Cytokine Production and Function in c-mpl–Deficient Mice: No Physiologic Role for Interleukin-3 in Residual Megakaryocyte and Platelet Production

NZB mice bearing mutant IL-3Rα genes (Il3raⁿ),23 c-mpl-deficient mice,20 and mice lacking a functional IL-3 gene24 have been described previously. Mice genetically defective for both c-Mpl and IL-3Rα were generated by breedingmpl^−/− and NZB mice to yield F1 offspring (mpl^+/-Il3ra^+/n), which were subsequently interbred. The peripheral blood platelet counts of 98 male F2 mice were measured at 6 weeks of age and 27 thrombocytopenic (mpl^−/−) mice were identified and analyzed. The IL-3Rα genotype of these mice was then determined using the polymerase chain reaction method previously described.23 As mpl^−/− mice were of mixed C57Bl/6 and 129/Sv genetic background, and IL-3^−/− mice were a mixture of C57Bl/6, 129/Sv, and Balb/c, in the mpl^−/−IL-3^−/− intercross experiments, wild-type controls included a combination of data from mice of each of these strains. All mice were housed in a conventional animal facility and analyzed at between 2 and 4 months of age.

Murine GM-CSF, IL-3, IL-5, and TPO were produced in purified recombinant form and kindly provided by Dr N. Nicola (The Walter and Eliza Hall Institute of Medical Research, Melbourne). Recombinant murine IL-6 was a kind gift of Dr R. Simpson (Joint Protein Structure Laboratory, The Walter and Eliza Hall Institute of Medical Research and The Ludwig Institute for Cancer Research, Melbourne). Recombinant human G-CSF and recombinant murine LIF were kindly provided by Amgen (Thousand Oaks, CA) and AMRAD (Boronia, Australia), respectively, and recombinant human IL-11 was purchased from R&D Systems (Minneapolis, MN).

Organs from male mpl^−/− mice or their wild-type littermates were collected, fragmented with scissors, and incubated in 2 mL of serum-free Dulbecco's modified Eagle's medium (DMEM) in a fully humidified atmosphere of 10% CO₂ in air. Supernatants conditioned by each tissue were collected after 4 days, filter sterilized, and stored at 4°C for analysis. Cytokine concentrations were determined by bioassay as described25,26 using parental Ba/F3 cells for IL-3 (detection limit, 20 pg/mL) or Ba/F3 cells transfected with the specific receptors for IL-6 (detection limit, 100 pg/mL), IL-11 (detection limit, 100 pg/mL), LIF (detection limit, 200 pg/mL), G-CSF (detection limit, 400 pg/mL), or TPO (detection limit, 100 pg/mL). As no organ-conditioned medium contained IL-3, FDC-P1 cells, which respond to IL-3 or GM-CSF, provided a specific bioassay for GM-CSF (detection limit, 100 pg/mL). Microwell assays were performed in 60-well microtiter dishes (Lux, Nashville, TN) as described.25 Briefly, 200 cells in 10 μL DMEM containing 10% newborn calf serum (NCS) were added to duplicate 5 μL volumes of serially diluted organ conditioned medium. The numbers of viable cells in each well were determined by manual inspection under phase-contrast microscopy after incubation for 2 days at 37°C in a fully humidified atmosphere of 10% CO₂ in air. Where cell counts exceeded 200 per well, the culture was scored as containing >200 cells. Each assay included a titration of a known concentration of the relevant purified recombinant cytokine, which was used as a standard for calculating the amount of that cytokine present in each conditioned medium. As previously described,25 the media conditioned by four organs, brain, salivary gland, kidney, and liver, contain components toxic to the assay cells. Because mixing experiments indicated that the toxic effects were eliminated at dilutions greater than 1:8, the limits of detection for cytokines in these conditioned media were eightfold higher than stated. Similarly, as serum samples were initially diluted fourfold, the limit of detection for cytokines in serum was fourfold higher than stated.

Groups of weight-matched male mpl^−/−mice or wild-type littermates were injected twice daily subcutaneously with 0.2 mL of cytokine solution or dilution medium (saline supplemented with 5% NCS). Mice injected with IL-6 (2 μg/day) and LIF (1 μg/day) were analyzed after 7 and 8 days, respectively.

Peripheral blood was collected by retro-orbital bleeding and diluted into 3% acetic acid containing methylene blue (white cells) or 1% ammonium oxalate (platelets) for manual cell counts using hemocytometer chambers and standard microscopy. Megakaryocytes were enumerated by microscopic examination of hematoxylin and eosin-stained histologic sections of sternal bone marrow and spleen, which had been cut to standard thicknesses of 1 and 2 μm, respectively. The clonal culture of hematopoietic progenitor cells was performed in 1 mL cultures of 2.5 × 10⁴ (bone marrow) or 10⁵ (spleen) cells in 0.3% agar in Iscove's modified Dulbecco's medium (IMDM) supplemented with 20% fetal calf serum (FCS), 10 ng/mL murine IL-3, 100 ng/mL murine SCF, and 4 U/mL human EPO as previously described.18 Single stimulus cultures included cytokines at the following final concentrations: murine TPO, 100 ng/mL; IL-5, 10³ U/mL; IL-6, 100 ng/mL; murine GM-CSF, 10 ng/mL; human G-CSF, 10 ng/mL; murine M-CSF 10 ng/mL. Agar cultures were fixed and sequentially stained for acetylcholinesterase, Luxol fast Blue, and hematoxylin, and the composition of each colony was determined at 100-fold to 400-fold magnifications.

The concentration of cytokines in the serum or in media conditioned by the organs of adult mpl^−/− mice was examined using factor-dependent cells ectopically expressing specific cytokine receptors (see Materials and Methods). Consistent with previous reports,19 the level of serum TPO, which was undetectable in wild-type mice, was elevated inmpl^−/− animals (640 ± 0 pg/mL, n = 4). Recent studies suggest that lack of platelet-mediated clearance of TPO is the predominant mechanism accounting for the elevated TPO levels in mpl-deficient animals.27 We were unable to determine whether changes in production also contribute, as TPO production was not sufficient to be detected in conditioned medium from any organs of normal or mpl^−/− mice. To investigate whether maintenance of the residual thrombopoiesis in mice lacking TPO signalling may be due to increased production of alternative cytokines with megakaryocytopoietic activities, we also examined production of IL-3, IL-6, IL-11, LIF, and GM-CSF inmpl^−/− mice. None of these regulators was detected in the serum of wild-type or mpl-deficient animals (Table 1). Similarly, IL-3 was undetectable in all conditioned media examined from mice of both genotypes. As previously observed,25 LIF, IL-6, and GM-CSF were detected in a wide range of organ-conditioned media from normal mice, in particular that of the lungs and muscle, and no significant differences in the concentrations of these cytokines was observed in analysis ofmpl^−/− animals (Table 1). Similarly, no difference in the organ distribution or level of IL-11 production, which was evident at lower concentrations and from a more restricted range of tissues, was observed between mpl-deficient and wild-type littermates (Table 1). Together, these data suggest that of the megakaryocytopoietic cytokines under study, dramatic elevation in production does not contribute to the mechanisms by whichmpl^−/− mice maintain residual platelet levels.

To investigate whether alternative cytokines are capable of stimulating megakaryocytopoiesis in the absence of TPO signalling, the response ofmpl^−/− mice to daily administration of IL-6 or LIF was compared with that of normal animals. Presumably reflecting their thrombocytopenic state, at analysis, severalmpl^−/− mice displayed evidence of subcutaneous hemorrhage at the injection site and two LIF-injected animals that exhibited hematocrit values below 40 were excluded from analysis.

Consistent with previous reports,16 IL-6 injections for 6 days in normal mice elevated spleen weight and induced a 1.6-fold increase in platelet numbers. Similar findings were observed inmpl^−/− mice, with the platelet count increasing by approximately twofold (Table2). In mice of both genotypes, the increase in platelet counts was accompanied by significant increases in megakaryocytes and megakaryocyte progenitor cells, particularly in the spleen, but also in the bone marrow (Table 2). Little alteration in the hematocrit or white blood cell count were observed in IL-6–injected mice of either genotype (Table 2).

Similarly, LIF administration for 7 days also increased the number of circulating platelets in both normal andmpl^−/− mice (Table 3). The magnitude of the increase, 1.6-fold, was consistent with that observed in previous studies of LIF activity in normal animals.17 In contrast to IL-6–treated mice, little if any significant elevation in the number of mature megakaryocytes was observed in LIF-injectedmpl^−/− mice. No significant elevation in megakaryocyte progenitor cells was observed in either wild-type ormpl^−/− mice receiving LIF (Table 3).

To determine whether IL-3 plays an essential role in the residual megakaryocytopoiesis observed in the absence of TPO signalling, mice deficient in both c-Mpl and the IL-3Rα chain were generated. NZB mice are homozygous for a mutation in the IL-3Rα chain gene (Il3ra), and their cells are markedly hyporesponsive to IL-3 in vitro23 (see Table 4). A total of 98 F2 mice from an intercross betweenmpl^−/− and NZB mice were generated. Twenty-seven (27.5%) of these animals were thrombocytopenic and were genotyped as mpl^−/−. As expected with normal Mendelian segregation of alleles, of the 27mpl-deficient animals, seven (26%) were also homozygous for the mutant IL-3Rα chain (Il3ra^n/n). Although one of the seven mpl^−/−Il3ra^n/n mice died before analysis of an undetermined cause at 3 weeks of age, the mutant IL-3Rα gene appeared to have no significant impact on the survival ofmpl^−/− mice. Moreover, thempl^−/−Il3ra^n/nanimals were no more thrombocytopenic thanmpl^−/−Il3ra^+/+ ormpl^−/−Il3ra^+/nmice (Table 4) suggesting that IL-3 does not contribute significantly to the maintenance of residual platelet numbers inmpl^−/− mice. As expected, cells from mice homozygous for the Il3raⁿ allele displayed only a residual response to IL-3, while normal colony numbers were observed in response to GM-CSF (Table 4).

Given that some residual IL-3 responsiveness may still exist in NZB mice (Table 4), mice created through gene targeting to definitively lack IL-324 were bred with mpl-deficient animals for analysis of hematopoiesis in mpl^−/−IL-3^−/− double mutant offspring. Genotyping of 276 weanlings from matings ofmpl^+/-IL-3^+/- parents showed normal ratios of offspring of each of the expected genotypes (Table 5). In addition, no adult lethality was observed, indicating that the loss of IL-3 had no significant effect on the survival of mpl^−/− mice. Peripheral blood analysis showed that, like thempl^−/−Il3ra^n/nmice, mpl^−/−IL-3^−/− animals displayed the thrombocytopenia typical of mpl^−/− mice. However, the superimposed lack of IL-3 did not reduce the platelet count further, or was any thrombocytopenia observed in mice lacking IL-3 alone (Table 6). Similarly, the reduction in megakaryocytes evident in histologic sections ofmpl^−/− mice was not exacerbated inmpl^−/− IL-3^−/−animals and megakaryocyte numbers were normal in IL-3^−/− mice (Table 6). Both IL-3^−/− andmpl^−/− IL-3^−/−mice also displayed normal hematocrits and total white blood cell counts, as well as levels of circulating lymphocytes, monocytes, neutrophils, and eosinophils that were within the normal range (Table6). The number of peritoneal cells, the cellularity of the femoral bone marrow and spleen, and the distribution of morphologically recognizable precursor cells in these populations was not significantly different in mutant mice of all genotypes from that in normal controls (Table 6). Moreover, flow cytometric analysis of cells from bone marrow, spleen and thymus, using antibodies directed against a range of T lymphoid, B lymphoid, myeloid, and erythroid markers,20 showed no perturbations in mutant animals of any genotype (data not shown).

To determine whether IL-3 plays a physiologic role in earlier stages of hematopoiesis, progenitor cells from mutant animals were assayed in semisolid agar cultures containing a combination of SCF, IL-3, and EPO, which stimulates a broad range of erythroid and myeloid colonies.20 The total numbers of hematopoietic progenitor cells were normal in IL-3^−/− mice and the low levels evident in mpl^−/−animals20 were not further lowered inmpl^−/− IL-3^−/−mice (Table 7). The enumeration of individual colony types also indicated that loss of IL-3 did not alter the numbers of progenitor cells committed to specific hematopoietic lineages, including megakaryocyte progenitors, either in normal ormpl^−/− mice (Table 7). As expected, in cultures stimulated solely by TPO, no colony formation was evident with the cells from mpl^−/− ormpl^−/− IL-3^−/−mice, while equal numbers of small megakaryocyte colonies developed from wild-type and IL-3^−/− bone marrow cells (wild-type: 9 ± 4; IL-3^−/−: 8 ± 5 colonies per 2.5 × 10⁵ cells plated, n = 3).

Interestingly, despite development in the absence of IL-3, normal numbers of IL-3–responsive cells had developed in IL-3^−/− bone marrow (wild-type: 88 ± 28; IL-3^−/−: 71 ± 30 colonies per 2.5 × 10⁵ cells plated, n = 3) and no further reduction from that already evident in mpl^−/− mice was evident in the mpl^−/−IL-3^−/− double mutants (mpl^−/−: 29 ± 9;mpl^−/− IL-3^−/−: 33 ± 12 colonies per 2.5 × 10⁵ cells plated, n = 3). In addition, colony formation in response to stimulation by single cytokines (IL-5, IL-6, GM-CSF, G-CSF, M-CSF, or SCF) was similar between bone marrow cells from mpl^−/−and mpl^−/−IL-3^−/− mice, as well as between IL-3^−/− and wild-type animals (data not shown). Consistent with these observations of bone marrow progenitor cells, the lack of IL-3 also failed to alter the numbers of erythroid, mixed erythroid, or myeloid colony-forming cells in the spleens of these animals (data not shown). Together, these analyses suggest that even in the absence of signalling from the major regulator of megakaryocytopoiesis, TPO, IL-3 does not contribute to the residual megakaryocyte and platelet production inmpl^−/− mice. Moreover, these data confirm and extend previous observations24 to suggest that IL-3 has no essential physiologic role in the maintenance of other mature hematopoietic cells or their committed progenitor cells, in otherwise normal animals or in mpl^−/−mice.

The dominant role of TPO in the regulation of platelet production has been convincingly demonstrated by the severe thrombocytopenia evident in mice lacking this regulator or its receptor.19-21However, as TPO^−/− andmpl^−/− mice retain the capacity to produce sufficient platelets to prevent bleeding, an important contribution to thrombopoiesis in vivo is likely to be made by other cytokines with megakaryocytopoietic potential. Candidate cytokines include IL-6, IL-11, IL-3, and LIF, which significantly elevate platelet numbers when injected into animals.15-18 Although gene targeting studies have shown that in the presence of TPO, the loss of such cytokines does not significantly affect platelet levels,28-31 subtle actions of these factors may exist and be shown or amplified in mice lacking the dominant TPO signalling system. We initially investigated this possibility by examiningmpl^−/− mice for evidence of elevated production of known megakaryocytopoietic factors. Although we found, as previously described,19 that serum TPO was elevated, we did not detect biologically active IL-6, IL-11, IL-3, LIF, or GM-CSF in the circulation of mice of either genotype. A measure of production of these cytokines from individual tissue sources was obtained by analyzing organ conditioned medium. As evident in a previous survey,25 factors were produced by multiple tissues of normal adult mice, in particular the lungs, muscle, thymus, and bone shaft. However, no significant differences in the organ distribution or level of cytokine production were observed in analyses of media conditioned by the organs of mpl^−/−mice (Table 1). Thus, dramatic elevation in production of alternative megakaryocytopoietic cytokines does not appear to be a major mechanism by which mice lacking an intact TPO signalling system maintain residual platelet production. An interesting characteristic ofmpl^−/− mice is that despite producing reduced numbers of progenitors cells for essentially all hematopoietic lineages, deficits in mature blood cells are restricted to platelets.19,20 This observation complements previous studies, including those of cytokine administration32 and other mutant mouse models of hematopoiesis,33 in which alterations in the number of committed progenitor cells are not necessarily reflected by the numbers of circulating blood cells. In the case of the granulocyte lineage, for example, we also found no evidence that elevated cytokine production provides a compensatory mechanism: the concentrations and spectra of organs producing G-CSF (data not shown) and GM-CSF (Table 1) were not significantly different between normal and mpl-deficient mice.

To determine the in vivo megakaryocytopoietic potential of alternative cytokines in the absence of TPO signalling, control andmpl^−/− mice were injected with LIF or IL-6. Previous studies with these cytokines in normal mice showed a capacity to elevate platelet numbers up to twofold.16,17Our studies confirmed these observations and demonstrated that LIF and IL-6 retain this activity in mpl^−/−mice. Although the platelet numbers in LIF- or IL-6–injectedmpl^−/− mice did not recover to normal levels, reflecting the reduced numbers of megakaryocytes and their progenitors available for stimulation, the magnitude of the platelet increase was at least as significant in mpl-deficient mice as in the wild-type controls (IL-6: mpl^+/+ 1.6-fold,mpl^−/− 2.2-fold; LIF:mpl^+/+ 1.7-fold,mpl^−/− 1.6 fold, Table 2).

Our results confirm previous studies demonstrating the thrombopoietic potential of IL-6, and also of IL-11 and SCF, inmpl^−/− and TPO^−/− mice22 and further demonstrate the potency of LIF in the absence of TPO signalling. These observations suggest that, in vivo, such cytokines act substantially independently of TPO, in contrast to their reported action in certain in vitro assays.34 The elevation in platelet numbers inmpl^−/− mice injected with IL-6 was also accompanied by significant increases in the numbers of mature megakaryocytes and their progenitors in hematopoietic organs, a response consistent with that in wild-type animals (Table 2). Thus, our data also extend previous observations to show that in vivo, alternative cytokines cannot only elevate platelet numbers, but can stimulate the full process of megakaryocytopoiesis in the absence of TPO signalling.

These analyses provide important proof that alternative cytokines have the capacity to stimulate megakaryocytopoiesis in the absence of TPO signalling, but do not provide direct evidence that any of these factors control the residual steady-state megakaryocyte and platelet development in mpl^−/− mice. To further address this issue, we are conducting genetic crosses betweenmpl^−/− mice and other genetically modified mice unable to produce alternative megakaryocytopoietic cytokines or their receptors. As IL-3 is the most potent single stimulus for megakaryocyte colony formation in vitro,1 we initiated studies with mice deficient in either IL-3 or the IL-3Rα chain. NZB mice have a naturally-occurring mutation in the IL-3Rα chain gene preventing normal cell-surface expression of the receptor and resulting in a markedly reduced response to IL-3.23Interbreeding of mpl^−/− and NZB mice yielded the expected number of mpl^−/−Il3ra^n/n double mutant offspring and platelet numbers in these mice were equivalent to those in theirmpl^−/−Il3ra^+/(+/n)littermates (Table 4), suggesting little or no role for IL-3 inmpl-deficient platelet development.

As our data and that of others23 suggested that some residual IL-3 responsiveness may exist in NZB mice (Table 4), we definitively addressed the role of IL-3 in normal and residualmpl^−/− megakaryocytopoiesis using animals genetically modified to lack this cytokine.24 The numbers of circulating platelets, mature megakaryocytes, and megakaryocyte progenitor cells were normal in IL-3^−/− mice (Tables 6 and 7). Moreover, even in the absence of TPO signalling, lack of IL-3 did not exacerbate the megakaryocyte deficiency or the thrombocytopenia of c-mpl^−/− mice (Tables 6 and 7). In previous studies, initial indications that IL-3 may not play a prominent role in megakaryocytopoiesis came from the normal megakaryocyte and platelet production evident in athymic nu/numice, which lack T lymphocytes, considered a prominent source of IL-3.35 Our data from genetically modified mice specifically lacking the cytokine provide definitive evidence that, despite its potent in vitro activity on megakaryocyte proliferation1 and its ability to stimulate this lineage on administration in vivo,15 physiologically IL-3 has no significant essential role in megakaryocytopoiesis, or is it involved in the residual megakaryocyte and platelet production inmpl^−/− mice. Of broader significance, IL-3^−/− mice displayed normal numbers of hematopoietic progenitor cells in the bone marrow and spleen, as well as their mature progeny in the circulation, marrow, spleen, and peritoneal cavity24 (Tables 6 and 7). Similarly, the reduction in committed progenitor cells characteristic ofmpl^−/− mice was not exacerbated in the absence of IL-3. Thus, in addition to providing no evidence of a role for IL-3 in megakaryocyte and platelet production, our studies confirm previous analyses24 to suggest that IL-3 plays little physiologic role in the development of blood cells of other lineages.

We thank Ladina DiRago and Sandra Mifsud for excellent technical assistance and Jodie deWinter and Anne Chow for animal husbandry.