Role of the Distal Half of the c-Mpl Intracellular Domain in Control of Platelet Production by Thrombopoietin In Vivo

Abstract

The cytokine thrombopoietin (TPO) controls the formation of megakaryocytes and platelets from hematopoietic stem cells. TPO exerts its effect through activation of the c-Mpl receptor and of multiple downstream signal transduction pathways. While the membrane-proximal half of the cytoplasmic domain appears to be required for the activation of signaling molecules that drive proliferation, the distal half and activation of the mitogen-activated protein kinase pathway have been implicated in mediating megakaryocyte maturation in vitro. To investigate the contribution of these two regions of c-Mpl and the signaling pathways they direct in mediating the function of TPO in vivo, we used a knock-in (KI) approach to delete the carboxy-terminal 60 amino acids of the c-Mpl receptor intracellular domain. Mice lacking the C-terminal 60 amino acids of c-Mpl (Δ60 mice) have normal platelet and megakaryocyte counts compared to wild-type mice. Furthermore, platelets in the KI mice are functionally normal, indicating that activation of signaling pathways connected to the C-terminal half of the receptor is not required for megakaryocyte differentiation or platelet production. However, Δ60 mice have an impaired response to exogenous TPO stimulation and display slower recovery from myelosuppressive treatment, suggesting that combinatorial signaling by both ends of the receptor intracellular domain is necessary for an appropriate acute response to TPO.

Hematopoiesis is a complex process in which functionally and morphologically very distinct blood cells originate from a common precursor, the hematopoietic stem cell. The whole-blood system of a vertebrate can be reconstituted in its entire diversity by a very small number of hematopoietic stem cells, illustrating that this process involves both massive proliferation and differentiation. It is established that these processes are, at least in part, controlled by hematopoietic cytokines that bind to receptors expressed on blood progenitor cells. Whether signals of cytokine receptors instruct the progenitor cell to commit to a specific lineage or simply provide a survival signal to an already committed progenitor cell is a matter of intensive research and debate. Furthermore, cytokine-induced receptor homo- or hetero-dimerization leads to the activation of a plethora of distinct downstream signaling pathways. Although knowledge concerning the biochemical mechanisms by which these pathways are activated is increasing, their role in mediating the action of specific cytokines is still relatively unclear.

Thrombopoietin (TPO) is the primary physiological regulator of platelet production. In vitro and in vivo experiments with recombinant TPO (rTPO) indicate that it stimulates both megakaryocyte progenitor proliferation as assayed by colony formation and megakaryocyte maturation (3, 9, 20, 39). TPO supports the formation of CFU-MK, both alone and in combination with early acting factors (4, 21) and stimulates the production of megakaryocytes and functional platelets from enriched murine or human stem cell populations (7, 41). Injection of rTPO into mice increases platelet counts 4- to 6-fold and causes up to a 20-fold increase in the number of bone marrow megakaryocytes (21, 26). Even though rTPO dramatically stimulates platelet production, it has only modest effects on platelet function. In vitro studies show that rTPO does not directly induce platelet aggregation but does enhance aggregation induced by other agonists (28, 30). Thus, TPO appears to sensitize platelets, making them more responsive to aggregation agonists.

Mice deficient in TPO have platelet and megakaryocyte counts reduced by approximately 90% compared to normal mice (8). This decrease in platelet number is accompanied by a reduction in megakaryocyte progenitors and megakaryocyte ploidy. Although these results point to TPO as the physiological regulator of platelet production, they also indicate that TPO is not required for the production of normal platelets and megakaryocytes, since these mice exhibit a low level of morphologically and functionally normal platelets (5). While the effects of TPO were originally thought to be lineage specific, TPO-deficient mice also have decreased progenitor numbers of both myeloid and erythroid lineages (1, 6). They also have a decreased number of hematopoietic stem cells, indicating that TPO has a more pleiotropic range of activities (35).

The action of TPO is mediated entirely through c-Mpl, a member of the cytokine receptor superfamily originally identified as the cellular homologue of a retroviral oncogene (36, 38). c-Mpl expression appears to be limited to tissues that support hematopoiesis, namely, bone marrow, spleen, and fetal liver (27), and is high in CD34⁺ cells and cells of the megakaryocyte lineage. Binding of TPO to c-Mpl is believed to induce receptor homodimerization and subsequently activation and tyrosine phosphorylation of JAK2. JAK2 activation leads to phosphorylation of c-Mpl on tyrosines followed by the recruitment and activation of signaling molecules to these phosphorylated docking sites. Downstream molecular targets of receptor activation include signal transducers and activators of transcription 3 and 5 (STAT3 and STAT5), CBL, Shc, Vav, Raf-1, mitogen-activated protein kinase (MAPK), phosphatidylinositol (PI) 3-kinase, and SHIP (11, 15, 29, 31–34). Activation of these various pathways by thrombopoietin has been extensively studied over the last few years. In vitro experiments with factor-dependent cell lines suggested that signaling pathways activated by the membrane-proximal half of the c-Mpl cytoplasmic domain were required for mediation of the proliferation function of TPO (15), while the distal half of the receptor, in particular its activation of the MAPK pathway, were involved in mediating megakaryocyte differentiation (17, 32). However, these cell lines only poorly recapitulate the complex process of megakaryocyte differentiation and platelet production. To address the role of these different signaling pathways in a more physiological context, we used a gene-targeting approach to generate mice with a c-mpl gene encoding a protein missing the C-terminal half of the intracellular domain. The phenotype of these mice indicates that the distal region of the receptor is not required for megakaryocyte differentiation but is necessary to potentiate the information delivered by the membrane proximal half of the receptor in response to a rapid increase in thrombopoietin levels.

MATERIALS AND METHODS

Targeting constructs.

Genomic clones of murine c-Mpl were isolated from a 129 genomic DNA library, and an EcoRI-KpnI fragment containing exons 7 and 8 was subcloned into pBS.SK(−). A PCR primer annealing 5′ upstream of the KpnI site present in intron 8 (AAT AGT ATC CCT GCT CGC AAA) was used in combination with a primer located at the end of exon 10 (TAG CAG CAG TAG GCC CAG) to PCR amplify a genomic DNA fragment containing exons 9 and 10. This fragment was then fused by PCR to a cDNA fragment containing the region encoding the entire c-Mpl intracellular domain (WT) or only the first 60 amino acids (Δ60) and subcloned into the KpnI site to generate the long arm. A simian virus 40 polyadenylation site and a PGK1-neo cassette flanked by lox sites were then added sequentially to the 3′ end of this long arm. A short arm consisting of a 1.4-kb fragment containing the 3′ untranslated region (UTR) of c-Mpl was obtained by PCR with a PmeI site engineered in at its 3′ end and inserted 3′ to the PGK1-neo gene to generate the final targeting construct (see Fig. 1).

FIG. 1.

Targeting of the c-mpl gene. (A) Structure of the c-mpl gene and knock-in construct. Exons are numbered and represented by boxes. Exons 1 to 9 (light grey boxes) encode the extracellular domain. Exon 10 encodes the transmembrane domain and is indicated by a black box. Exon 11 and the beginning of exon 12 (dark grey) encode the intracellular domain. Sites relevant to the construction of the targeting vector and screening for homologous recombination are indicated. (B) Targeting of the c-mpl gene in ES cells. After homologous recombination, the mutated allele was detected by PCR across the short arm (data not shown) and confirmed by Southern blotting with a probe located outside the targeting construct, 3′ of the short arm following digestion of the genomic DNA with EcoRI. Following homologous recombination, the size of the WT allele is reduced from 12 to 4.5 kb. (C) PCR analysis of genomic DNA isolated from mouse tail snips of the offspring of interbreeding heterozygous mice. The DNA was amplified with a primer set specific for the neo gene (mutant) and a primer set which amplifies a fragment of intron 9 which is deleted in the targeted allele and therefore specific for the WT allele (wt). (D) Western blot analysis of the c-Mpl receptor expressed on platelets obtained from WT and Δ60 animals. The blot was probed with a hamster monoclonal antibody directed against the extracellular domain of the mouse receptor. The reduction in size observed for the receptor present on the platelets from Δ60 animals demonstrates the successful targeting of the c-mpl gene.

ES cell work.

The embryonic stem (ES) GS cells, derived from the 129Sv mouse strain, were electroporated with 20 μg of PmeI-linearized targeting vector. At 24 h later, the cells were placed for 9 to 11 days in selection medium containing 400 μg of G418 per ml. Single colonies were picked, and pools of 12 clones were subjected to PCR analysis for homologous recombination with a primer annealing to the PGK1-neo cassette (ATG CGG TGG GCT CTA TGG CTT CT) and a primer specific for the 3′ UTR chromosomal sequence external to the 3′ homology region of the targeting vector (TGG GTC TGG GGT GGC AAA CA). A 1.6-kb PCR fragment is generated from clones having undergone homologous recombination. Positive clones were further confirmed for homologous recombination by Southern blotting of EcoRI-digested ES cell genomic DNA. The blots were probed with a ³²P-labeled overlapping oligonucleotide probe recognizing a region located outside of the targeting vector, in the 3′ UTR of c-mpl. Two independent ES clones were injected into blastocysts of C57BL/6J mice. Chimeric males were identified by agouti coat color and were mated with C57BL/6J females. Genomic tail DNA derived from the offspring was analyzed for the c-Mpl-KI allele by PCR. The progeny of F₁ × F₁ intercrosses were used in all experiments for both KI and WT controls. All the results presented were generated by using both clones, with similar results.

Preparation of washed murine platelets.

Whole blood was collected into 3.8% sodium citrate (9:1, vol/vol) in microcentrifuge tubes and centrifuged for 3 s at 12,000 × g to make platelet-rich plasma (PRP). The PRP was collected, and the residual whole blood was recombined and recentrifuged to collect a second batch of PRP, which was added to the initial collection. Platelets were counted on a Baker System 9000 Diff Model cell counter (Serono-Baker, Allentown, Pa.). The PRP was collected and centrifuged at 1,500 × g for 5 min. The platelets were resuspended in 1 ml of Tyrode's buffer (0.14 M NaCl, 2.7 mM KCl, 12 mM NaHCO₃, 0.4 mM Na₂HPO₄, [pH 7.4]) supplemented with 5.5 mM glucose, 10 mM HEPES, and 2% bovine serum albumin (BSA), i.e., Tyrode's BSA) containing 300 ng of prostaglandin I₂ per ml (PGI₂). After centrifugation, the platelets were resuspended in 1 ml of Tyrode's BSA containing 300 ng of PGI₂ per ml, centrifuged, and finally resuspended in Tyrode's BSA. To allow recovery from the PGI₂ treatment, the washed platelets were allowed to stand in the dark for a minimum of 2 h at room temperature (RT) before use. The platelets were then stimulated with rTPO for 10 to 15 min, lysed, analyzed by immunoprecipitation and Western blotting.

Western blot analysis.

To detect tyrosine phosphorylation of signaling molecules, washed murine platelets were stimulated with TPO (10 ng/ml) for 15 min or with buffer alone then washed twice with cold phosphate-buffered saline (PBS). Platelets were then incubated in lysis buffer (10 mM Tris [pH 7.5], 50 mM NaCl, 5mM EDTA, 30 mM sodium pyrophosphate, 5 mM sodium fluoride, 100 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 1% Nonidet P-40) on ice for 1 h. The insoluble fraction was removed by centrifugation. Extracts were precleared by incubation with protein A-agarose for 1 h at 4°C and then incubated overnight at 4°C with 10 μl of agarose-conjugated antiphosphotyrosine monoclonal antibody 4G10 (UBI, Inc., Waltham, Mass.) per 10⁸ cells. Western blot analysis of tyrosine phosphorylation was performed essentially as recommended by the manufacturer. Antibodies against JAK-2, STAT3, and Akt were from Santa Cruz Biotechnology (Santa Cruz, Calif.), antibodies against STAT5 and Shc were from UBI; and antibody against phospo-ERK was from Promega (Madison, Wis.).

¹²⁵I-fibrinogen binding to washed platelets.

Platelets (2.5 × 10⁷ per ml) were incubated in the presence of ¹²⁵I-fibrinogen (20 nM), CaCl₂ (2 mM), and TPO (2 μg/ml) for 10 min at RT. Various concentrations of ADP were then added, and the binding was allowed to proceed for 60 min at RT. The binding solution was then layered on a 500-μl step of Tyrode's BSA–20% sucrose in a microcentrifuge tube and centrifuged at 12,000 × g for 4 min. The liquid was discarded, and the platelet pellet was counted.

Colony assays.

Age-matched mice (C57BL/6, WT, Δ60 and c-mpl KO) were sacrificed, and femurs were harvested. Bone marrow was flushed with PBS–2% fetal calf serum, and a single-cell suspension was made. Nucleated cells were counted on an inverted microscope by using a hemactoytometer. Methylcellulose-based colony assays were performed in 35-mm plates at a density of 5 × 10⁴ cells per plate in complete myeloid methylcellulose medium (Stem Cell Technologies Inc., Vancouver, British Columbia, Canada). The plates were incubated at 37°C and 5% CO₂ for 14 days. Colonies were then counted and phenotyped on an inverted light microscope. CFU-Mk assays were performed by using the Megacult-C system (Stem Cell Technologies), essentially as described by the manufacturer. Briefly, 2.2 × 10⁶ cells/mL were seeded in double-chambered slides (10⁵ cells/slide) in medium plus collagen. Interleukin-3 (IL-3) (10 ng/ml), IL-6 (20 ng/ml), IL-11 (50 ng/ml) (R&D Systems, Minneapolis, Minn.), and rmTPO (50 ng/ml) (Genentech, South San Francisco, Calif.) were then added to the cells and medium. Slides were placed at 37°C for 7 days, fixed with acetone, and visualized for megakaryocytes by acetylcholinesterase staining (18).

Flow cytometric analysis of megakaryocyte frequency and ploidy.

Megakaryocyte frequency and ploidy were determined by two-color flow cytometry as detailed previously (2). Briefly, megakaryocytes in bone marrow cell suspensions prepared from one femur and tibia of each mouse or derived from short-term culture as described above were labeled with a platelet-specific rat anti-mouse monoclonal antibody (4A5) (kindly provided by Samuel Burstein, University of Oklahoma, Oklahoma City, Okla.) followed by fluorecein isothiocyanate-conjugated goat anti-rat immunoglobulin G F(ab′)₂, and DNA labeled with propidium iodide as previously described (2). The DNA content of 4A5-positive cells was analyzed by two-color flow cytometry (19) on a FACScan flow cytometer (Becton-Dickinson, San Jose, Calif.). The percentage of cells in each ploidy class was obtained by integrating the number of cells under each ploidy peak. Megakaryocyte frequency was calculated from the flow cytometric analyses as the percentage of total cells recognized by the 4A5 antibody.

RESULTS

Generation of c-Mpl Δ60 mice.

To analyze the contribution of signaling pathways activated by different regions of the c-Mpl intracellular domain to megakaryocyte and platelet production in vivo, we have generated mice with a c-mpl gene encoding a protein lacking the last 60 amino acids (one-half of the intracellular domain). We used a knock-in approach (16) to replace exons 11 and 12 encoding the intracellular region of c-mpl with a cDNA fragment encoding only the first half of the cytoplasmic signaling domain. The targeting vector (Fig. 1A) was electroporated into ES cells. Gene targeting was detected by PCR in 2 of 200 colonies (Fig. 1B), and these clones were injected into blastocysts. Both clones gave germ line transmission, and founders were bred to generate homozygous gene-targeted mice (Fig. 1C). These mice were viable and healthy and displayed no overt abnormalities.

To verify that the targeting construct had properly modified the c-mpl gene, PRP was obtained from Δ60 mice and their control littermates, lysed, and subjected to Western analysis with a hamster monoclonal antibody directed against the extracellular domain of mouse c-Mpl. A strong signal of approximately 90 kDa was detected as expected in PRP from WT mice, while a band of slightly reduced size was present in PRP from Δ60 mice (Fig. 1D). The size difference is consistent with a 60-amino-acid deletion in a total of 629 residues. As observed in Fig. 1D, the truncated version of c-Mpl is expressed at the same level as the WT molecule, indicating that modification of the protein did not perturb its expression levels.

TPO signaling in platelets from Δ60 mice.

Analysis of tyrosine-phosphorylated proteins in response to TPO stimulation of platelets was performed to determine which signaling pathways were disrupted in c-Mpl Δ60 mice compared to wild type (Fig. 2). Tyrosine-phosphorylated proteins were immunoprecipitated from PRP stimulated with 10 ng of TPO per ml for 15 min then analyzed by Western blotting with antibodies directed against specific signaling molecules. JAK2, STAT3, and STAT5 phosphorylation was still observed in these platelets but at different levels from those in the platelets from WT mice. The levels of JAK2 phosphorylation induced upon stimulation with rhTPO was identical in platelets from WT and Δ60 mice, confirming that the truncated receptor is correctly expressed at the cell surface. Phosphorylation of both STAT molecules was reduced in Δ60 mice, as previously observed in tissue culture. However, STAT5 phosphorylation was decreased more than STAT3 phosphorylation, in contrast to what is observed in cell lines, where STAT3 phosphorylation appears to be more strongly affected by a similar deletion of the receptor (10). This suggests that the cellular context may influence the signaling pathway recruited to a particular docking site, as already observed in other systems. Akt phosphorylation upon TPO stimulation was also abrogated in platelets from Δ60 mice, indicating that activation of this pathway, probably via PI 3-kinase, is mediated by the distal half of the cytoplasmic domain of c-Mpl. As predicted from experiments in cell lines (15), Shc phosphorylation was abolished in platelets isolated from Δ60 mice and ERK phosphorylation was strongly reduced. These results indicate that we have successfully modified c-Mpl in these mice into a receptor that can still activate the JAK-STAT pathway but not the Shc/ras/MAPK and the anti-apoptotic Akt pathways.

FIG. 2.

Activation of downstream signaling by tyrosine phosphorylation in platelets from WT or Δ60 mice stimulated with rTPO. Washed platelets from WT or Δ60 mice were treated for 10 min with rTPO or buffer only, lysed, and immunoprecipitated with an antiphosphotyrosine antibody coupled to Sepharose beads. Immunoprecipitated proteins were analyzed by SDS-PAGE and probed with antibodies against JAK2, STAT3, STAT5, Akt, and Shc. For Erk, the whole platelet lysate was run and analyzed with an anti-phospho-ERK antibody.

Blood cell counts and progenitor analysis.

To investigate the effect of the Δ60 deletion on the production of platelets, complete blood cell counts were performed. While the complete absence of c-Mpl receptor leads to a ∼90% reduction in platelet count (1, 14), mice expressing the Δ60 form of the receptor have normal platelet counts (Fig. 3A) as well as normal levels of all other circulating blood cells (data not shown). TPO levels assayed by a HU-3 cell proliferation assay, in which we were recently able to detect a small increase in circulating TPO level in NF-E2-deficient mice (23), indicate that Δ60 mice do not have increased levels of circulating TPO compared to WT mice (data not shown). This result is in accordance with the proposed mechanism of regulation of TPO levels by platelet mass (12, 22). Histopathologic analysis of Δ60 mice did not reveal any decrease in the number of megakaryocytes in the spleen or bone marrow (Fig. 3C and D). Similar results were obtained when megakaryocyte frequency was analyzed by fluorescence-activated cell sorting with an anti-CD41 antibody (see Fig. 5A). Furthermore, the ploidy of megakaryocytes present in the bone marrow of Δ60 mice was identical to that of megakaryocytes from WT littermates (Fig. 4A). Together, these data suggest that signaling by the last 60 amino acids of c-Mpl is not required for the production of a normal platelet count. Interestingly, although platelet and megakaryocyte counts are normal in Δ60 mice, megakaryocyte progenitor levels were reduced by approximately 50% whereas progenitor cell counts from other lineages were not decreased compared to controls (Fig. 3B).

FIG. 3.

Platelet and megakaryocyte analysis in Δ60 mice. (A) Platelet counts in WT and Δ60 mice. Blood was collected by retroorbital bleed venous puncture and analyzed in a hematology analyzer (System 9000 Diff; Serono-Baker Diagnostics, Allentown, Pa.) to determine the platelet count (five mice per group). (B) Comparative analysis of bone marrow megakaryocyte progenitors from WT and Δ60 mice. (C and D) Megakaryocytes were counted in 5-μm bone marrow (C) and spleen (D) sections stained with hematoxylin and eosin. All results shown are the mean and one standard error of the mean.

FIG. 5.

Megakaryocyte and platelet counts in WT and Δ60 mice after rTPO injection or myelosupressive regimen. (A) Megakaryocyte frequency in the marrow of WT and Δ60 mice was evaluated 2 and 3 days after intraperitoneal injection with PBS alone or 5 μg of rTPO per mouse. The cells were stained with the 4A5 antibody and analyzed by FACS. The average of day 3 and day 4 data is presented for PBS-injected mice. (B) Time course analysis of the platelet counts in WT and Δ60 mice following intraperitoneal injection with PBS alone or 5 μg of rTPO per mouse. Blood was collected by retroorbital bleed venous puncture and analyzed in a Serono-Baker Diagnostic System 9000 Diff model hematology analyzer (five mice per group). (C) Platelet numbers in mice following myelosuppression. All mice received 500 rads of whole-body irradiation from a cesium irradiator plus 1.2 mg of carboplatin intraperitoneally. Platelet counts were determined at various time points as indicated. P < 0.05 on day 10, P < 0.003 on day 14, and P < 0.011 on day 17 between WT and Δ60 mice. (D) Embryonic day 20 platelet counts from WT and Δ60 mice. Statistical analysis was performed by analysis of variance followed by Fisher's protected least-significant-difference test. All results shown are the mean and one standard error of the mean for six mice in each group.

FIG. 4.

Ploidy analysis. (A) Ploidy analysis of bone marrow megakaryocytes from WT and Δ60 mice. Bone marrow cells were collected from the femur and tibias of five mice in each group and stained with the 4A5 antibody (a generous gift of Sam Burstein) and propidium iodide. (B) Ploidy analysis of megakaryocytes from liquid cultures. Bone marrow from WT and Δ60 mice was grown in cultures for 5 days in the presence of rTPO and then analyzed for DNA content.

Response to TPO.

To compare the capacity of bone marrow cells from Δ60 mice and WT animals to respond to exogenous TPO, we first used a liquid culture assay. Bone marrow cells harvested from femurs and tibias of Δ60 mice and their WT littermates were grown for 5 days in the presence of rTPO and then analyzed for megakaryocyte markers by using a rat anti-mouse platelet (4A5) or anti-CD41 antibody and for megakaryocyte ploidy. Fewer megakaryocytes were generated in cultures from Δ60 mice compared to those from controls (data not shown), and the megakaryocytes generated were of lower ploidy (Fig. 4B).

To extend these observations in vivo, Δ60 mice or littermate controls were given a single intravenous injection of rTPO, and their bone marrow was harvested 72 or 96 h later, when megakaryocyte ploidy and frequency are maximal (2). At this time, normal animals exhibited a dramatic increase in both megakaryocyte frequency and ploidy as expected (Table 1; Fig. 5). In contrast, Δ60 mice showed no increase in the frequency of bone marrow megakaryocytes at 72 h and only half the increase in megakaryocyte frequency at 96 h. They also had a less pronounced increase in modal ploidy, which was one ploidy class less than that of WT mice. To evaluate the impact of this reduction on platelet production, we used the same single-dose injection of rTPO in Δ60 mice or WT controls and measured the time-dependent increase in platelet counts. As seen in Fig. 5B, the increase in platelet counts was significantly smaller in Δ60 mice than in WT mice. On day 5, the average platelet count had risen only 116% in Δ60 mice compared to day 0, while it had risen 206% in WT mice (P = 0.001). Δ60 mice were also more sensitive to myelosuppression. Use of a combination regimen of sublethal irradiation plus 1.2 mg of carboplatin intraperitoneally showed that Δ60 mice had a more pronounced nadir and had delayed recovery from thrombocytopenia, suggesting that their megakaryocytes were more sensitive to apoptosis and that they were not responding as well as WT mice to increased levels of endogenous TPO levels (Fig. 5C). Finally, we analyzed platelet counts in these mice on embryonic day 20, a time of elevated platelet synthesis, and found a significant reduction in the platelet counts of Δ60 mice compared to WT mice, suggesting that it takes longer for Δ60 mice to reach their normal platelet count.

TABLE 1.

Megakaryocyte ploidy in TPO-treated mice^a

Mouse	Treatment	% of megakaryocytes in:
		8N	16N	32N	64N	128N
WT	PBS	8.6	43.8	13.1	1.2	0.6
	TPO	5.5	16.8	32.9	15.0	1.7
Δ60	PBS	12.8	51.0	8.7	1.4	0.7
	TPO	15.1	20.5	18.7	5.5	1.5
KO	PBS	17.8	31.6	3.6	1.4	0.8
	TPO	18.3	31.3	5.9	1.6	0.7

^aPloidy analysis of bone marrow megakaryocytes from WT and Δ60 mice treated with a single dose of rTPO (5 μg/mouse) or PBS only is shown. Bone marrow cells were collected from the femurs and tibias of five mice for each group and stained with the 4A5 antibody and propidium iodide. 

^bThe percentage of megakaryocytes in each ploidy class is indicated. 

Functional evaluation of Δ60 platelets.

To find whether platelets from Δ60 mice were functionally different from platelets from normal mice, we first measured their ability to upregulate α_IIb β₃ (GP-IIb-IIIa) in response to platelet agonist stimulation. Platelets become activated when exposed to extracellular matrix or soluble agonists such as thrombin or ADP. During activation, the platelet receptor α_IIb β₃ undergoes conformational changes, enabling it to bind fibrinogen and von Willebrand factor, leading to platelet aggregation and formation of a hemostatic plus in vivo, an important primary step in the arrest of bleeding. The binding of fibrinogen by α_IIb β₃ at the platelet surface is therefore a direct measure of platelet activation. In this assay, the level of activation of platelets from Δ60 mice by ADP was not significantly different from that of platelets from WT mice (Fig. 6A). Addition of exogenous TPO to platelets potentiates the fibrinogen binding induced by ADP or other known agonists, decreasing the 50% effective concentration by about twofold (28, 30). Interestingly, this effect of TPO has been shown to be mediated through PI 3-kinase (40). Consistent with the involvement of PI3 kinase, the potentiating effect of TPO on platelets from Δ60 mice was less pronounced than that for the WT controls (Fig. 6A). However, this effect was not completely abolished, suggesting that other pathways contribute as well. This difference in responsiveness does not seem to affect the capacity of platelets to form a hemostatic plug in Δ60 mice, since the bleeding times of these mice evaluated by a tail cut technique were virtually identical to those of WT controls (Fig. 6B). Together, these data indicate that platelets from Δ60 mice are functionally normal.

FIG. 6.

Functional evaluation of platelets from WT and Δ60 mice. (A) Stimulation of fibrinogen binding to platelets from WT or Δ60 mice by ADP in the presence or absence of exogenous TPO. (B) Bleeding time measured by tail cut in WT and Δ60 mice.

DISCUSSION

Receptor activation by hematopoietic cytokines is known to activate multiple signaling pathways. Over the last decade, the specific molecular interactions leading to activation of downstream targets have been dissected. However, the precise role of these various pathways in translating the biological activity of these cytokines remains to be determined. Multiple studies of cell lines have provided evidence that independent signaling pathways were mediating the proliferation and the differentiation signals generated by these receptors. For example, the membrane-proximal region of the cytoplasmic domain of both the granulocyte colony-stimulating factor and the TPO receptor appears to deliver a proliferation signal while the distal region contains the information necessary for differentiation signaling (10, 13, 15, 17, 29, 32). However, although both TPO (or c-Mpl) and granulocyte colony-stimulating factor knockout mice display a dramatic decrease in platelet and neutrophil counts, respectively, the basal levels of these cells are still present in the circulation (1, 8, 14, 24, 25). The remaining cells appear to be properly differentiated, suggesting that these cytokines control the number of their target cells but are not required for their differentiation. Consistently, recent mouse studies indicating that the intracellular domains of these two receptors were interchangeable suggested that cytokine receptors do not provide an instructive signal but merely a nonspecific survival and/or proliferation signal (37).

The data presented herein question the differentiation role established in vitro for the carboxy-terminal region of c-Mpl. Since cell lines do not accurately recapitulate the process of megakaryocytopoiesis and thrombopoiesis in vitro, we have used a knock-in approach to generate mice carrying a c-mpl gene encoding a protein lacking the last 60 amino acids (one-half of the intracellular domain) in order to define its function. This deletion removes two regions that have been proposed to mediate a differentiation effect of TPO in cell lines: Tyr 112 is necessary for activation of the MAPK pathway through Shc (17), and a domain located between residues 71 and 94 is involved in the prolonged activation of MAPK (32). Consistent with in vitro data, Shc is no longer substantially activated in Δ60 mice upon TPO stimulation and the extent of ERK phosphorylation is strongly diminished. Interestingly, the compromised signaling does not result in an overt phenotype in unstimulated Δ60 mice. The numbers of platelets, as well as other blood cells, were within the normal range. Similarly, the frequency and ploidy distribution of bone marrow megakaryocytes from Δ60 mice were identical to those of megakaryocytes from WT mice. The platelets of Δ60 mice appear to function normally, since they were capable of upregulating fibrinogen binding sites upon agonist stimulation and of forming a hemostatic plug to arrest bleeding. Together, these data indicate that the first 61 amino acids of the c-Mpl receptor intracellular domain are sufficient for the generation of a normal platelet count, providing the signals leading to the proliferation of megakaryocyte progenitors and to megakaryocyte endomitosis. These data also strengthen the idea that the c-Mpl receptor does not deliver a differentiation signal or that this signal is not required in vivo.

Interestingly, although the membrane-proximal region of the receptor intracellular domain is capable of providing all the signals necessary for generating normal numbers of megakaryocytes with normal ploidy under steady state conditions, it is not as efficient as the full-length receptor in response to a rapid increase in endogenous or exogenous TPO levels. Treatment of Δ60 mice with a single dose of recombinant TPO led to a smaller increase in platelet number than in WT mice. The reduced effect of TPO on the platelet count of Δ60 mice appears to result from a smaller increase both in megakaryocyte ploidy and in megakaryocyte frequency compared to WT mice. Similarly, Δ60 mice exhibited a more pronounced nadir and slower recovery from thrombocytopenia induced by a combination of sublethal irradiation and carboplatin injection, indicating an impaired response to an increase in endogenous TPO levels. Finally, Δ60 mice have low platelet levels compared to WT mice on embryonic day 20, indicating a delay in reaching their definitive platelet level. Together, these results suggest that loss of signaling from the carboxy-terminal region results in a muted response to TPO. Signals emerging from the distal portion therefore determine only the quantitative but not the qualitative response to TPO on megakaryocyte numbers and ploidy.

Since several signaling pathways are affected by the deletion of the last 60 amino acids, it is not possible to conclude from these experiments the exact contribution of each one of them to the TPO response. The slower response to TPO stimulation could be due to decreased protection against apoptosis caused by the inability of mutant c-Mpl to activate Akt through PI3 kinase. Therefore, it may take longer for the proliferation signal emanating from the first 61 amino acids to lead to the production of an adequate number of megakaryocytes. Alternatively, activation of the MAPK pathway through Shc may be required for a maximal ploidy increase, since small molecule inhibitors of MAPK inhibit endomitosis in vitro (P. Rojnuckarin, J. G. Drachman, and K. Kaushansky, Blood 92[Suppl. 1], abstr. 2777). The reduced MAPK activation through the Δ60 mutation could therefore result in a decreased rate of endomitosis, decreased proliferation of megakaryocyte progenitors, or increased apoptosis. Finally, as yet undiscovered pathways activated by the distal region of c-Mpl may be implicated.

In summary, these data demonstrate that platelet production by TPO does not require a differentiation signal emanating from the distal half of the cytoplasmic domain of its receptor. The membrane-proximal half of the c-Mpl cytoplasmic domain is sufficient to activate all the pathways necessary to establish a normal steady-state platelet count. However, a combination of signals from the proximal and distal regions of the receptor intracellular domain is necessary for an appropriate acute response to TPO.