(Received for publication, May 25, 1996, and in revised form, February 10, 1997)
From the Graduate Program in Biochemistry and Molecular Biology, Center for Gene Regulation and the Department of Veterinary Science, The Pennsylvania State University, University Park, Pennsylvania 16802
Studies of proliferative signaling via type 1 cytokine receptors have revealed a three-step activation mechanism.
Cytokine-induced receptor dimerization mediates the
trans-phosphorylation of Jak kinases, Jaks phosphorylate receptors at
tyrosine sites, and SH2 domain-encoding effectors then are
recruited to these sites. Signaling factors that associate with
activated erythropoietin (Epo) receptor complexes include
phospholipase C-, phosphatidylinositol 3-kinase, SHIP, Shc, Grb2,
Cbl, Crk-l, HCP, Syp, and STAT5. While at least certain of these
factors modulate proliferative signaling, mutated Epo receptor forms
lacking Tyr(P) sites retain substantial mitogenic activity. Presently
we show that a highly truncated Epo receptor form that retains box-1,
yet lacks the conserved box-2 domain (and all Tyr(P) sites) nonetheless
effectively promotes mitogenesis, survival, and Myc and Pim-1
expression. In addition, mitogenesis and Myc expression are shown to be
supported by a direct Epo receptor-Jak2 kinase domain chimera. Thus,
Epo-dependent mitogenesis and inhibition of apoptosis each
depend critically upon only the Epo receptor box-1 domain, with no
essential role exerted in these response pathways by the box-2
domain.
Primary regulation over hematopoietic cell growth is exerted
through cytokine activation of receptors of the type 1 superfamily. Included in this family are the receptors for
Epo,1 interleukins-2, -3, -4, -5, -7, and
-9, ciliary neurotrophic factor, oncostatin M, prolactin, growth
hormone, G-CSF, granulocyte-macrophage-colony stimulating factor, and
thrombopoietin (1, 2). Within extracellular domains of these receptors,
co-distributed cysteine residues, predicted paired seven-fold helical domains, and an intervening WSXWS motif comprise the
most highly conserved structural features. In addition, within
cytoplasmic domains a membrane-proximal eight-amino acid box-1 motif
typically is represented, and in at least certain receptors (Epo, IL-2,
IL-3, G-CSF, growth hormone, prolactin receptors, for example) an
adjacent 10-amino acid box-2 motif is discernible (3-10).
Functionally, type 1 receptors possess no known enzymatic activities.
However, delimited receptor domains containing these box motifs have
been shown to mediate the recruitment of one or more of the Janus-type
protein-tyrosine kinases Jak-1, -2, and -3 and Tyk2 (11, 12). In this
receptor family, the catalytic activation of Janus kinases is thought
to be driven by trans-phosphorylation as mediated by ligand-induced
receptor oligomerization. The activation of primary signaling factors
then proceeds from the phosphorylation of receptor tyrosine residues,
to the recruitment of an increasingly complex set of SH2
domain-encoding effectors. In our laboratory, work has focused
primarily on signaling in the Epo receptor system. This type 1 cytokine
receptor comprises an attractive minimal model in that unlike many
related receptors the Epo receptor apparently exists as a single
transmembrane subunit (13), is activated via simple homodimerization
(14), and activates a single member of the Jak kinase family, Jak2 (15,
16). Moreover, disruption of Epo receptor gene expression in chimeric
mice recently has provided evidence that the physiological role for the
Epo receptor apparently is also strictly defined, and is restricted in
its essential activities to supporting the development of late
erythroid progenitor cells (17). By comparison, cellular response
pathways that are activated by Epo are diverse, and important roles for Epo have been demonstrated in the mitogenesis, survival, and terminal differentiation of erythroid progenitor cells (18). To date, the
majority of studies of Epo receptor signaling have focused on
mitogenesis. This is largely due to the practical consideration that it
is relatively straightforward to reconstitute this signaling pathway in
stably transfected IL-3-dependent murine lymphoid and myeloid cell lines. Studies in these systems have revealed that eight
cytoplasmic tyrosine motifs in the murine Epo receptor function to
recruit and/or facilitate the tyrosine phosphorylation of the following
set of SH2 domain-encoding effectors: p95vav (19),
phospholipase C-
(20), p85 (p110), phosphatidylinositol 3-kinase
(21), SHIP (22), Grb2, Shc (23), Cbl (24), Crk-l (24), Cis (25), HCP
(26), Syp (27), and STAT5 (and possibly STATs 1 and 3) (28, 29). In
addition, activation of a mSos1, Ras, Raf-1 pathway to MAP kinases has
been described (30), and Epo stimulation of the transcription of
myc, oncostatin-M, cis, and fos genes
has been demonstrated (31, 32). Finally, Epo-induced serine
phosphorylation of Bcl2 (33) and the induced expression of the
cytosolic serine-threonine kinase Pim-1 (19) have been demonstrated.
Despite this above progress in identifying Epo-activated transducing factors, for the majority of these factors the extent to which their activation is important for Epo-stimulated proliferative signaling remains a central, unresolved issue. This point is underlined by studies from several laboratories, which recently have demonstrated mitogenic signaling by truncated and/or (phospho)tyrosine-deficient Epo receptor forms (2, 21, 28, 29). Work from our laboratory, for example, has assisted in mapping essential membrane-proximal cytoplasmic domains (34), has defined the box-1 domain as a specific subdomain for Jak2 activation (2), and has evidenced a requirement for Jak2 kinase activation in Epo-stimulated proliferation (16, 34, 35). Presently, to more directly assess roles in this process for Jak2 kinase activation, the abilities of two novel minimal Epo receptor forms to mediate Epo-dependent mitogenesis and inhibition of apoptosis have been studied. These include a highly truncated form that retains box-1, but lacks the conserved box-2 domain and all distal residues (ER-Bx1), and a form in which the extracellular and trans-membrane domains of the Epo receptor are fused directly to the PTK-like and PTK domains of Jak2 (ER-J2KK). Notably, ER-Bx1 is shown to efficiently support Epo-induced mitogenesis, inhibition of apoptosis, and the expression of Myc and Pim-1. The chimera ER-J2KK likewise was active in supporting Epo-stimulated mitogenesis and myc gene expression, and detectably mediated the induced expression of Pim-1 protein. Findings are discussed in the context of Epo receptor minimal subdomains (and associated effectors), which are essential for the activation of mitogenic and survival response pathways.
IL-3-dependent murine myeloid FDCP1 cells have been characterized previously as a model for studies of Epo receptor function (2, 9, 16, 34, 35). The subline used (FDC-WEHI2, or FDC2) has low potential for transformation to factor-independent growth. FDC2 and derived lines were maintained at 37 °C, 5% CO2 in 8% FBS, 10 µM 2-mercaptoethanol, Opti-MEM medium supplemented with 3% conditioned medium from WEHI-3B cells as a source of IL-3 (36). FDC2 lines ectopically expressing select Epo receptor forms (see below) were prepared by stable co-electrotransfection with pXM expression vector constructs (50 mg) and pCINeo (2.5 mg) as described previously (34). Polyclonal lines were established by selection in G418 (1 mg/ml) and Epo (50 units/ml).
Epo Receptor ConstructsEpo receptor forms studied include: the murine wt Epo receptor (13), ER-396, a carboxyl-terminal deletion mutant truncated at Met396 (34) (i.e. residue 372 of the mature Epo receptor), ER-Bx1, a form truncated to delete the box-2 subdomain (and all distal cytoplasmic residues), and ER-J2KK, a chimeric construct in which the extracellular and transmembrane domains of the Epo receptor (Met1-Lys280) were fused directly to the PTK and kinase-like domains (Glu775-Val1160) of murine Jak2 kinase (15). Expression of wild-type and ER-396 constructs (prepared by Exo III-mediated mutagenesis) has been described previously (34). ER-Bx1 and ER-J2KK receptor forms were constructed as follows. For ER-Bx1, an XbaI site was introduced by PCR (CACCTAGAGCATCTAGAG; His328-Leu329-Glu330) immediately preceding the first leucine of the box-2 domain. A stop codon then was introduced at this position using an XbaI amber stop cassette (Pharmacia Biotech Inc.). For ER-J2KK, a 2000-base pair fragment of a murine Jak2 cDNA (16) was ligated to a natural BglII site within a murine Epo receptor cDNA. Each cDNA construct then was cloned into a modified pXM expression vector. Levels of Epo receptor construct expression in stably transfected FDC2 lines were assayed using 125I-Epo (324 Ci/mmol, Amersham) in equilibrium binding assays as described previously (37).
Assays of Cytokine-induced ProliferationCytokine-induced mitogenesis in FDC2 and derived cell lines was assayed based on rates of stimulated reduction of the tetrazolium compound MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl-2H-tetrazolium)) to formazan (Promega) (2), or on rates of stimulated incorporation of [3H]thymidine (34). Cells (3 × 105 cells/ml, 50 ml/well, 96-well plate) were exposed to cytokines (50 ml) for 48 h (37 °C, 5% CO2). MTS and phenazine methosulfate were added, and at 2 h of incubation absorbance (A490) was measured (microplate reader, model 550, Bio-Rad). In [3H]dT incorporation assays, cells were plated (2 × 104 cells/well) exposed to cytokines (48 h) and incubated with [3H]thymidine (1 mCi) for 2 h. Scintillation counting of harvested cells was performed using a 1205 Betaplate counter (Pharmacia).
Jak2 Activation AssaysFDC2-derived cell lines (at 8 × 105 cells/ml) were washed in Opti-MEM and incubated for 12 h at 37 °C, 5% CO2 in medium containing 1% fetal calf serum. Na3VO4 (500 mM) then was added for a 20-min period, followed by Epo (± 50 units/ml, 8 min). Cells were chilled to 0 °C, washed once with Opti-MEM, and lysed in 0.2 ml of lysis buffer (1.1% Nonidet P-40, 150 mM NaCl, 50 mM Tris, 1 mM Na2EDTA, 0.02% NaDOC, 1 mM Na3VO4, 2 mM NaF, pH 7.5, 0.5 mg/ml leupeptin, 0.7 mg/ml pepstatin, 50 mg/ml phenylmethylsulfonyl fluoride). Cleared lysates were incubated with Jak2 antiserum (1.5 h, 4 °C) (Upstate Biotechnology, Inc. (UBI), Lake Placid, NY) and with protein A-Sepharose CL4B (90 min, 4 °C). Immune complexes were washed twice with wash buffer (lysis buffer containing Nonidet P-40 at 0.55%) eluted in sample buffer (3.4% SDS, 0.2 M dithiothreitol, 0.05 mM bromphenol blue, 10% glycerol, and 0.12 M Tris-HCl, pH 6.8 (100 °C, 5 min)), electrophoresed (7.5% acrylamide, 0.2% bisacrylamide SDS gels), and transferred to Nitro-Plus membranes (MSI, Westboro, MA). Membranes were blocked (1% powdered milk, 3% BSA, 0.1% Tween 20, in 10 µM Tris base, 150 mM NaCl, 0.05% Tween 20) and probed with phosphotyrosine antibody 4G10 (UBI). Complexes were detected by ECL (Amersham) (38). For reprobing, membranes were stripped in 100 mM 2-mercaptoethanol, 2% SDS, 62.5 mM Tris-HCl, pH 6.7 (50 °C, 20 min).
Western Blotting of Bcl2, Pim-1, and Epo Receptor Forms in FDC2-derived Cell LinesIn analyses of Bcl2 and Pim-1 expression, cell lines were allowed to accumulate in G1 phase by incubating exponentially growing cells (8 × 105 cells/ml) in Opti-MEM, 1% FBS, 10 µM 2-mercaptoethanol in the absence of IL-3 or Epo for 9 h. Cells then were exposed to Epo, and at 0, 3, and 9 h of exposure, lysates were prepared for Western blotting. Briefly, cells were collected, washed once in Opti-MEM, and lysed in RIPA buffer (1% Nonidet P-40, 0.5% NaDOC, 0.1% SDS, 0.5 mg/ml leupeptin, 0.7 mg/ml pepstatin, 50 mg/ml phenylmethylsulfonyl fluoride, in 145 mM NaCl, 12 mM Na2HPO4, 2.5 mM NaH2PO4, pH 7.2). Electrophoresis was in 12% acrylamide, 0.3% methylene bisacrylamide SDS gels. Antisera to Bcl2 (Santa Cruz Biotechnology) and Pim-1 (see "Acknowledgments") were used at 0.1 and 2.5 mg/ml, respectively. Antibody-antigen complexes were detected using a horseradish peroxidase-conjugated secondary antibody and ECL (38). Western blotting of Epo receptor constructs was performed as described previously (34).
Apoptosis-associated Nucleosomal Fragmentation of DNAIn assays of apoptosis-associated DNA fragmentation, exponentially growing FDC-WEHI2-derived cell lines (6-9 × 105 cell/ml) were washed twice in Opti-MEM and resuspended in pre-equilibrated Opti-MEM, 1% fetal calf serum, 10 µM 2-mercaptoethanol at 8 × 105 cells/ml. Cytokines (IL-3, Epo, SCF) then were added as indicated. At 7 h, DNA was isolated using a Trevigel (Gaithersburg, MD) extraction system. Inter-nucleosomally fragmented DNA was then labeled using [32P]GATP (800 Ci/mmol) and DNA polymerase I (Klenow fragment) (5-min reaction, 23 °C). Reactions were terminated by the addition of Na2EDTA (5 mM) and heating (15 min at 70 °C). Products were electrophoresed on a 1.4% TrevigelTM 500 agarose gel in TAE (4 mM Tris, 2 mM NaC2H3O2, 1 mM Na2EDTA, pH 7.8) and were transferred to Nytran membranes. Fragmented DNA was analyzed by autoradiography, and quantitatively by phosphorimaging.
Northern BlottingRNA was isolated from FDC2 and derived
cell lines by the method of Chomczynski and Sacchi (39). Total RNA (20 µg/sample) was electrophoresed in denaturing gels (3% formaldehyde)
and blotted to nylon membranes. Probes used in hybridizations were
labeled by random priming using the Klenow fragment of DNA polymerase I
and [32P]-dATP (3000 Ci/mmol). These included
full-length cDNAs for murine Pim-1 (40), c-Myc, and
glyceraldehyde-3-phosphate dehydrogenase.
In initial
studies, two novel Epo receptor constructs were developed toward
advancing an understanding of possible direct contributions of Jak2
kinase to Epo-dependent mitogenesis and inhibition of
apoptosis. These included ER-Bx1, a form in which the conserved box-2
subdomain (and all distal carboxyl residues) of the Epo receptor were
deleted, and ER-J2KK, a form in which the extracellular and
transmembrane domains of the Epo receptor were fused directly to the
kinase-like and PTK subdomains of murine Jak2 kinase. These receptor
forms are diagramed in Fig. 1, together with the
previously described truncation mutant ER-396 (31), the wt Epo
receptor, and wt Jak2 kinase.
In proliferative signaling studies, the above Epo receptor constructs
were transfected into myeloid FDC2 cells, and lines expressing the
above receptor forms were established by selection in G418, and Epo.
Receptor expression was confirmed by Western blotting, and in FDCER-Bx1
and FDCER-J2KK cells receptor forms of expected molecular mass were
detected (Mr 34,000 and 72,000, respectively)
(Fig. 2). For FDCER-Bx1, the occurrence of two
Mr forms is thought to be attributable to
differential glycosylation as has been observed previously for several
alternate, active Epo receptor constructs (13, 37). For the wt Epo
receptor and ER-396 receptor forms, expression has been characterized
previously (34). In addition, levels of expression of each receptor
form were assayed in 125I-Epo equilibrium binding assays.
As shown in Table I, receptor densities were highly
comparable among cell lines (approximately 400-600 receptors/cell) and
approximated normal levels of Epo receptor expression in erythroid
progenitor cells (18). In parental FDC-WEHI2 cells, no expression of
endogenous Epo receptors was detected by Western blotting, Northern
blotting (34), or reverse transcription-PCR (2).
|
Experiments next were performed to assess the mitogenic activities of
the above minimal receptor forms. In previous studies, our laboratory
(2, 16, 34, 35) and others (10, 34, 41, 42) have developed Epo receptor
forms with mutations in, or proximal to, the box-2 cytoplasmic
subdomain that are either inactive or substantially compromised in
mitogenic signaling. Based on these studies, it has been concluded that
this receptor subdomain is critical to mitogenic function. However, and
as shown in Fig. 3, Epo-induced proliferation proved to
be supported efficiently by the box-2-deficient Epo receptor construct,
ER-Bx1. Specifically, concentrations of Epo that were required to
support proliferation of FDCER-Bx1 cells at 50% of maximal rates (1-2
units/ml) were essentially equivalent to FDCER-396 cells. In accordance
with previous studies (34), mitogenic signaling via ER-396 was
approximately 50% as efficient as via the wt Epo receptor in FDC-wtER
cells. For FDCER-Bx1 and FDCER-396 cells, maximal levels of
proliferation also were comparable yet again were somewhat lower than
levels supported by the wt Epo receptor (approximately 65% of maximal versus FDCwtER cells; average of three independent
experiments). Thus, these primary analyses reveal that the box-2
subdomain of the Epo receptor is nonessential for mitogenic function in
FDC cells, and suggest that inactivity of at least certain previously studied Epo receptor forms with mutations in (and/or immediately flanking) this domain may be the trivial consequence of misfolding.
Based on the observed efficient mitogenic signaling via the above
ER-Bx1 construct (and on the defined role of the box-1 subdomain in Jak
kinase recruitment/activation among related type1 receptors) (5, 8,
42), the ability of a direct Epo receptor/Jak2 kinase fusion construct,
ER-J2KK, to function mitogenically in FDC2 cells next was tested. In
preliminary MTS-based assays, activity reproducibly was indicated for
this construct. To better quantitate this response, in FDCER-J2KK
cells, rates of Epo-stimulated [3H]thymidine
incorporation were assayed and were compared directly to FDCER-Bx1
cells as a control (Fig. 4). Notably, this Epo
receptor-Jak2 chimera proved to support Epo-induced
[3H]dT incorporation at significant levels, although,
maximal mitogenic rates were attenuated (40% versus ERBx1,
and 25% versus wtER, see Fig. 3). Whether this restricted
activity is attributable to folding events for this ER-J2KK chimera, or
to the loss of an ability to activate at least certain downstream
mitogenic signaling factors is addressed below. It is clear, however,
that this chimera effectively supports Epo-induced proliferation, and
evidence therefore is provided that activation of Jak2 (in the absence
of Epo receptor cytoplasmic domains) promotes mitogenesis in a
ligand-regulated fashion.
Jak2 Kinase Activation via the Minimal Epo Receptor Form ER-Bx1, and Apparent Constitutive Tyrosine Phosphorylation of ER-J2KK
To
initially address mechanisms by which activity of the above Epo
receptor forms is regulated by Epo, tyrosine phosphorylation of the
ER-Bx1 and ER-J2KK receptor forms was studied in FDC cells. As shown in
Fig. 5 (panel A), ER-Bx1 mediated Epo-induced
Jak2 activation at levels comparable to the wt-Epo receptor. In
contrast, for the chimeric construct ER-J2KK, no effect of Epo exposure on tyrosine phosphorylation was detectable in repeated independent experiments performed in parallel with FDCwtER cells. Instead, tyrosine
phosphorylation of ER-J2KK appeared to be constitutive (Fig. 5,
panel B). This finding suggests that Epo-induced
dimerization of Jak2 (in addition to its tyrosine phosphorylation) is
probably critical to its activity in proliferative signaling. These
experiments, however, were compromised to an extent by the limited
reactivity of Jak2 antibodies with ER-J2KK, and by the lack of useful
antibodies to residual Epo receptor domains (for use in
immunoprecipitation). Therefore, to further confirm the identity of the
indicated Tyr(P)-ERJ2KK protein in FDCER-J2KK cells, this chimeric
construct also was expressed in COS cells (Fig. 5, panel B).
ER-J2KK from COS cells likewise was constitutively
tyrosine-phosphorylated, and this result is consistent with the
observation by Duhe et al. (43) that catalytic activity of
Jak2 increased upon the removal of amino-terminal homology domains.
Activity of ER-Bx1 and ER-J2KK Receptor Forms in the Epo-dependent Inhibition of Apoptosis
Effects of Epo
on red cell production have been well evidenced to include primary
effects on progenitor cell survival (44). However, little is known
regarding Epo-activated upstream effectors of this response pathway.
Thus, it was of interest to test whether apoptosis might be inhibited
by the above-described minimal Epo receptor forms in FDCER-Bx1, and
FDCER-J2KK cells. As controls, FDCwtER cells also were analyzed, and
inhibition of apoptosis by IL-3 and SCF was compared in each cell line.
Here, apoptosis was assayed based on internucleosomal fragmentation of
DNA (35). As shown in Fig. 6, Epo efficiently protected
FDCER-Bx1 cells against apoptosis. As compared with levels of apoptotic
inhibition supported by the wt Epo receptor (FDCwtER cells), however,
protection in FDCER-Bx1 cells was slightly diminished. By comparison,
IL-3-inhibition of DNA fragmentation was highly efficient, while
inhibition by SCF reproducibly was partial (therefore providing two
useful internal controls). Quantitative assays next were performed to
more precisely assess the ability of the ER-Bx1 receptor form (as
compared with the wt Epo receptor) to inhibit DNA fragmentation (Fig.
7). Here, Epo concentrations were varied and
fragmentation was assayed quantitatively via phosphorimaging. These
analyses confirmed the high activity of the ER-Bx1 receptor form in
mediating Epo-dependent survival. In contrast, the ability
of the Epo receptor-Jak2 chimera ER-J2KK to mediate Epo-inhibited
apoptosis was marginal, at best. This was observed in three independent
experiments and at least suggests that differences may exist in the
ability of this latter receptor form to promote mitogenesis
versus survival.
Activity of ER-Bx1 and ER-J2KK Receptor Forms in Mediating the Epo-induced Expression of Bcl2, Myc, and Pim-1
Possible
mechanisms underlying the ability of ER-Bx1 ER-J2KK Epo receptor forms
to support Epo-stimulated proliferation next were addressed. In at
least certain related type 1 receptor systems, cytokine-induced
expression of Bcl2 has been implicated in proliferative signaling (45).
In FDCwtER, FDCER-Bx1, or FDCER-J2KK cells, however, little if any
effect on levels of Bcl2 was observed over a 9-h time course of Epo
exposure (Fig. 8). By comparison, Pim-1 expression
recently has been shown to be induced by Epo receptor forms containing
both box-1 and -2 subdomains, and Tyr343 (19).
Interestingly, analyses of the above samples from Epo-exposed FDCwtER,
FDCER-Bx1, and FDCER-J2KK cells revealed the Epo-induced expression of
Pim-1 in each cell line (Fig. 9). Levels of Pim-1 expression as stimulated via ER-J2KK, however, were limited. In these
studies, two forms of Pim-1 reproducibly were detected
(Mr 34,000 and 35,000), and this is consistent
with studies from other laboratories (46). Therefore, to further assess
these effects (and associated signaling mechanisms), transcriptional
activation of pim-1 and myc genes via ER-Bx1 and
ER-J2KK receptor forms also were studied (Fig. 10). In
FDCER-Bx1 cells, Epo rapidly and efficiently activated the
transcription of both c-myc and pim-1 genes. As compared directly to FDC2-wtER cells, rates and levels of
transcriptional activation via ER-Bx1 receptors were detectably but not
markedly attenuated. In FDCER-J2KK cells, c-myc
transcription also was activated upon Epo exposure. pim-1
gene transcription, however, was not supported at significant levels
via this chimeric Epo receptor form. pim-1 and
c-myc gene transcription were stimulated strongly by IL-3 in
FDCER-J2KK cells, and pathways to pim-1 transcriptional activation in this subline therefore are not perturbed. These studies
serve to confirm the activity of the ER-Bx1 construct in mediating
Pim-1 expression, demonstrate c-myc activation by ER-Bx1 as
well as ER-J2KK receptor forms, and at least suggest that effectors of
pim-1 transcription might depend upon domains that are
represented in ER-Bx1, but lacking in ER-J2KK.
Since the original identification of type 1 cytokine receptors as
single-transmembrane, non-enzymatic proteins (1), this family of
receptors has been revealed to function in concert with an increasingly
diverse set of co-assembled primary signaling factors. In the Epo
system, ligand-induced receptor dimerization has been shown to promote
the binding, (trans-)phosphorylation, and catalytic activation of Jak2
(12). Jak2-mediated phosphorylation of Epo receptor tyrosine motifs
then leads to the SH2 domain-mediated recruitment of the
phospholipid-modifying enzymes phospholipase C-,
phosphatidylinositol 3-kinase, and SHIP; the tyrosine phosphatases HCP
and Syp; the molecular adaptors Grb2, Shc, Cbl, Crk-l, and Cis; and the
latent transcription factor STAT5. Accordingly, significant attention
has focused on defining mechanisms of factor recruitment, interaction,
and function in this and related receptor systems. However, in an
increasing number of type 1 cytokine systems the ability of
membrane-proximal (phospho)tyrosine-deficient receptor domains to
efficiently signal growth has been demonstrated (5, 7, 8, 42). These
findings limit the likelihood that distal cytoplasmic receptor domains
and associated factors contribute more than modulating roles, and
reinforce the apparent essential nature of Jak kinases in
cytokine-stimulated proliferation. The present studies of Epo receptor
function were performed to further define minimal Epo receptor domains
(and effectors), which contribute critically to
Epo-dependent mitogenesis and inhibition of apoptosis.
A highly truncated Epo receptor construct ER-Bx1 first was developed that retains the conserved membrane-proximal box-1 domain, but lacks box-2 and all distal (phospho)tyrosine sites for effector recruitment. Notably, this receptor form efficiently mediated both mitogenesis, and Epo-dependent inhibition of apoptosis. These findings underline a necessary and sufficient role for the Epo receptor box-1 domain (and the associated activation of Jak2 kinase) in these proliferative signaling pathways. Mechanistically, this ER-Bx1 receptor form also was shown to mediate the Epo-induced expression of Myc and of the serine-threonine kinase Pim-1. To further test contributions of Jak2 kinase activity in Epo-stimulated growth, a chimeric construct also was developed in which the Epo receptor extracellular and transmembrane domains were fused directly to the kinase and kinase-like domains of Jak2. Interestingly, this chimera likewise effectively mediated mitogenesis and Epo-induced myc transcription. In contrast, Pim-1 expression was only nominally supported, and this chimera did not function detectably in inhibiting apoptosis.
Signaling properties of the Epo receptor forms ER-Bx1 and ER-J2KK are
of interest to consider based on comparisons to previously studied,
related Epo receptor mutants. As mentioned above, several Epo receptor
forms that are mutated within the conserved box-2 domain or immediate
carboxyl flanking regions have been described previously, and these
typically are either highly compromised or inactive in mitogenic
signaling (see Table II). This receptor domain therefore
has been concluded to be important, if not essential, to this activity.
The high proliferative activity of the presently studied ER-Bx1 mutant,
however, demonstrates that this is not the case and suggests that
mutations in the box-2 domain and adjacent residues simply may
compromise functions of the essential box-1 domain. These conclusions
are supported by recent studies by Hilton et al. (47) in
which essential mitogenic activity was demonstrated for highly
truncated Epo receptor forms expressed in BaF/3 cells. Unfortunately,
however, BaF/3 cells recently have been revealed to express endogenous
Epo receptors at levels sufficient to confound interpretations of the
activities of ectopically expressed Epo receptor forms (48, 49).
Presently, the apparent lack of an essential role for the Epo receptor
box-2 domain in proliferative signaling raises the obvious question as
to the role of this conserved domain among type 1 receptors. One
possibility that previously has been suggested is a supporting role in
Jak2 binding and activation (12). For the Epo receptor, however,
substitution of the box-1 domain into a corresponding box-1 region in
the IL-2- receptor converts the IL-2
/
/
receptor from a
Jak1,3 to a functional Jak2,3 signaling complex (2). Thus, the
selective activation of Jak2 apparently depends only upon the box-1
domain. In addition, highly truncated forms of the receptors for G-CSF
and growth hormone likewise have been recently shown to efficiently
support mitogenic signaling in the absence of the box-2 domain (5-8,
50). Notably, both Epo and G-CSF are known to synergize in mitogenic
signaling with the receptor for SCF, c-Kit (51). It therefore is
interesting to speculate that this synergy might depend at least in
part on features in the box-2 region. In recent work by Wu et
al. (52), in fact, the direct binding of c-Kit to a box-2
associated domain of the Epo receptor has been demonstrated. Thus, it
should be of interest to test whether this possible role for the Epo
receptor box-2 domain (and flanking residues) is evident in functional analyses.
|
To assess whether ligand-mediated activation of Jak2 per se
might support proliferative signaling, the kinase (and kinase-like) domain of Jak2 presently was fused directly to the extracellular and
trans-membrane domains of the Epo receptor, and signaling activities of
this ER-J2KK chimera in FDC2 cells were studied. Interestingly,
ligand-dependent mitogenic signaling was supported by
ER-J2KK, albeit at reduced rates (40% of levels supported by the
ER-Bx1 construct). In contrast, this construct did not exert detectable
activity in the Epo-dependent inhibition of apoptosis. With
regard to signaling mechanisms, ER-J2KK was shown to efficiently mediate Epo-induced Myc expression, but activated Pim-1 protein expression at only low levels did not appreciably stimulate
pim-1 transcription. In keeping with findings in related
systems (53), this latter result at least suggests that Pim-1
expression may be regulated in the Epo receptor system at both
transcriptional and post-transcriptional levels. These findings also
are consistent with the prospect that Epo-induction of pim-1
gene expression might depend upon as yet unidentified effectors that
are activated by the ER-Bx1 construct, but not by ER-J2KK. In a broader
context, factors that act to regulate Pim-1 expression generally are
not well understood, and FDC2-ER-Bx1 cells should provide a tractable system to investigate this Epo signaling pathway. Finally, while the
present studies were in progress, properties of two distinct receptor-Jak2 chimeric constructs were reported. First a CD16-Jak2 fusion construct was developed, and the tyrosine phosphorylation of
this construct was shown to be ligand regulated (i.e.
stimulated by antibody-induced cross-linking) (54). However, this
CD16-Jak2 chimera apparently was inactive mitogenically. Second, Jak2
PTK and PTK-like domains were fused to the trans-membrane and
extracellular domains of the EGF receptor (55). Interestingly, this EGF
receptor-Jak2 chimera efficiently supported EGF-dependent
mitogenesis as well as inhibition of apoptosis in 32D cells. Moreover,
and despite the absence of canonical sites for the recruitment of
either Grb2-mSos complexes or STAT5, MAP kinase, and STAT5 activation
was supported in this system. One interpretation of these results is
that novel and as yet undefined pathways exist through which Jak2 can
stimulate Ras, MAP kinase kinase, MAP, and STAT5 activation. However,
the possibility also should be considered that abnormally increased Jak2 activation levels might result in the tyrosine phosphorylation of
uncoupled effectors, which then promote growth and survival. In
addition, this stimulation would proceed in the absence of inhibitory
effects, which normally are exerted on Jak2 by the SH2 tyrosine
phosphatase, HCP (26). Increased stimulation of Jak2 might result from
the overexpression of chimeric constructs, or overstimulation with
ligand. In studies by Nakamura et al. (55), levels of EGF
receptor-Jak2 chimera expression in 32D cells were not addressed, and
ligand concentrations 1 µg/ml were used. In contrast, in the
present studies of ER-J2KK signaling, FDC2 cells expressing normal
levels of receptors were used, and ligand concentrations were limited
(2-3 orders of magnitude lower). This likely explains why tyrosine
phosphorylation and activation of SH2 domain-encoding effectors
(including STAT5) are not detectable in FDC-ERJ2KK or FDC-ERBx1
cells,2 as would be predicted based on the
lack of (phospho)tyrosine recruitment sites.
A final property of the wt Epo receptor and of the minimal construct
ER-Bx1 that merits discussion is their ability to efficiently mediate
Epo inhibition of apoptosis. Among type 1 cytokine receptors, the Epo
receptor and IL-2 -receptor structurally are relatively closely
related, and their homology in part led to the discovery of this
receptor superfamily. In the context of apoptosis, IL-2 has been shown
to activate Bcl-2 expression (45), and this observation promoted the
present analyses of Epo effects on Bcl-2 expression. However, Bcl-2
levels were not affected by Epo in FDC2 cells, and in recent analyses
of Bcl-xL no significant effects of Epo on this Bcl-2
homologue were apparent.3 For Bcl-2, this
result is consistent with studies by May et al. (33), in
which Epo likewise did not affect Bcl-2 expression (yet interestingly
did stimulate Bcl-2 serine phosphorylation). By comparison, in HCD57
cells Epo recently has been shown to stimulate increases in the
expression of Bcl-xL (56). Thus, it seems likely that
Bcl-2-related factors may act to regulate Epo effects or survival, and
this prospect presently is being investigated further in our FDC-ER and
FDC-ER-Bx1 models.
Last, and as mentioned above, roles may exist for Pim-1 in Epo-dependent inhibition of programmed cell death. Support for this notion derives from three considerations. First, in the context of erythropoiesis, Pim-1 expression has been associated with erythroleukemias (57) and has been correlated with modulated erythrocyte volume in transgenic models (58). Second, and as studied in lymphocytes from lpr mice, the forced expression of Pim-1 leads to hyperproliferation and interestingly this appears to depend upon effects on cell survival rather than mitogenesis (59). Third, Pim-1 expression commonly is activated by hematopoietic cytokines which promote cell survival, yet little presently is known about its regulated expression or the targets of Pim-1 phosphorylation. In the present study, Epo-induced expression of Pim-1 clearly is shown to depend upon the conserved box-1 domain of the Epo receptor. Future studies that address mechanisms of Pim-1 expression and function in this minimal model should be of significant interest.
We thank Amgen for the generous provision of recombinant Epo, Dr. Steven Pelech (University of British Columbia, Vancouver) for antibodies to Pim-1, Ning Jiang (The Pennsylvania State University) for assistance with Jak2 kinase assays, Ron Gmerek for contributions in apoptosis studies, and Loretta Muchinsky for valued typing efforts.