BclxL overexpression in megakaryocytes leads to impaired platelet fragmentation

Megakaryopoiesis is a complex process, which is completed by platelet formation in bone marrow (BM) sinusoids accompanied by apoptotic death of senescent cells.1 Mpl ligand, also referred to as thrombopoietin (TPO) or megakaryocyte growth and development factor (MGDF),2,3 elevates the number and ploidy of megakaryocytes and increases platelet levels in vivo. Although the terminal differentiation of mammalian megakaryocytes into platelets has great relevance for blood homeostasis, the mechanism of platelet formation is still a matter of controversy.4 5 

Apoptotic cell death is a highly regulated genetic program,6 and although it has been observed in mature megakaryocytes, it remains to be determined how the induction of apoptosis relates to normal megakaryocyte maturation and to the release of functional platelets. Recent studies attempted to investigate the relationship existing between megakaryocyte maturation, apoptosis, and proplatelet formation in culture of CD34⁺-derived megakaryocytes.7,8 Although unable to investigate in depth the subordination of events, this study showed that apoptosis occurred predominantly in mature megakaryocytes rather than in immature megakaryoblasts. Moreover, the kinetics of platelet release into the culture supernatants corresponded to the onset of apoptosis in these cells. This temporal correlation suggested that maximal platelet production and megakaryocyte apoptosis are closely related events. In a recent study, it was shown that the antiapoptotic protein BclxL is abundant in differentiating megakaryocytes in culture, but absent from mature prefragmenting cells.9 

To explore the role of deregulated expression of BclxL in megakaryocyte development and platelet fragmentation in vivo, we generated transgenic mice overexpressing the antiapoptotic human BclxL gene in megakaryocytes.10 The transgenic phenotype observed included a decrease in the number of megakaryocytes undergoing apoptosis and a change in the morphology of the demarcation membrane structure in megakaryocytes, associated with a low frequency of cells able to produce proplatelets. The platelets produced, however, displayed no change in reactivity in response to different agonists. These results support the hypothesis that cellular changes associated with BclxL overexpression in megakaryocytes are important for thrombopoiesis.

The plasmid PF4-hBclxL-SV40 was constructed by using gene fragments from several different plasmids: pPF4GH, which contains the rat 1.1-kb PF4 promoter in which the original PF4 5′ untranslated region (UTR) was included, linked to human growth hormone gene (HGH)11; PF4SV40 construct, which contains the rat platelet factor 4 (PF4) promoter and 400 bp of SV40 small t intron and polyA tail12; and pSFFV-hBclxL-NEO plasmid, which contains 800 bp of human full-length BclxL cDNA (a generous gift from Dr D. D. Chaplin, Howard Hughes Medical Institute, Washington University School of Medicine, St Louis, MO). The construction involved conventional methods, as we used before.13 The resulting plasmid, referred to as PF4-hBclxL-SV40, was confirmed by DNA sequencing, and subsequently digested with NdeI to excise the PF4-hBclxL-SV40 fragment used for microinjections.

The above construct was used for the generation of transgenic mice (on FVB strain), which were then examined for transgene integration and expression, as described before.13 

Bone marrow (BM) was harvested from femurs of control or transgenic mice as described before.11 The isolated cells were plated at a concentration of 5 × 10⁶ cells/mL in Iscove modified Dulbecco medium (IMDM; Gibco BRL, Gaithersburg, MD), supplemented with 20% horse serum (Gibco BRL), 2%l-glutamine (Gibco BRL), 50 U/mL penicillin (Gibco BRL), 50 U/mL streptomycin (Gibco, BRL), and 25 ng/mL of the Mpl ligand, pegylated recombinant murine MGDF (PEG-rmMGDF; generous gift from Amgen, Thousand Oaks, CA; referred to as MGDF) and cultured at 37°C in humidified chamber with 5% CO₂.

For proplatelet formation assay, BM was harvested and cultured as described above. For the first 3 days of culture, 25 ng/mL MGDF was added. Day 3 BM cultures were used to obtain an enriched population of megakaryocytes, using indirect immunomagnetic negative separation with Dynabeads M-450 in accordance with the manufacturer's protocol with some modifications (Dynal, Oslo, Norway). Briefly, BM cultured cells were resuspended at a concentration of 5 × 10⁶ cells/mL in 0.1% bovine serum albumin (BSA; Sigma, St Louis, MO), 0.6% sodium citrate and 2% fetal bovine serum (FBS; Gibco, BRL) in phosphate-buffered saline (PBS). Purified rat antimouse IgG: Mac-1α (Pharmingen, San Diego, CA, no. 01711A), TER-119 (Pharmingen, no. 09081), and Gr-1 (Pharmingen, no. 01211A) were added at a final concentration of 5 μg/mL and incubated for 30 minutes at 4°C on a rotor. Cells, covered with primary antibodies were washed in 0.1% BSA, 0.6% sodium citrate, and 2% FBS in PBS. Washed Dynabeads M-450 covered with sheep antirat IgG were added to a final concentration of 2 × 10⁷ beads/mL in 0.1% BSA, 2% FBS in PBS and incubated for 40 minutes at 4°C with rotation. Negatively selected megakaryocytes were pipetted off with the supernatant on a Dynal magnetic particle concentrator (MPC), washed twice in IMDM, and plated at a concentration of 1 × 10⁵cells/well in 24-well plates in IMDM, supplemented with 10% FBS, 2%l-glutamine in the presence or absence of 25 ng/mL MGDF. Proplatelets were detected and scored by phase contrast microscopy (inverted microscope Olympus IX 70, Olympus America, Bellingham, MA) on days 5 and 6 of culture. The total number of megakaryocytes making proplatelets was determined and normalized to the total number of megakaryocytes in the well.

The BM megakaryocytes were identified by in situ staining for acetylcholinesterase as described elsewhere.14,15 For immunohistochemistry, both primary BM cells and negatively selected BM megakaryocytes were used after 3-day culture in the presence of MGDF (see above). BM cells were cytospun on microscope slides at a concentration of 1 × 10⁶ or 1 × 10⁵cells/slide respectively, fixed, and stained as described before.16 Cy3-conjugated goat antirabbit IgG (Jackson Immunoresearch Laboratories, West Grove, PA) was used in conjugation with 10 μg/mL primary rabbit anti-BclxL antibody (Santa Cruz Biotechnology, Santa Cruz, CA, no. sc 7195), which specifically recognizes BclxL in rat BM cell extract. Rat anti-CD41 conjugated to fluorescein isothiocyanate (FITC; Pharmingen) or Cy3 (Jackson Laboratories) was used for staining megakaryocytes. DNA was stained with 20 μM Hoechst (Sigma).

To further confirm transgene expression, RNA isolated from total BM was subjected to reverse transcription (RT) and polymerase chain reaction (PCR)–based cDNA amplification. For this purpose BM was harvested from femurs of control or transgenic mice and cultured for 3 days as described. Total RNA was isolated with TRIzol reagent (Gibco BRL). First-strand cDNA synthesis was carried out following a protocol of SMART cDNA Library Construction Kit (Clontech, Palo Alto, CA), using specific sense: 5′-GCAGTGGTAACAACGCAGAGTACGCGGG-3′ (with 3′ annealing to CAP sequences in mRNAs), and antisense: 5′-GCAGTGGTAACAACGCAGAGTACTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTVN primers. For cDNA amplification, the Expand Long Template PCR system (Boehringer Mannheim, Indianapolis, IN), the above sense primer and 20 cycles of PCR reactions were used. Amplified DNA was electrophoresed on 1% agarose gel, transferred to nylon membrane, and tested with ³²P-labeled probes as indicated.

Platelet counts and volume and blood cell counts were all determined as described elsewhere.17 18 The level of TPO in the serum of mice was determined using the Quantikine immunoassay kit (R & D Systems, Minneapolis, MN) following the manufacturer's instructions.

Age- and sex-matched transgenic and wild-type mice (8-week-old mice each weighing about 25 g) were injected in the tail vein with 25 μg of the antibody pOp3 (dissolved in PBS/BSA), a rat IgG2a against the N-terminal 45-kDa region of mouse glycoprotein Ibα (GPIbα). This antibody is platelet/megakaryocyte specific and was used before to induce immune thrombocytopenia (ITP).19Following the injection, platelet levels were determined as a function of time. Injection of mice with PBS/BSA did not cause a significant change in platelet level.

The DNA content of megakaryocytes was analyzed by using a statistical package on a FACScan flow cytometer (Becton Dickinson, Rutherford, NJ), as described before.13 20 

Apoptosis was determined in an enriched population of BM megakaryocytes, obtained by magnetic beads separation as described above, on day 6 of culture. The ApopTag Plus Peroxidase in situ Apoptosis Detection Kit (Intergene, Purchase, NY) was used to assay apoptosis, following the manufacturer's instructions for suspension cell cultures.

Bone marrow was flushed from femurs with fixative (4.3% glutaraldehyde in 0.1 M Veronal buffer, pH 7.4), washed once in fixative, and incubated for 60 minutes at 4°C. After fixation, cells were washed once in 0.1 M Veronal buffer and postfixed in 1% osmium tetroxide in 0.1 M Veronal buffer for 10 minutes at 4°C.21 The samples were dehydrated through a series of graded ethanol, infiltrated with resin composed of Araldite 502, DDSA (dodecenyl succinic anhydride), and BDMA (benzyldimethylamine; Ted Pella, Redding, CA) and finally embedded in the above-mentioned resin. Thin sections were cut and stained with uranyl acetate and lead citrate and viewed with a transmission electron microscope (Phillips 300 TEM).

Single injections of PEG-rmMGDF (50 μg/kg) into 3-month-old mice via the lateral tail vein were carried out as described previously.22 Control mice received a carrier injection of 1% normal mouse serum in PBS.

For aggregation studies, platelets were prepared as described elsewhere.23 Briefly, blood was drawn from the abdominal aorta of 4-month-old mice into a 3.15% sodium citrate solution, keeping the ratio of 1 volume of anticoagulant to 9 volumes of blood. To obtain platelet-rich plasma (PRP), samples were centrifuged for 35 seconds at low speed (2700 rpm) in a Sorvall centrifuge (brakes off) at room temperature. To obtain platelet-poor plasma (PPP), the rest of the sample was centrifuged for 15 minutes at 1500g in a tabletop microfuge at room temperature. Platelets were stored at 37°C throughout the experiments and the cell count was adjusted in the final suspension to a concentration of 5 × 10⁵ cells/μL in PRP. Aggregation was measured at 37°C by a turbidimetric method in a dual-channel Platelet Aggregation Profiler (Biodata, Horsham, PA). A 200-μL aliquot of platelet suspension was stirred at 1100 rpm and activated by the addition of 5 to 15 μL agonist.

To explore the role of processes associated with programmed cell death in megakaryocyte development and platelet fragmentation, we created a transgenic mouse model in which the antiapoptotic humanBclxL gene was overexpressed in megakaryocytes via the rat PF4 tissue-specific promoter.11 15 We confirmed the expression of the transgene in megakaryocytes of 4 transgenic lines produced (Tg12, Tg13, Tg16, Tg19) by PCR-based cDNA amplification of BM cells (Figure 1A, see “Materials and methods”). This analysis revealed that all 4 transgenic founder lines expressed the transgene. This was concluded based on detection of SV40 (unique to the transgene) or BclxL, as compared with endogenous PF4 (a lineage-specific mRNA), or to glyceraldehyde-phosphate dehydrogenase (GAPDH). Based on this method of determination, Tg19 and Tg16 had significant levels of transgenic mRNA as compared with Tg13 and Tg12.

Expression on the protein level was verified by immunohistochemistry, using an antibody specific to BclxL. It should be pointed out that in studying megakaryocyte biology, one often has to rely on careful in situ examinations because of the rarity of this cell type in the marrow (< 0.05%). We noted that not all of the large mature megakaryocytes in the wild-type or transgenic mice stained positive for BclxL; for example, among 40 transgenic megakaryocytes surveyed in Tg19, only 70% displayed a strong staining with anti-BclxL, and only 50% of the wild-type megakaryocytes displayed a weak, but detectable, staining for endogenous BclxL. The remainders of the cells in each sample were BclxL negative. Similar results were obtained with Tg16 analyzed (data not shown), further suggesting that wild-type or PF4-driven BclxL levels are regulated on the protein level (representative fields are shown in Figure 1B-C and Figure 2B). In the transgenic or wild-type samples, we occasionally noted small-sized megakaryocytes with intense staining for BclxL (Figure 1C). This is also supported by a recent report, according to which BclxL is expressed in small size, developing megakaryocytes in culture, but not in mature cells.9 The fact the immunohistochemistry did not reveal a significant difference in the level of BclxL protein in different transgenic lines (eg, Tg12 and Tg16 shown in Figure 1B) could be due to saturating levels of BclxL mRNA in the transgenic megakaryocytes.

We also followed apoptosis in megakaryocytes in one of the transgenic lines using the TUNEL assay. This assay cannot be performed simultaneously with immunohistochemistry for BclxL because of quenching and overlap of colors. BM cultures were enriched for megakaryocytes as described in “Materials and methods.” The transgenic sample displayed a significant number of apoptotic cells, in accordance with the above conclusion that BclxL is absent in a fraction of the cells. This assay presented with a relatively high variability. For instance, in 3 experiments performed, a range of 1.5- to 1.9-fold reduction in the number of apoptotic cells was observed in the transgenic samples as compared with control (Figure 2A). Hence, we could only conclude that there is a trend of reduced numbers of apoptotic megakaryocytes in these transgenic mice. Figure 2B focuses on a culture of BM cells enriched for megakaryocytes in which BclxL is detected, showing that its level in the transgenic sample is higher than the typical level detected in a control cell. Note also in Figure 2B cells in the control culture (arrows), in which proplatelet formation is in process, and endogenous BclxL levels are low. Processes typical of proplatelet formation were routinely further confirmed by visualizing the cells at high magnification (eg, Figure 7A). Our survey of several fields (approximately 500 megakaryocytes) indicated that not all megakaryocytes with low level of BclxL are engaged in proplatelet formation, but those with high abundance of BclxL do not display proplatelet formation (eg, Tg19 in Figure 2B, arrowheads).

Megakaryocyte number could be under a tight homeostasis in vivo so that a change in the rate of apoptosis may cause a small or no increase in the steady-state level of megakaryocytes in the marrow. To examine this possibility, BM samples cytospun on slides were subjected to in situ staining for acetylcholinesterase, a unique marker for rodent and cat megakaryocytes14 (Figure3A). We determined a constant tendency of augmented megakaryocyte number in BM of transgenic animals, as compared with control, as follows: a 1.62-fold increase in Tg12 (P > .05), a 1.54-fold increase in Tg13 (P < .05), a 1.20-fold increase in Tg16 (P > .05), and a 1.60-fold increase in Tg19 (P < .05). The differences observed for Tg13 and Tg19, as compared with control, are statistically significant, as derived by t tests. MGDF increased megakaryocyte number in wild-type mice by 2.4-fold (P = .05; Figure 3A). Tg13 and Tg19 were also subjected to histologic evaluation of the spleen, as we described before.16 A blinded evaluation of 3 different splenic sections from each group revealed that the total number of megakaryocytes in each of these lines was increased by approximately 2-fold compared with control mice (not shown). The overall cellularity of the spleen in control and transgenic animals was similar. Spleen sections of MGDF-treated mice (analyzed as a control for this assay) displayed a 2.5-fold increase in the number of megakaryocytes, as also observed before.16 Taken together, this assay and the above in situ staining evaluation indicated a moderate, but significant, increase in megakaryocyte number in the transgenic lines 13 and 19. These changes were not accompanied by any significant shift in the ploidy level, as derived by t test evaluation of the data (Table 1). Because the number of megakaryocytes may affect the serum levels of TPO, we examined this parameter in Tg16 (which displayed no significant change in megakaryocyte number) and in Tg19 (which displayed an increase in megakaryocyte numbers) and found it to be similar to control mice (in the range of 3.7-4.1 × 10³ pg/mL). This indicated that the phenotype of the transgenic mice was not based on secondary effects on the levels of TPO in the circulation.

Whereas MGDF injection into control mice increased platelet counts, platelet levels in all the PF4-hBclxL transgenic lines were not significantly different from those in control mice (Figure 3B). Please note that platelet levels were monitored in mice at age 6 to 12 weeks and the data were compiled, as shown in Figure 3. This was pursued to analyze platelet counts in most of the mice that were used for other analyses. This range of age, although small, may have contributed to the variability observed (SD) in platelet counts, in addition to the well-known variability that exists between individual mice of the same strain. However, when data were analyzed by comparing smaller groups of age-matched mice, there was still no significant difference in platelet counts in the transgenic and nontransgenic mice. By both types of group analysis, MGDF injection significantly increased platelet counts. Bleeding time was comparable in control and transgenic mice (not shown). The extent of platelet aggregation in response to adenosine diphosphate or arachidonic acid was similar in the control and transgenic mice (not shown). Also, platelet volumes in transgenic mice were not significantly different from those in control mice (Table2). Finally, the transgenic phenotype was not associated with any change in white or red blood cell counts (Table2). The statistical differences between groups were derived byt tests.

Platelet levels may be influenced by megakaryocyte ploidy and number as well as by the integrity of the membrane system within the mature megakaryocyte. Because the increased number of megakaryocytes in BclxL transgenic mice (Tg19) was not associated with any increase in the numbers of platelets, we speculated that under normal conditions, a tight regulation of platelet homeostasis prevents a significant change in platelet levels on BclxL overexpression. If this were the case, platelet elevation under stress might be different in the control and transgenic mice. To test this hypothesis, age- and sex-matched transgenic and wild-type mice were subjected to experimental ITP and platelet recovery was monitored. Tg19 and Tg16 were chosen for these experiments because both strains expressed the transgene (Figures 1 and2B), but only Tg19 showed a significant increase in megakaryocyte number, hence allowing us to examine if platelet recovery also depends on changes in the baseline level of megakaryocytes. As shown in Figure4, platelet recovery was significantly and similarly lower in both transgenic lines compared with control (P < .01), suggesting that megakaryocytes in the transgenic mice were impaired in their ability to fragment into platelets. In this type of analysis we observed a lesser variation in platelet counts as compared with the data in Figure 3B, most likely because emphasis was given to elect perfectly matching mice (age and sex) for this injection protocol.

There are several facets to the study of the mechanism by which BclxL overexpression may induce impairment in platelet fragmentation. These include, for instance, identification of the signaling pathways mediating the effect (which would typically require biochemical assays on abundant cells), structural changes within the cells (which could be performed in situ), and so forth. In the current study we elected to focus on examining potential changes in the intracellular demarcation membrane system, which have been described as important for platelet fragmentation (see “Discussion”). Electron microscopic analysis indicated that the majority of control megakaryocytes (> 90%) exhibited well-demarcated cytoplasmic territories with extensive dilation of the demarcation membrane system (Figure5A-B). As described in “Discussion,” this system has been described before as the origin of platelets. In contrast, the majority of transgenic megakaryocytes (> 75%) showed poorly developed demarcation territories that consisted of a very congested demarcation membrane system (Figure 5C-F, especially panels E and F). Although a small population of the megakaryocytes from the transgenic line (< 25%) did contain some areas of platelet territory delineation, the majority of megakaryocytes from these mice had a significant change in these membranes. Also, megakaryocytes of transgenic mice exhibited an extensive outer ring of cytoplasm with no typical platelet extensions (Figure 5D,F, arrows). In control megakaryocytes the peripheral sheet unfolds from the cell core and displays an extensive platelet-shedding pattern (Figure 5B, arrow). It is important to note that our analyses were conducted in 3 different mice for each group (control and transgenic), showing the same results. At least 50 megakaryocytes from these 3 different mice (control and transgenic) were analyzed. It should be also pointed out that the BclxL antibodies available thus far are not yielding a signal when used under the fixation methods necessary for electron microscopic analysis.

We next sought to investigate if a treatment of BclxL transgenic mice with MGDF would affect megakaryocyte expansion and platelet levels to the same extent as it does in control mice. MGDF caused a 2-fold increase in circulating platelets in control mice as well as in Tg19 (Figure 6A). The number of megakaryocytes did not change significantly in the transgenic line (P > .05; which had already an elevated baseline), whereas it increased almost by 2.5-fold in control animals on treatment with MGDF (Figure 6B). Electron microscopic analysis showed that MGDF-treated transgenic mice displayed more megakaryocytes (about 75% of 20 analyzed) containing membranous structures that were dilated as compared with nontreated transgenic mice (Figure 6C).

To further study the consequences of BclxL overexpression on thrombopoiesis, we used an in vitro model of proplatelet formation as used by others.24 As others, we noticed that control megakaryocytes form several types of extensions: extended cytoplasmic processes constricted at platelet-sized intervals (Figure7A), pseudopodialike processes (Figure7B), and mainly long, thin extensions (Figure 7C). As shown in Figure7E, cultures from Tg19 displayed a 3-fold decrease in the number of proplatelet-forming megakaryocytes (Figure 7D) as compared with control (P < .05).

The process of megakaryocyte fragmentation into platelets in vitro has been clearly associated with apoptosis.7,25 Apoptosis has been described in mature megakaryocytes in BM26 and CD34⁺-derived cultured cells.8 An interest in apoptosis as a possible mechanism of platelet shedding emerged from the observation that the process of platelet fragmentation is reminiscent of the course of programmed cell death in its multiple features: ruffling, blebbing, and condensation of the plasma and nuclear membranes; reorganization and disruption of the cytoskeletal architecture; DNA fragmentation; and, finally, cell shrinkage and formation of apoptotic bodies.27 One of the major regulators of cytoskeletal organization and apoptosis is BclxL,10 the level of which is reduced in mature cultured megakaryocytes.9 To establish the role of apoptosis in platelet fragmentation in vivo, one may need to explore effects of deregulated expression in megakaryocytes of different proapoptotic as well as antiapoptotic proteins. In the current study, we have developed the first tool needed to approach this question in vivo. We aimed at addressing whether deregulated expression of BclxL has an effect on megakaryocyte development or fragmentation into functional platelets. To this end, transgenic mice were generated in which BclxL was overexpressed specifically in megakaryocytes via the PF4 promoter.

The transgenic phenotype of PF4-BclxL mice included a moderate increase in the number of megakaryocytes, as was also observed in MGDF-treated control mice (single-dose injection). In contrast to the latter case, however, the platelet number did not increase in mice overexpressingBclxL. This led us to suspect that more megakaryocytes in the transgenic mice have an impaired ability to release platelets. Interestingly, a different study described a 50% reduction in platelet counts in transgenic mice constitutively expressing theBcl-2 gene in the whole hematopoietic compartment.25 Because in this model Bcl-2overexpression was in all hematopoietic lineages, it was not determined how lineage redistribution may affect the number of mature megakaryocytes in vivo and platelet levels. The role of Bcl-2 family members in platelet fragmentation was also implicated in a recent study,28 showing that platelet levels are reduced by half in mice that lack the gene for Bim, a negative regulator of Bcl-2–like prosurvival factors. In this study, too, the lack ofBim in all lineages could have affected cytokine production and hence platelet levels. Our current investigation is novel in that it links directly changes in BclxL expression in megakaryocytes in vivo to megakaryocyte ultrastructure and platelet levels.

At the present time 2 models of thrombopoiesis exist. The platelet territory model has been suggested based on freeze-fracture and thin-section electron microscopic analysis of the internal membranes of megakaryocytes. According to this model, internal membranes demarcate fields or territories of prepackaged platelets, and fragmentation of the cytoplasm releases these platelets.29,30 An alternative flow model incorporates the observation that mammalian platelets assemble through intermediate long cytoplasmic extensions, called proplatelets, consisting of multiple platelet-sized beads along their length.24,31-33 Several studies have captured mammalian megakaryocytes extruding proplateletlike processes into the BM sinusoids.34,35 Additionally, cultured megakaryocytes were shown to develop various degrees of the demarcation membrane system in addition to proplatelet extensions.8,36 These 2 models are not viewed as mutually exclusive,4,37 because the position of long cytoplasmic extensions and proplatelet formation may be determined by the structure of intracellular membranes that protrude and connect to the cell surface. A linkage between poorly developed demarcation membranes and a lack of platelet fragmentation was observed in several genetically modified mouse models in which an increase or no change in megakaryocyte numbers was detectable.16,38 In our current study, examination of the sections of mouse BM cells by transmission electron microscopy revealed that the majority of control megakaryocytes exhibited well-demarcated cytoplasmic territories with extensive dilation of the demarcation membrane system and with large cisterns and vacuoles present. All of the above are believed to lead to demarcation of platelet fields in mature megakaryocytes.36 

In contrast to control megakaryocytes, the majority of transgenic cells showed poorly developed demarcation territories that consisted of a very congested demarcation membrane system, which may prevent delineation of future platelet fields. In addition, megakaryocytes of transgenic mice exhibited an extensive outer ring of cytoplasm with poorly developed platelet extensions. We speculate that these significant ultrastructural changes did not result in a parallel decrease in platelet counts because they were only prevalent in about 75% of the cells, and because the total number of megakaryocytes in some of the transgenic mice was increased (in accordance, and as discussed below, only about 70% of the transgenic megakaryocytes displayed high levels of BclxL). For example, Tg19 demonstrates a 1.6-fold higher number of megakaryocytes, of which about 25% display normal demarcation membranes. Fragmentation from these normal cells, in addition to a potential fragmentation from the cells with impaired demarcation membranes, may account for maintaining an adequate platelet level in the transgenic mice. In agreement with the above-described ultrastructural characteristics, the BclxL transgenic mice had a poor rate of platelet recovery after experimentally induced ITP, as compared with control mice. Also in accordance, megakaryocytes isolated from BM of transgenic mice had a low frequency of cells able to produce proplatelets in vitro. We could not collect a sufficient number of platelets from this system for proper evaluation of level and activation potential.

Injection of MGDF reversed to a large extent the effects of BclxL on the ultrastructure of megakaryocytes and on platelet levels, suggesting that the profile of development of the demarcation membrane system in megakaryocytes is relevant to the determination of platelet levels. It is conceivable that high levels of this cytokine initiate signaling in mature megakaryocytes that promote destabilization of Bcl-2 proteins (endogenous and transgene) or its inactivation via posttranslational modifications.39,40 It is also possible that at this stage of cell maturation, a boost of MGDF increases Bax levels or induces its dephosphorylation and translocation to the mitochondrial membranes, resulting in a low ratio of BclxL/Bax, which would promote apoptosis.41 Future studies will address these possibilities.

Our current model clearly indicates that BclxL levels are regulated in mature megakaryocytes in vivo and that deregulated, high-level expression of BclxL impairs the ability of the cells to fragment into platelets. In addition, this study also provides a facet related to the mechanism responsible for this phenomenon by linking it to significant changes in the development of demarcation membrane structures.

We thank Dr Ramesh Shivdasani for reading the manuscript and for valuable insights on the assay of proplatelet formation, Dr Beatrice Hechler for assistance with platelet aggregation studies, and Dr Hou-Xiang Xie and Robin MacDonald at our BUSM Central Transgenic facility.