to show that SHP-2 can directly bind to Y704 as well as to Y764, whereasboth Shc and GRB2 can only bind directly to Y764 of G-CSF-R (1,2). All thesefindings suggest an important role for Y764 of G-CSF-R in G-CSF-mediated p2 Rsignaling. This was confirmed by p2 R -Ioading assays showing that G-CSF-inducedp21 activation is severely reduced by substitution of Y764 (Chapter 5).The same BAF3 cell transfectants expressing the various G-CSF-R mutants wereused to determine which cytoplasmic regions of G-CSF-R are involved in STATactivation. In Chapter 3, it is shown that the C-terminal region of G-CSF-R containing

4


115/136

Y704 -----. ~ ~ ~

< t = e l e1 -.-J She GRB eY729

~ ? 9 1Y764

o - - - -8 -= 1 prolifer tion - - - - - - - - - - - - - - - - - - -igure 1 Signal transduction pathways activated by G CSF R. Data described in this thesis

are summarized in a schematic diagram.the cytoplasmic tyrosine residues is important for G-CSF-induced activation ofSTAT3 but not for STAT activation. Like the formation of SHP-2/GRB2 complexes,STAT3 activation is mediated via Y704 as well as via other domain(s) of theC-terminal region of G-CSF-R (Figure I . To further study these alternativemechanisms of STAT3 activation, morc detailed experiments were recentlyperformed using G-CSF-R mutants in which all four tyrosine residues weresubstituted or which retain just one of the cytoplasmic tyrosines. In the BAF3 celltransfectants, the C-terminal region of G-CSF-R appeared to mediate STAT3activation in a phosphotyrosine-dependent manner, via direct binding of STAT3 toeither Y704 or Y744, and in a phosphotyrosine-independent manner, via as yetunknown interactions (3). Although not included n this thesis, we and others havealso shown that G-CSF-induced activation of STATS, like STATI activation, doesnot depend upon the C-terminal region and therefore does not requirephosphotyrosine residues of G-CSF-R 4,S). Thus, the membrane-proximal region ofG-CSF-R that is involved in Iak binding and activation is essential and sufficient foractivation of STAT I and STATS (Figure I). In Chapter 3 the possibility is discussedthat, in these cases, the Iak kinases may specifically recruit the STAT proteins. Insupport of this, it has recently been shown that the kinase-like JH2 domain of Iaks caninteract with STAT proteins, resulting in direct STAT activation independent ofreceptor phosphorylation 6).

liS


116/136


117/136

Evidence is provided that STAT3 induces cell cycle exit via upregulation of the cdkinhibitor p27K1pl (Figure I). Since disruption of the STA T3 gene results in very earlyembryonic lethality due to unknown deficiencies, these observations could not beconfirmed in STAT3-deficient cells (16). Besides the effects of STAT3 onproliferation, several recent studies have suggested involvement of STAT3 in survivalsignaling (17-19). Interestingly, preliminary experiments with 320 cells expressingG-CSF-R mutants in which all four tyrosine residues were substituted or which retainjust one of the cytoplasmic tyrosines show a clear correlation between the ability ofthese mutants to activate STAT3 and to prevent apoptosis (Ward et al., unpublisheddata). Interference of dominant-negative STAT3 mutants with G-CSF-inducedsurvival signals is currently unller investigation.

Whereas STAT3 seems to have a negative effect on mitogenesis, ST TSactivation has been implicated in proliferation induction. For instance, it was shownthat expression of a constitutively active ST TS mutant in IL-3-dependent BAF3cells is capable of inducing cytokine-independent growth (20). Conversely, adominant-negative STATS protein partially inhibited IL-3-induced proliferation ofBAF3 cells (21). Furthermore, STATSAIB-deficient mice were recently generatedthat showed a reduced response of bone marrow cells to JL-3 as well as to G-CSF inill vitro colony assays, suggesting involvement of STATS in IL-3- and G-CSFmediated cell growth (22). Whether STATS is important for G-CSF-inducedproliferation of 320 cells is currently examined by transfection of constitutivelyactive or dominant-negative ST TS mutants.

In Chapter S the role of the cytoplasmic tyrosine residues of G-CSF-R in thetransduction of proliferation and differentiation signals was determined by expressingtyrosine-to-phenylalanine substitution mutants in 320 cells that lack endogenousG-CSF-R. All tyrosines could be replaced essentially without affecting theneutrophilic differentiation signaling properties of G-CSF-R. However, substitution ofone specific tyrosine, i.e. Y764, markedly influenced proliferation induction, whereassubstitution of Y704, Y729, or Y744 had no effect. G-CSF-induced cell cycleprogression from the G I to the S phase as well as activation of Shc and p21 R and theinduction of C-IIl) C expression were abrogated by mutation of Y764. The potentialinvolvement of these signaling events in proliferation induction is extensivelydiscussed in Chapter S. Additional evidence supporting the observation that Y764 ofG-CSF-R is important for proliferation signaling is provided by recent experimentswith a triple tyrosine-to-phenylalanine substitution mutant which retains only Y764(mO), and a quadruple null mutant with no cytoplasmic tyrosines (mO). Strikingly,mutant mO appeared incapable of transducing proliferation signals in 320 celltransfectants, whereas 320 cells expressing mutant mO proliferated continuously inresponse to G-CSF (I). Besides confirming the importance of Y764 for proliferationinduction, these results also suggest that proliferation inhibitory signals are given via(one of) the other cytoplasmic tyrosines of G-CSF-R. This becomes clear bycomparing the transient with the sustained proliferation from wild-type G-CSF-R andmutant mO, respectively. Because STAT3 activation is mediated via interaction with

117


118/136

JY704 Y704

~ T A Y729 Y729 Y 7 ~ l YliY76 Y764 . 9

f \~ 0 ~ ~


119/136

pathway that results in ubiquitin-mediated p27K P degradation (Figure 2) (24). Duringprolonged G-CSF stimulation of myeloid cells, p21 , activation might be inhibited,possibly resulting in decreased cyclin D expression and decreased p27 K1pldegradation, and thereby in growth anest. One possible mechanism of p21 Rinhibition is via induction of GTPase activating protein (GAP) family members. Forinstance, the neurofibromatosis type I NFl) tumor-suppressor gene encodes aprotein (neurofibromin) that accelerates GTP hydrolysis on p21 R proteins.Interestingly, expression of neurofibromin is induced during myoblast differentiation(25). Furthermore, loss of NFl function results in increased p21 R , activation andleads to abelTant growth of myeloid cells (26,27). However, no evidence was obtainedso far that GAP proteins are recmited by the cytoplasmic tyrosines of G-CSF-R.

In addition to p21 R"" Myc has been implicated in the control of myeloid cellproliferation (discussed in Chapter 5). Myc is a transcription factor, which mustdimerize with Max proteins in order to regulate the activation of genes involved incell proliferation. One product of an as yet unknown Myc/Max target gene(s) hasbeen suggested to induce sequestration of p27 K1pI in a form unable to bind cyelin edkcomplexes (28). In this way, Myc could counteract growth anest by p27 K1Pl withoutp27K P degradation, which represents an essential aspect of Myc's mitogenicfunctions (Figure 2). During cellular differentiation, a reduction of Myc protein and aconcomitant increase in Mad or Mxi proteins is frequently observed (29,30). Mad andMxi proteins also form heterodimers with Max proteins and subsequently bind tosimilar sites as Myc/Max dimers. However, Mad/Max and MxilMax dimers represstranscription by ternary complex formation with the repressor Sin3, and therebyinhibit proliferation (31). During growth anest of 32D cells after prolonged G-CSFstimulation, a significant increase of Mad I expression was detected, whereas Mycexpression only slightly decreased (Soede-Bobok ef al., unpublished data). IncreasedMad expression may result in the repression of Myc/Max target genes, one of whichencodes the as yet unidentified protein involved in sequestration of p27K P away fromcyclin/cdk complexes. Besides the observed G-CSF-induced increase in p27Kipiprotein expression, also more p27Kipl protein can then bind to, and thereby inhibit,cyelin/cdk complexes. This may provide an additional mechanism of G-CSFmediated growth control. In agreement with this, granulocytic precursors of Mad 1deficient mice show an increased proliferative response to G-CSF at a late stage ofneutrophil development (32).

7,3 Differentiation signaling by GCSFRAlthough significant progress has been made in understanding G-CSFR

controlled myeloid cell proliferation, it is still unclear which signal transductionpathways are involved in GCSFinduced neutrophilic differentiation. In Chapter 5, itwas shown that mutation of Y704, Y729, Y744, or Y764 did not affect the ability ofG-CSF-R to transduce differentiation signals in 32D cells. Furthermore, G-CSF also

119


120/136

induccs terminal neutrophilic dilferentiation in 32D cells cxpressing a triple mutantwhich retains only Y704 (mA), indicating that differentiation signaling by theC-terminal domain of G-CSF-R is tyrosine-independent (33). Additional experimentswill be required to determine the signaling substrates that are recruited by this domainand initiate the differentiation process.

Because differentiation-specific signal transduction pathways have not yet beenidentified, the rolc of HGFs in the regulation of cellular differentiation remainscontroversial. wo general models for hematopoietic differentiation have beenproposed (34,35). In the stochastic model, the ability to differentiate is an intrinsicfeature of hematopoietic cells that s independent of HGFs. According to this model,HGFs only provide growth and survival signals that are required for the execution ofthe differentiation programs. In the instructive model, differentiation is activelyinduced by HGFs. The observation that the C-terminal domain of G-CSF-Rspecifically transduces neutrophilic differentiation signals in several myeloid cell linemodels (32D, L-GM, and FDC-P cells) provides evidence for the latter model(36,37). Notably, the same G-CSF-R domain s not functional n the pro-B cell lineBAF3, suggesting that signaling for neutrophilic differentiation by G-CSF-R requiresthe appropriate intracellular environment or genetic program of the responding cells(36).

Recently, it was shown that truncation of the C-terminal dilferentiation domainof G-CSF-R causes neutropenia n a knock-in mouse model, providing ll vivoevidence that the G-CSF-R C-terminus is essential for normal neutrophil production(38). However, bone marrow cells from these mice differentiated ll vitro tomorphologically mature neutrophils in response to G-CSF. Similarly, G-CSF-Rdeficient mice have chronic neutropenia but do not show accumulation o immaturegranulocytic cells n the bone marrow, suggesting normal differentiation of theresidual granulocytic precursors (39). These results indicate that G-CSF-R is a majorregulator of granUlopoiesis, although, at least n mice, a G-CSF-R-independentmechanism of neutrophil development also appears to exist. This conclusion is stillcompatible with the cell line data and the instructive model, since altelllativegranulocytic differentiation signals (e.g. from other HGFs) that are absent n the cellline models may compensate for the loss of G-CSF-R signals in primary cells.

Additional evidence supporting the instructive model of differentiation ispresented n Chapter 5. In the presence of the cell cycle inhibitor cytosine arabinoside(Ara-C), 32D cells expressing wild-type G-CSF-R atTest in G I, gradually loseviability, and die after 4 to 5 days. Whereas cells cultured with IL-3 remainmyeloblastic during this period, G-CSF induces terminal neutrophilic differentiationafter 2 to 4 days, indicating that enforced cell cycle arrest does not inducedifferentiation of 32D cells n the absence of G-CSF. In agrcement with this,suppression of apoptosis by overexpression of bcl-2 or bel-Xc is not sufficient tosupport myeloid differentiation of 32D cells (40). Furthermore, Ara-C-treated 32Dcells expressing a G-CSF-R mutant lacking the C-terminal differentiation-domainremain myeloblastic in both IL-3- and G-CSF-containing medium (De Koning et al.120


121/136

unpublished data). Taken together, these results suggest that neutrophilicdifferentiation requires both downmodulation of proliferation and, via the C-terminalcytoplasmic region of G-CSF-R, induction of specific differentiation signals Figure3).

M

Y7 4

Y729I Z 4 ~Y764

specific ?differentiation .signals

G2 r 1 arrest jS

neutrophilicdifferentiation

Y7 4

Y729

~ : < ~ 0 __ -@

proliferationigure 3. A model for the balance between G-CSF-R-controlled proliferation and

differentiation induction during neutrophil development. Signaling via Y764 of G-CSF-R,which involves p 2 1 R a ~ and Myc activation, is essential for G-CSF-induced proliferation. Incontrast, STAT3 inhibits proliferation of myeloid cells via upregulation of p27Kipt Besidesaccumulation of cells in the G I phase of the cell cycle which is a prerequisite fordifferentiation, specific differentiation signals induced via the C-terminal cytoplasmic regionof G-CSF-R are required for terminal neutrophilic differentiation.

2


122/136

7.4 Function of mutant G-CSF-R iu severe congeuital neutropeniaIn 5 to 20 of patients with severe congenital neutropenia (SCN). mutations

are found in the G CSF R gene (41,42). These mutations introduce premature stopcodons resulting in the truncation of 82 to 98 C-terminal amino acids (Figure 4). SCNpatients who develop AML almost invariably acquired a G-CSF-R mutation,suggesting that this genetic alteration represents a key step in leukemogenesis. Uponectopic expression in myeloid cell lines, the truncated G-CSF-R consistently fails totransduce differentiation signals, whereas proliferation signaling by these receptors issustained and enhanced. Moreover, the truncated receptors interfere with signalinginduced via wild-type G-CSF-R in a dominant-negative manner (41). This dominanthyperproliferative function of truncated G-CSF-R was recently confirmed in a'knock-in' mouse model with an equivalent G-CSF-R mutation (38). Daily treatmentof these mice with G-CSF results in an increased production of neutrophils, leading toa sustained neutrophilia. This indicates that the G-CSF-R mutation itself is sufficientfor a hyperproliferative response to G-CSF, supporting the idea that G-CSF-Rtruncation represents a preleukemic event

To identify the mechanisms mediating the dominant hyperproliferative functionof the truncated G-CSF-R, receptor activation and internalization studies wereperformed with both primary cells and 32D cell transfectants (33,43). G-CSF-induced

W seN

} }Y704

Y729differentiation: N 1 4 ~ 1internalizationY76

Figure 4. Schematic diagram o truncated G-CSF-R from Se patients. Nonsense mutationsin the G CSF R gene o seN patients result in truncations o the C-terminal cytoplasmicregion. The domains o wild-type \VT) and truncated G-CSF-R involved in induction odifferentiation. internalization and STAT activation are indicated.

22


123/136

intemalization of truncated G-CSF-R appeared to be severely impaired which resultsin prolonged receptor activation. Importantly, mutant receptors were found to act in adominant-negative manner over wild-type receptors with regard to both receptoractivation and intemalization. Currently, experiments are in progress to determine thesignaling mechanisms activated via the C-terminal domain of G-CSF-R that mediatereceptor internalization. These findings suggest that extended signaling by truncatedG-CSF-R due to defective internalization results in hyperproliferation. In support ofthis, activation of STAT5, implicated in proliferation induction, is significantlyprolonged due to receptor truncation 33,43). n contrast, the duration of STAT3activation is hardly affected, because wild-type G-CSF-R already induces sustainedactivation of STAT3 Chapter 6). However, the level of STAT3 activation bytruncated G-CSF-R is clearly reduced compared to wild-type G-CSF-R. Most likely,this is due to the fact that the truncated G-CSF-R can only activate STAT3 via Y704,whereas STAT3 recruitment by wild-type G-CSF-R can also occur via Y744 ofG-CSF-R Figure 4). Whether the reduced levels of STAT3 activation by the G-CSFR mutants prevent induction of p27 K;pl protein expression during prolonged G-CSFstimulation remains to be determined. Although the increased STAT5 and thedecreased STAT3 activation may already result in enhanced proliferation, it isobvious that sustained activation of other signaling pathways induced by the truncatedG-CSF-R may also contribute to the observed hyperproliferation.

In summary, these results implicate G-CSF-R truncations as possiblepreleukemic events. t is conceivable that cells with an acquired G-CSF-R mutationmay have a growth advantage due to prolonged receptor activation in the presence ofligand.

7.5 uture directionsAlthough it is still unclear which signal transduction pathways that are activated

by G-CSF-R mediate neutrophilic differentiation, recent studies have suggested thatseveral transcription factors are important for myeloid differentiation. Evidencesupporting the involvement of transcription factors in myelopoiesis include theobservations that many genes cloned at the site of leukemic translocation breakpointsare transcription factors, and that knockout mice of these factors show severe defectsin myeloid development. The transcription factors shown to be important forinduction of myeloid-specific genes are PU.1 Spi-I), CIEBPa CIEBPE, AMLlCBF, c-Myb, RARa PLZF, and MZFI 44). Their activity is modulated by coactivator and co-repressor proteins that create a physical bridge between transcriptionfactors and the basal transcriptional machinery. Whether transcription factors eitheractivate or repress transcription might be regulated by their ability to associate withhistone acetyltransferases co-activators; e.g. p300/CBP, pCAF) or histone acetylasesco-repressors; e.g. N-CoR, mSin3), respectively. Furthermore, it has recently been

shown that the activity of a single transcription factor, Pit-I can be modulated by123


124/136

distinct signal transduction pathways, not through modification of Pit I itself, butthrough regulation of the recruited co activator complex (45). How transcriptionfactors are induced during myeloid differentiation and how they fit into G CSF Rsignaling cascades remains an important goal for future research.

ReferencesI Ward AC, Smith L De Koning JP, Van Aesch Y Touw LP: MUltiple signals for

proliferation differentiation anp survival from the granulocyte colony-stimulating factorreceptor in myeloid 32D ceJls. In preparation

2. Ward AC, Monkhouse JL, Hamilton JA, Csar XF: Direct binding of shc, grb2, SHP-2and p40 to the murine granulocyte colony-stimulating factor receptor. Biochim BiophysActa 1448:70, 1998

3. \Vard AC, Hennans MBA, Smith L Van Aesch YM. Schelen AM, Antonissen C, TouwIP: Tyrosine-dependent and -independent mechanisms of STAT3 activation y thehuman granulocyte colony-stimulating factor G-CSF) receptor are differentially utilizeddepending on G-CSF concentration. Blood 93: 113 1999

4. Tian S-S Tapley P, Sincich C Stein RB, Rosen J Lamb P: Multiple signaling pathwaysinduced by granulocyte colony-stimulating factor involving activation of JAKs STATS,and/or STAT3 are required for regulation of three distinct classes of immediate earlygenes. Blood 88:4435, 1996

5. Dong F Liu X De Koning JP, Touw lP Henninghausen L Lamer A Grimley PM:Stimulation of StatS by granulocyte colony-stimulating factor G-CSF) is modulated bytwo distinct cytoplasmic regions of the G-CSF receptor. J Immunol 161 :6503, 1998

6. Fujitani Y Hibi M Fukada T Takahashi-Tezuka M Yoshida H Yamaguchi TSugiyama K Yamanaka Y Nakajima K Hirano T: An alternative pathway for STATactivation that is mediated by the direct interaction between JAK and STAT. Oncogene14:751,1997

7. Meydan N, Grunberger T, Dadi H Shahar M Arpaia E Lapidot Z Leeder JS, FreedmanM Cohen A Gazit A Levitzki A Roifman CM: Inhibition of acute lymphoblasticleukaemia by a Jak-2 inhibitor. Nature 379:645, 1996

8. Nelson KL Rogers JA, Bowman TL Jove R Smithgal TE: Activation of STATI by thec-Fes protein-tyrosine kinase. J Bioi Chem 273:7072, 1998

9. Saharinen P Ekman N, Sarvas K Parker P Alitalo K Silvennoinen : The Bmx tyrosinekinase induces activation of the Stat signaling pathway which is specitically inhibited byprotein kinase Cdelta. Blood 90:4341, 1997

10. Chaturvedi P Reddy MY, Reddy EP: Src kinases and not JAKs activate STATs duringIL-3 induced myeloid cel proliferation. Oncogene 16:1749, 1998

II. Corey SJ, Burkhardt AL, Bolen JB Geahlen RL Tkatch LS, Tweardy DJ Granulocytecolony-stimulating factor receptor signaling involves the formation of a three-componentcomplex with Lyn and Syk protein-tyrosine kinases. Proc Natl Acad Sci USA 91 :4683,1994

24


125/136

12. Corey SJ Dombrosky-Ferlan PM Zuo S Krohn E Dannenberg AD Zorich P RomeroG Takata M Kurosaki T: Requirement of Src kinase Lyn for induction of DNsynthesis by granulocyte colony-stimulating factor. J Bioi Chem 273:3230, 1998

13. Boyer MJ, Romero G, Gomez-Cambronero J, Xu S, Dombrosky-Ferlan P Kurosaki T,Corey SJ: Submitted

14. Parganas E, Wang D Stravopodis D, Topham DJ, Marine JC, Teglund S, Vanin EF,Bodner S, Colamonici OR, Van Deursen JM, Grosveld G hIe IN: Jak2 is essential forsignaling through a variety of cytokine receptors. Cell 93:385, 1998

15. Rodig SJ, Meraz MA, White JM, Lampe PA, Riley JK, Arthur CD King KL, SheehanKC Yin L Pennica D Johnson EM Schreiber RD: Disruption of the l ki genedemonstrates obligatory and nonredundant roles of the Jaks in cytokine-induced biologicresponses. Cell 93:373, 1998

16. Takeda K Noguchi K Shi W Tanaka T Matsumoto M Yoshida N Kishimoto T AkiraS: Targeted disruption of the mouse Stat gene leads to early embryonic lethality. ProcNat Aead Sci USA 94:3801, 1997

17. Zushi S Shinomura Y Kiyohara T Miyazaki Y Kondo S Sugimachi M HigashimotoY Kanayama S Matsuzawa Y: STAT3 mediates the survival signal in oncogenic ras-transfeeted intestinal epithelial cells. Int J Cancer 78:326, 1998

18. Takeda K Kaisho T Yoshida N Takeda J Kishimoto T Akira S: Stat3 activation isresponsible for IL-6-dependent T cell proliferation through preventing apoptosis:generation and characterization of T cell-specific Stat3-deficient mice. J Immunol161:4652,199819. Fukada T Hibi M Yamanaka Y Takahashi-Tezuka M Fujitani Y Yamaguchi TNakajima K Hirano T: Two signals are necessary for cell proliferation induced by acytokine receptor gp130: involvement of ST TJ in anti-apoptosis. Immunity 5:449, 1996

20. Onishi M Nosaka T Misawa K Mui AL Gorman D McMahon M Miyajima AKitamura T: Identification and characterization of a constitutively active STAT5 mutantthat promotes cell proliferation. Mol Cell BioI 18:3871, 1998

21. Mui AL-F Wakao H Kinoshita T Kitamura T Miyajima A: Suppression of interleukin-3-induced gene expression by a C-terminal truncated Stat5: role of Stat5 in proliferation.EMBO J 15:2425, 199622. Teglund S, McKay C, Schuetz E, Van Deursen JM, Stravopodis D Wang D Brown M,Bodner S Grosveld G Ihle IN: Stat5a and Stat5b proteins have essential andnonessential or redundant roles in cytokine responses. Cell 93:841, 1998

23. Okuda K Ernst TJ, Griffin JD: Inhibition of p21ras activation blocks proliferation butnot differentiation of interleukin-3-dependent myeloid cells. J Bioi Chem 269:24602,1994

24. Weber JD, Hu W Jefcoat SC, Jr., Raben DM, Baldassare JJ: Ras-stimulated extracellularsignal-related kinase and RhoA activities coordinate platelet-derived growth factorinduced Gl progression through the independent regulation of cyelin DI and p27. BioiChem 272:32966, 1997

25. Gutmann DH, Cole JL, Collins FS: Modulation of neurotlbromatosis type I geneexpression during ll vitro myoblast differentiation. J Neurosci Res 37:398 1994

125


126/136

26. Bollag G Clapp DW, Shih S, Adler F Zhang YY, Thompson P Lange BJ, FreedmanMH. McConnick F lacks T Shannon K: Loss of NFl results in activation of the Rassignaling pathway and leads to aberrant growth in haematopoietic cells. Nat Genet12:144,1996

27. Largaespada DA, Brannan CI, Jenkins NA, Copeland NG: Nfl deticiency causes Rasmediated granulocyte/macrophage colony stimulating factor hypersensitivity and chronicmyeloid leukaemia. Nat Genet 12: 137, 1996

28. Vlach J Hennecke S, Alevizopoulos K Conti D, Amati B: Growth arrest by the cyclindependent kinase inhibitor p27Kip I is abrogated by c-Myc. EMBO J 15:6595, 1996

29. Ayer DE Eisenman RN: A switch from Myc:Max to Mad:Max heterocomplexesaccompanies monocyte/macrophage differentiation. Genes Dev 7:2110 1993

30. CuItraro eM Bino T Segal S: Function of the c-Myc antagonist Mad I during amolecular switch from proliferation to differentiation. Mol Cell Bioi 17:2353 199731. Bemards R: Transcriptional regulation. Flipping the Myc switch. Curr Bioi 5:859, 1995

32. Foley KP, McArthur GA, Queva C, Hur/in PJ, Soriano P Eisenman RN: Targeteddisruption of the MYC antagonist MADI inhibits cell cycle exit during granulocytedifferentiation. EMBO J 17:774, 1998

33. Ward AC Van Aesch YM Schelen AM T llW lP: Defective internalization andsustained activation of tmncated granulocyte colony-stimulating factor receptor found insevere congenital neutropenia/acllte myeloid leukemia. Blood 93:447 1999

34. Enver T, ileyworth CM, Dexter TM: Do stem cells play dice? Blood 92:348, 199835. Metcalf D: Lineage commitment and maturation in hematopoietic cells: the case forextrinsic regulation. Blood 92:345, 199836. Dong F Van Buitenen C, Pouwels K Hoefsloot LH, Lowenberg B, Touw P: Distinct

cytoplasmic regions of the human granulocyte colony-stimulating factor receptorinvolved in induction of proliferation and maturation. Mol Cell BioI 13:7774 1993

37. Fukunaga R Ishizaka-Ikeda E Nagata S: Growth and differentiation signals mediated bydifferent regions in the cytoplasmic domain of granulocyte colony-stimulating factorreceptor. Cell 74: 1079, 1993

38. Hemlans MH \Vard AC Antonissen C Karis A Lowenberg B Touw IP: Perturbedgranulopoiesis in mice with a targeted mutation in the granulocyte colony-stimulatingfactor receptor gene associated with severe chronic neutropenia. Blood 92:32 1998

39. Liu F Wu HY Wesselschmidt R Kornaga T Link DC: Impaired production andincreased apoptosis of neutrophils in granulocyte colony-stimulating factor receptordeficient mice. Immunity 5:491, 1996

40. Rodel JE, Link DC: Suppression of apoptosis during cytokine deprivation of 32D cells isnot sufticient to induce complete granulocytic differentiation. Blood 87:858 1996

41. Dong F Brynes RK, Tidow N Welte K Lowenberg B, Touw P: Mutations in the genefor the granulocyte colony-stimulating-factor receptor in patients with acute myeloidleukemia preceded by severe congenital neutropenia. N Engl J Med 333:487, 199542. Dong F Hoefsloot LH, Schelen AM, Broeders CA, Meijer Y Veerrnan AJ, Touw P,Lowenberg B: Identification of a nonsense mutation in the granulocyte-colony-

26


127/136

stimulating factor receptor in severe congenital neutropenia. Proc Natl Acad Sci USA91 :4480, 1994

43. Hermans MHA, Antonissen C Ward AC, Mayen AEM, Ploemacher RE, Touw IP:Sustained receptor activation and hyperproliferation in response to granulocyte colonystimulating factor G-CSF) in mice with an SCN/AML-derived mutation in the G-CSFreceptor gene. J Exp Med, 189: in press, 1999

44. Tenen DO Hromas R Licht JD Zhang DE: Transcription factors normal myeloiddevelopment, and leukemia. Blood 90:489, 1997

45. Xu L Lavinsky RM, Dasen JS, Flynn SE, Mcinerney EM, Mullen TM, Heinzel T, SzetoD, Korzus E, Kurokawa R, Aggarwal AK, Rose DW, Glass CK, Rosenfeld MG: Signalspecific co-activator domain requirements for Pit-l activation. Nature 395:30 I 1998

27


128/136


129/136

SUMMARY

SAMENV ATTING Summary in Dutch)


130/136

Summary

G-CSF is the most important growth factor involved in granulopoiesis andmaintenance of neutrophil levels in the peripheral blood. t stimulates proliferationand survival of myeloid progenitor cells, and their differentiation towards neutrophilicgranulocytes. G-CSF exerts its function via activation of a membrane receptor,G-CSF-R. In Chapter I an overview is given of the cunent understanding of thefunction of G-CSF-R in normal granulopoiesis as well as in some patients with severecongenital neutropenia and acute myeloblastic leukemia, diseases characterized bydisturbed myeloid differentiation. The cytoplasmic domain of G-CSF-R containsdiscrete functional regions, suggesting the existence of multiple signaling pathwaysthat are activated via these regions to elicit distinct biological effects. The C-terminalregion of the human G-CSF-R contains four conserved tyrosine Y) residues, locatedat positions 704, 729, 744, and 764, that could be involved in G-CSF-R-mediatedsignal transduction via interaction with SH domain-containing signaling proteins.In Chapter 2, the ability of wild-type, C-terminal deletion mutants, and tyrosineto-phenylalanine substitution mutants of G-CSF-R to activate signaling intermediatesof the p2 R>< pathway was examined. Several different Shc- and/or GRB2-containingcomplexes were formed after activation of G-CSF-R, namely p 14S/Shc/GRB2,p90/GRB2, and SHP-2/GRB2 complexes. Neither of these complexes was detectedafter activation of a C-terminal deletion mutant of G-CSF-R that lacks all fourconserved cytoplasmic tyrosine residues. G-CSF induced the formation ofSHP-2/GRB2 complexes in all the tyrosine-substitution mutants, suggesting that thisassociation did not depend on the presence of single specific tyrosine residues inG-CSF-R. In contrast, Y764 of G-CSF-R appeared to be indispensable for theformation of pI4S/She/GRB2 and p90/GRB2 complexes, suggesting a prominent rolefor Y764 ofG-CSF-R in G-CSF signal transduction.In Chapter 3, the cytoplasmic regions of G-CSF-R involved in activation ofSTAT I and STAT3 were determined. Studies with C-terminal deletion mutants andtyrosine-to-phenylalanine substitution mutants of G-CSF-R indicated that Y704 ofG-CSF-R, located within a YXXQ consensus sequence for STAT3-SH2 binding, isinvolved in the recruitment and activation of STAT3 by G-CSF-R. However, STAT3binding is not exclusively mediated by Y704, but also by other domains of theC-terminal region of G-CSF-R via as yet unknown interactions. In contrast, G-CSFinduced activation of STAT I did not depend upon the membrane-distal cytoplasmicregion of G-CSF-R, and therefore does not require phospho tyrosine residues of thereceptor.

In Chapter 4, the role of Jak kinases in activation of the p2 R pathway wasinvestigated by constructing a mutant in which tryptophan 650 in the membraneproximal region of G-CSF-R was replaced by arginine. This replacement abolishedJak phosphorylation and abrogated the mitogenic response to G-CSF. Significantly,the ability of G-CSF-R to induce formation of pI4S/Shc/GRB2, p90/GRB2, and130


131/136

SHP-2/GRB2 complexes was also completely blocked as a result of this amino acidsubstitution. These data indicate that the membrane-proximal cytoplasmic region ofG-CSF-R is not only crucial for proliferative signaling and activation of Jak kinases,but is also required for activation of signaling complexes of the p21 R pathway,which occurs via the membrane-distal region of G-CSF-R.In Chapter 5, the role of the cytoplasmic tyrosine residues of G-CSF-R intransduction of proliferation and differentiation signals was determined by expressingtyrosine-to-phenylalanine substitution mutants in a differentiation competent subcloneof 32D cells that lacks endogenous G-CSF-R. All tyrosines could be replacedessentially without affecting the neutrophilic differentiation signaling properties ofG-CSF-R. However, substitution of one specific tyrosine, i.e. Y764, markedlyinfluenced proliferation induction as well as the timing of differentiation, whereassubstitution of Y704, Y729, or Y744 had no effect on proliferation signaling. MutantY764F failed to support G-CSF-induced cell cycle progression from the OJ to the Sphase, resulting in accelerated differentiation and significantly reduced net productionof mature neutrophils. Importantly, G-CSF-mediated activation of She and p21 R andthe induetion of C-III) C expression were severely redueed by substitution of Y764.These findings indicate that Y764 of G-CSF-R is crucial for maintaining theproliferation/differentiation balance during G-CSF-driven neutrophil development,and suggest a role for multiple signaling mechanisms in maintaining this balance.

In Chapter 6, the involvement of STAT3 in G-CSF-indueed proliferation andneutrophilic differentiation was examined by overexpressing dominant-negativeSTAT3 mutants in differentiation competent 32D eells expressing the wild-typeG-CSF-R. STAT3 appeared to be essential for G-CSF-indueed neutrophiliedifferentiation by inducing a growth arrest that is a prerequisite for myeloid preeursoreells to proceed with differentiation. The splice variant STAT3P, previously shown tobe a dominant -negative form of STAT3 on eertain promoters, did not affect cellgrowth. To obtain an indication as to how STAT3 mediates a growth arrest of 32Dcells, the role of several cyclin-dependent kinase cdk) inhibitors was analyzed. Thelevel of the cdk inhibitor p27Kipi appeared to increase during G-CSF stimulation of32D cells. An oligonucleotide derived from the promoter region of p27 containing aputative STAT-binding site shifted with STAT3 in electrophoretic mobility shiftassays. Dominant-negative STAT3 reduced G-CSF-induced p27 promoter activity inluciferase reporter assays and interfered with the expression of both p27 mRNA andprotein. Furthermore, bone marrow and spleen mononuclear cells of p27 -deficientmice showed and increased proliferative capacity in response to G-CSF. These resultssuggest that STAT3-controlled cell cycle exit of myeloid precursors is mediated viadirect upregulation of the cdk inhibitor p27.

In Chapter 7 the findings described in this thesis are summarized and theirsignificance for the understanding of the function of G-CSF-R is discussed.

J3


132/136

Samenvatting Summary in Dutch)

De productie van neutrofiele granulocyten granulopoiese) wordt gereguleerddoor een netwerk van hematopoietische groeifactoren. Een van die groeifactoren,G-CSF, speelt een essentiele rol in dit proces. G-CSF stimuleert de proliferatie enoverleving van de granulocytaire voorlopercellen en induceert hun differentiatie totneutrofiele granulocyten. Deze verschillende effecten worden gelnitieerd na bindingvan G-CSF aan specifieke receptor eiwitten G-CSF-R) op het celoppervlak, watresulteert in activatie van intracellulaire signaal transductie routes. In Hoofdstllkwordt een overzicht gegeven van de huidige inzichten in de functie van G-CSF-R inzowel normale granulopoiese als bij patienten met emstige aangeboren neutropenic enacute myeloide leukemie. In het cytoplasmatische domein van G-CSF-R kunnenmeerdere subdomeinen met specifieke functies worden herkend. Dit suggereert datdoor deze subdomeinen verschillende signaalwegen worden geactiveerd om deafzonderlijke biologische effecten te induceren. Het C-terminale deel van de humaneG-CSF-R bevat 4 geconserveerde tyrosine residuen Y704, Y729, Y744 en Y764), diebetrokken zouden kunnen zijn bij G-CSF-R g e m e i e ~ r e signaal transductie doorinteracties aan te gaan met SH2 domeinen van signaal eiwitten.

In Hoofdstllk 2 wordt beschreven welke eiwitten van de p2 R route na G-CSFstimulatie worden geactiveerd. G-CSF-R activatie blijkt de vorming van verschillendeShc en/of GRB2 bevattende complexen pI45/Shc/GRB2, p90/GRB2 en SHP-2/GRB2) te bewerkstelligen. Deze complexen worden niet gevormd na activatie vaneen C-terminale deletie mutant van G-CSF-R, waarin de 4 cytoplasmatische tyrosinesontbreken. Met behulp van tyrosine-naar-phenylalanine substitutie mutanten isvervolgens aangetoond dat G-CSF gelnduceerde SHP-2/GRB2 associatie nietafhankelijk is van een specifieke tyrosine van G-CSF-R. Daarentegen blijkt Y764essentieel te zijn voor de vonning van zowel p l45/Shc/GRB2 als p90/GRB2complexen. Dit suggereert een belangrijke rol voor Y764 van G-CSF-R in G-CSFgelnduceerde signaal transductie.

In Hoofdstllk 3 wordt uiteengezet welke cytoplasmatische subdomeinen vanG-CSF-R een rol spelen in de activatie van STATI en STAT3. Y704 van G-CSF-R isgelegen in een zogenaarnde YXXQ consensus sequentie, waarvan bekend is dat heteen interactie aan kan gaan rnet het SH2 dornein van STAT3. Uit experimenten metC-terminale deletie en tyrosine-naar-phenylalanine substitutie mutanten van G-CSF-Rblijkt dat Y704 inderdaad betrokken is bij activatie van STAT3. Het C-terminale deelvan G-CSF-R bevat naast Y704 ook andere dorneinen die, via nog onbekendernechanismen, STAT3 activatie bewerkstelligen. G-CSF afllankelijke activatie vanSTATl verloopt echter niet via het rnembraan-distale cytoplasrnatische dornein vanG-CSF-R en is dus onafhankelijk van de tyrosine residuen van de receptor.In Hoofdstllk 4 is de betrokkenheid van lak kinases bij de activatie van de p2 Rroute onderzocht door een G-CSF-R mutant te maken, waarin tryptofaan residu 650 inhet rnembraan-proximale dornein van G-CSF-R vervangen is door een arginine132


133/136

residu. Deze substitutie blijkt zowel G-CSF-R gemedieerde tyrosine-fosforylering vanJak2 als proliferatie te verhinderen. Tevens is deze mutant niet in staat om de vonningvan p I45/Shc/GRB2 p90/GRB2 en SHP 2/GRB2 complexen te induceren. Dezeresultaten tonen aan dat het membraan-proximale cytoplasmatische domein vanG-CSF-R niet aileen essentieel is voor proliferatie inductie en activatie van Jakkinases maar oak voar activatie van signaal eiwitten van de p21 rrs route \Vat via hetmembraan-distale domein van G-CSF-R verloopt.

In Hoofdslllk 5 is de rol van de cytoplasmatische tyrosine residuen van G-CSF-Rin de transductie van proliferatie en differentiatie signalen onderzocht door tyrosinenaar-phenylalanine substitutie mutanten van G-CSF-R te expresseren in myeloide32D cellen. AflOnderlijke mutatie van de 4 tyrosines blijkt geen invloed te hebben opde differentiatie-inducerende eigenschappen van G-CSF-R. Daarentegen heeft desubstitutie van een specifieke tyrosine, nl Y764, een dramatisch effect op lOwe I deinductie van proliferatie als de timing van differentiatie. Mutatie van Y704, Y729 ofY744 bei nvloedt de proliferatie signalering niet. G-CSF stimulatie van de mutant vanY764 induceert geen celcyclus progressie van de G I naar de S fase, wat resulteert inversnelde differentiatie en een significant verlaagde productie van neutrof

the role of granulocyte colony stimulating factor receptor signalling in neytrophil dev

Documents