Megakaryocytic (Mk) cells mature adjacent to bone marrow (BM) sinus walls and subsequently release platelets within the sinusoidal space or in lung capillaries. In contrast, primitive stem and Mk progenitor cells reside the furthest away from the BM sinus walls. The existence of pH gradients in the BM raises the question of whether pH affects Mk maturation and differentiation. We generated Mk cells from peripheral blood CD34+ cells in a serum-free medium at different pH levels (7.2, 7.4, and 7.6) and found that higher pH resulted in an earlier and higher polyploidization of CD41+ Mk cells and an earlier onset of Mk-cell apoptosis. The peak day of high ploidy was correlated well with the onset day of Mk apoptosis, thus suggesting that a decline in the fraction of high-ploidy Mk cells at the late culture stage is caused by Mk-cell apoptosis. We further explored the relationship between Mk-cell maturation and apoptosis by employing an antiapoptotic agent Z-Val-Ala-Asp(Ome)-FMK (zVAD). Addition of zVAD led to an average 30% higher and 2.8-day delayed polyploidization, while apoptosis was delayed by 2.4 days. Faster depletion of CD34+ cells and an earlier peak in the fraction of larger colony-forming Mk cells (BFU-Mks) were also observed at higher pH. Taken together, these data suggest that higher pH promotes Mk-cell differentiation, maturation, and apoptosis.
The bone marrow hematopoietic compartment (BMHC) is a complex tissue where hematopoietic cells interact with a heterogeneous microenvironment consisting of diverse hematopoietic and stromal cells, cytokines, extracellular matrix, and variable physiological conditions, such as pO2 and pH levels [1,, 2]. While much is known about the roles of stromal cells and cytokines [3,, 4], little is known about the effects of pH on megakaryocytic (Mk) differentiation and maturation. Hematopoietic stem and progenitor cells, including Mk progenitor cells, reside in the core of the BMHC located the farthest away from the sinuses, while Mk cells mature adjacent to bone marrow sinus walls, whereby they may release platelets directly into circulation through gaps of the sinusoid wall , or they may enter the circulation and release platelets in the pulmonary microvasculature . As cells differentiate, they move closer to the sinus lining. BM microphotographs [2,, 7,, 8] show that there are as few as four to eight to as many as 16-20 cells tightly packed between two sinuses in any one direction. Thus, many hematopoietic cells (and especially the most primitive stem and progenitor cells) are two to eight cells (or 20 to 80 μm, assuming a conservative 10 μm cell diameter) away from the closest sinus. Another measure of the likely distances of cells from sinuses is the size (80 to 1,200 μm in diameter) of the lymphocyte nodules, in which lymphocytes are organized in the human BMHC . Along with pO2 gradients, it has been demonstrated experimentally that pH drops from 7.35 to around 7.1 within about 25 μm from a blood vessel in normal subcutaneous tissue , and that the lowest pH (and pO2) values are observed the farthest away from the vessel wall . Furthermore, it has been shown that arterial blood has a pH of 7.4, while venous blood has a pH of 7.35 . These observations suggest that there exist spatial pH variations in the BMHC, and that primitive stem and progenitor cells are exposed to and differentiate in a lower (≤7.1) pH environment, while differentiated Mk cells mature at a higher pH (7.35-7.40) environment.
The effects of oxygen tension (pO2) on the granulocytic (G), erythroid (E), and Mk lineages have been extensively studied in our laboratory [13–, 17]. We reported that these effects were physiologically relevant and consistent with the expected exposure of the various cell types to variable pO2 levels in the BMHC. In addition to pO2, pH is also a powerful regulator of cell proliferation, differentiation, and cytokine secretion . We reported that culture pH was a potent proliferation and differentiation factor of E and G lineages, and that the cloning efficiencies of primitive erythroid progenitors (BFU-E) were ninefold higher at low pH (7.1) compared with high pH (7.6) [19,, 20]. A small pH increase of 0.2 units over the physiological value (7.4) yielded significant reductions (42%-85%) in cloning efficiencies for all E and G progenitor types and cytokine combinations tested. Differentiation of BFU-E in semisolid assays progressively increased as pH was raised from 6.95 (no colonies detected) to 7.4 (maximum colonies detected) to 7.6 (maximum hemoglobin content). In liquid cultures of peripheral blood (PB) CD34+ cells, E differentiation proceeded faster at high pH and was blocked at an intermediate stage by low pH . In G cultures, cell expansion was enhanced at pH 7.1-7.25, with twice as many total cells and G precursors produced as at pH 7.4 . For T-cell proliferation, a greater than threefold increase in the proliferation capacity was reported in pH 7.0 and 7.2 cultures compared with pH 7.4 cultures . Rich  has shown that optimal proliferation of murine macrophages occurs at pH 7.6-8.0, and that erythropoietin secretion by macrophages varies with pH.
The aim of this study is to understand the effects of pH on Mk differentiation, maturation, and apoptosis, and to test the hypothesis that pH regulates these processes in a physiologically relevant manner. It is expected that Mk progenitors will differentiate under conditions of low pH (and pO2), while maturation and platelet release will take place under high pH (and pO2) conditions. We cultured mobilized PB CD34+ cells in a serum-free medium supplemented with interleukin-3 (IL-3), Flt-3 ligand (Flt-3), and thrombopoietin (TPO) at three different pH levels (pH 7.2, 7.4, and 7.6). As in prior studies , the high pH value of 7.6 was used to enhance any possible pH effects and to contrast such effects with those at the low and physiologically relevant pH values of 7.2 and 7.4. Furthermore, the antiapoptotic agent Z-Val-Ala-Asp(Ome)-FMK (zVAD) was used to explore the relationship between Mk maturation (polyploidization) and apoptosis.
Materials and Methods
CD34+ Cell Collection and Cultures
Cultures were initiated with adult human CD34+ cells selected from apheresis products obtained from normal donors (AllCells; Berkeley, CA; http://www.allcells.com) or cancer patients (Northwestern Memorial Hospital, Chicago, IL) following stem cell mobilization with G-CSF. Patient blood samples were collected after informed written consent using protocols approved by the Institutional Review Boards. Positive selection of CD34+ cells was performed using MiniMACS columns as recommended by the manufacturer (Miltenyi Biotec; Auburn, CA; http://www.miltenyibiotec.com). The purity of the CD34+ cell population ranged from 94%-98%. All cultures were performed with serum-free medium X-VIVO 20 (BioWhittaker; Walkersville, MD; http://www.cambrex.com/default.asp) supplemented with 50 ng/ml TPO (Genentech; South San Francisco, CA; http://www.gene.com), 50 ng/ml Flt-3 (R&D Systems; Minneapolis, MN; http://www.rndsystems.com), and 5 ng/ml IL-3 (R&D Systems) in either 75 or 150 cm2 T flasks with a respective medium volume of 30 or 60 ml, corresponding to a liquid height of 0.4 cm. To ensure saturation, TPO (50 ng/ml) was added to cultures every 5 days . Cultures were seeded with a cell density ranging from 35,000 to 58,000 cells/ml and incubated for up to 21 days at 37°C in a fully humidified environment containing 5% CO2 and 20% O2. The cell density was kept under 200,000 cells/ml throughout all cultures by dilution feeding with medium at the same pH (see below) to minimize pH changes due to cell metabolism.
Culture pH Maintenance
Three different culture pH levels (low: 7.2; medium: 7.4; and high: 7.6) were used in this study. Each pH level was established based on a predetermined titration curve by addition of 1 N NaOH or 1 N HCl . The pH-preadjusted media were equilibrated in the same incubator used for cultivation for at least 12 hours prior to culture inoculation. The pH of each culture was regularly checked using a blood gas analyzer (Instrumentation Laboratory; Lexington, MA; http://www.ilww.com) and was found to be within ±0.05 pH units of the target level.
zVAD Addition and DMSO Controls
In order to examine the relationship between Mk maturation and apoptosis, the apoptotic inhibitor zVAD (Enzyme Systems; Livermore, CA; http://www.enzymesys.com), dissolved in dimethyl sulfoxide (DMSO) (purity ≥99.9%; Sigma; St. Louis, MO; http://www.sigmaaldrich.com) to give a zVAD stock solution of 80 mM, was added to cultures at a final zVAD concentration of 40 μM in medium . Due to the possible instability of zVAD in the medium, zVAD was supplemented every 2 days at the concentration of 20 μM according to the manufacturer's suggestion. Cultures with addition of the same amounts of DMSO, but without zVAD, were used as controls.
Cells were plated in a collagen gel matrix (Stem Cell Technologies; Vancouver, BC, Canada; http://www.stemcell.com) supplemented with TPO (50 ng/ml) and IL-3 (10 ng/ml) and incubated in a 5% CO2 and 5% O2 atmosphere . On days 0 and 3, the plating density was 3,000 cells/ml, and by day 12 of the culture, the plating density was linearly increased to 12,000 cells/ml. Cells were plated in sterile chamber slides (Fisher Scientific; Pittsburgh, PA; http://www.fisherscientific.com). After 14 days of incubation, cells were fixed with acetone:methanol at a ratio of 3:1. Slides were then sequentially incubated with a mouse anti-CD41 antibody, a Histomark streptavidin-horseradish peroxidase system (normal goat serum, biotinylated goat anti-mouse IgG, and streptavidin-peroxidase) (Kirkegard & Perry Labs; Gaithersburg, MD; http://www.kpl.com), and a peroxidase chromogen solution (Biomeda; Foster City, CA; http://www.biomeda.com), and finally visualized and counted using a Diastar microscope (Reichert-Jung; Buffalo, NY) . Mk cells and platelets, which express glycoprotein IIb/IIIa (CD41), appear pink. Mk colonies were subdivided into two groups: large (≥50 cells per colony) and small (<50 cells per colony) colonies. Large Mk colonies arise from more primitive Mk progenitors (BFU-Mk), whereas small Mk colonies are produced from more mature colony-forming units-Mk progenitors (CFU-Mk).
Flow Cytometric Detection of CD41, CD11b, CD15, and CD34
Cells were washed with phosphate-buffered saline (PBS) containing 2 mmol/l EDTA for detaching platelets and 0.5% bovine serum albumin (BSA), incubated with phycoerythrin (PE)-labeled anti-CD11b and fluorescein isothiocyanate (FITC)-labeled anti-CD15, or PE-labeled anti-CD41 and FITC-labeled anti-CD34 (Becton Dickinson; San Jose, CA; http://www.bd.com), and then analyzed on a Becton Dickinson FACScan flow cytometer using CellQuest software (Becton Dickinson). Propidium iodide ([PI] 200 μg/ml) was added to samples shortly prior to acquisition on the flow cytometer to exclude dead cells from living cells. Depletion of CD34+ cells was described by the depletion rate, which is the average drop in the number of CD34+ cells over time.
Flow Cytometric Detection of Mk Cell Apoptosis
Simultaneous staining for CD41, annexin V, and PI was used to recognize nonapoptotic Mk cells (CD41+, annexin V−, and PI−), apoptotic Mk cells (CD41+, annexin V+, and PI−), and dead cells (PI+) . Briefly, cells were washed with PBS containing 2 mmol/l EDTA and 0.5% BSA, incubated sequentially with PE-labeled anti-CD41, FITC-labeled annexin V (in the presence of Ca2+) (PharMingen; San Diego, CA; http://www.bdbiosciences.com/pharmingen), and PI (50 μg/ml), and then analyzed by flow cytometry. The ratio of the number of apoptotic Mk cells to the total number of PI− Mk cells was used to calculate the fraction or percentage of apoptotic Mk cells. The time point for onset of apoptosis was determined as follows: we first calculated the time derivatives of the fraction of apoptotic Mk cells for all pairs of neighboring time points; the first time point, which gave the steepest increase in the derivative, was then taken as the onset day of Mk apoptosis.
Flow Cytometric Analysis of Mk-Cell Ploidy
Cells were washed with PBS containing 2 mmol/l EDTA and 0.5% BSA and then incubated sequentially with FITC-labeled CD41 antibody (Becton Dickinson), 0.5% paraformaldehyde (Fisher Scientific) for fixation, cold 70% methanol for permeability, RNAse (Sigma), and PI (50 μg/ml) to stain DNA . Samples were analyzed by flow cytometry, and data were acquired using CellQuest software and subsequently analyzed using ModFit 3.0 software (Becton Dickinson). Non-Mk cells (CD41−) were gated out. The fraction of higher ploidy Mk cells was defined by the ratio of the number of Mk cells with 8 N or higher DNA ploidy to the total number of CD41+ Mk cells.
Benzidine Staining for Estimation of Erythrocytes
Cells were exposed to a fresh solution consisting of 10 μl of 30% H2O2 (Fisher Scientific) in 2.5 ml of a benzidine stock solution made up of 0.2% benzidine dihydrochloride (Sigma) in 0.5 M acetic acid . The final solution (10 μl) was added to 90 μl of cell suspension before counting the cells in a hemacytometer. Cells appearing blue contain hemoglobin and include cells ranging from the late basophilic normoblast stage through mature erythrocytes.
Statistical data analysis was conducted using the single-sample t test. For comparison of ratio values between two different conditions (two pH levels or with/without zVAD addition at the same pH), the null hypothesis was that the mean of the ratio was not significantly different from the value of one, while the alternative hypothesis was that the mean of the ratio was significantly greater than one. For comparison of the difference of a parameter between two different conditions, the null hypothesis was that the mean of the difference was not significantly different from zero, while the alternative hypothesis was that the mean of the difference was significantly greater than zero. Points or numbers marked with an asterisk (*) in tables indicate the rejection of one of the above null hypotheses and the acceptance of the respective alternative hypothesis at a confidence level of >95% (p < 0.05).
The effects of different pH levels were evaluated on total cell expansion, Mk-cell expansion, high ploidy content, Mk-cell apoptosis, BFU-Mk and CFU-Mk expansion, and non-Mk cell content. Four sets of Mk experiments with four different donor samples were carried out. CD34+ cells were cultured at three different pH levels (7.2, 7.4, and 7.6) with or without addition of zVAD.
Total and Mk Cell Expansion
Large variations in total and Mk-cell fold expansion, but no consistent pH effects, were observed among donor samples (data not shown). The total cell fold expansion varied from 1.5 to 80, and the Mk-cell fold expansion ranged between 0.34 and 12 among the four sets of experiments. The highest fraction of Mk cells varied from 31%-52% in the four culture sets.
pH Effects on Mk Maturation and Apoptosis
As Mk cells differentiate, they switch from cell division (mitosis) to polyploidization (endomitosis). As a result, the fraction of Mk cells with high ploidy increases (Fig. 1). Mk-cell apoptosis and proplatelet formation are observed toward the end of Mk maturation . Compared with medium (7.4) and low (7.2) pH, Mk cells matured faster and underwent apoptosis earlier at high pH (7.6) (Fig. 1, Table 1). The insignificant differences between pH 7.4 and 7.2 (Table 1) may be due to the large sampling time interval of 2-3 days that makes small differences (e.g., about 1 day) difficult to detect. We also found that higher pH resulted in a higher fraction of Mk cells with high ploidy (Table 1). The onset day of apoptosis at different pH values was correlated well with the peak day of the high-ploidy Mk fraction (r2 = 0.93), thus suggesting that the reduction, after reaching a peak, in the fraction of high-ploidy Mk cells was due to their apoptotic death.
Table Table 1.. pH effects on Mk polyploidization and apoptosis (n = 4)
Mean ratio of peak values of % high-ploidy Mk cells
Mean difference in onset day of Mk apoptosis
zVAD Effects on Mk Maturation and Apoptosis
In order to further explore the relationship between Mk maturation and apoptosis, we used the pseudosubstrate caspase inhibitor zVAD. zVAD inhibits the activation of a number of caspases by binding with their active sites, thus delaying apoptosis . Addition of zVAD delayed apoptosis and resulted in a higher fraction of high-ploidy Mk cells (Fig. 2). zVAD significantly delayed the peak in the fraction of high-ploidy Mk cells by an average of 2.8 days (p = 0.001) and the apoptosis onset by 2.4 days (p = 0.008). Furthermore, zVAD resulted in a significantly higher (30%) fraction of high-ploidy Mk cells (p = 0.018). These data further confirm the suggestion that Mk apoptosis is responsible for the decrease in the fraction of high-ploidy Mk cells. Once the apoptosis of high-ploidy Mk cells was inhibited, more Mk cells with high ploidy could be obtained. The average increase in the fraction of high-ploidy Mk cells due to zVAD addition correlated well with the average decrease in the fraction of apoptotic Mk cells (r2 = 0.97) for three of the four samples. Sample III was the exception for which the increase in DNA content was considerably larger than the decrease in the fraction of apoptotic cells. The discrepancy might be due to a much greater increase in proliferation after zVAD addition for Sample III (∼ twice as great as for the other samples). The solvent DMSO has been shown to have negative effects on cultured cells at higher concentrations . In this study, the final DMSO volumetric content was about 0.3% volume by volume. At this concentration, no significant effects of DMSO on Mk cultures were found in terms of Mk fold expansion, fraction of high-ploidy Mk cells, Mk-cell apoptosis, or CFU-Mk (data not shown).
pH Effects on Differentiation of Primitive Mk Progenitor Cells
Although the total number of Mk progenitor cells (BFU-Mk + CFU-Mk) did not vary in a consistent pattern with pH, high pH resulted in an earlier and higher fraction of more primitive Mk progenitor cells (BFU-Mk) (Fig. 3). Furthermore, BFU-Mk cells were depleted faster at high pH, whereas lower pH yielded an extended expansion of BFU-Mk cells (Fig. 3). Although only the kinetics of BFU-Mk progenitors were affected by pH, these data demonstrate that in addition to Mk maturation and apoptosis, high pH appears to alter Mk differentiation at its early stages (BFU-Mk) as well. Interestingly, Mk-cell expansion was found to be proportional to the total number of Mk progenitor cells, which is the cumulative density of total CFU-Mk cells (Fig. 4), indicating that the total number of Mk progenitors, but not the initial number of Mk progenitor cells, was the decisive parameter for Mk expansion. Thus, the finding that pH did not affect total Mk expansion was consistent with the finding that pH did not affect total Mk-progenitor expansion.
Non-Mk Cell Populations at Different pH Values
The observation that pH promotes differentiation of Mk progenitor cells was further confirmed by the finding that CD34+ cells were depleted faster at higher pH. The depletion rate of CD34+ cells at pH 7.6 was significantly higher (by 25% and 43%, respectively, n = 3) than at pH 7.4 and 7.2. These CD34+ cell-initiated Mk cultures contained several other cell types. Monocytic cells did not appear in Mk cultures at a significant level (<5%) on any culture day and at any pH value. Higher pH resulted in a higher percentage of G cells at the earlier stage and a higher percentage of E cells at the later stage (data not shown). Granulocytic cells expanded at the earlier stage, then steadily declined and never matured, probably due to the lack of necessary G proliferation/maturation-promoting cytokines. The E-cell population increased with culture time (21% on day 9; 32% on day 12).
Our results show that high pH promotes differentiation, maturation, and apoptosis of Mk cells. Mk cells at high pH matured faster and underwent apoptosis earlier. High pH also promoted faster depletion of primitive Mk progenitor cells (BFU-Mk), while lower pH extended their expansion. It has been reported recently that higher (20%) pO2 promoted maturation and apoptosis of Mk cells compared with lower (5%) pO2 [13,, 16]. Higher pO2 also yielded a higher fraction of high-ploidy Mk cells and a higher fraction of apoptotic Mk cells. In contrast, 5% pO2 resulted in a higher number of total CFU-Mk cells and a higher fraction of BFU-Mk cells than 20% pO2 [13,, 16]. Although there are no comparable studies of pH effects on Mk cultures in the literature, these oxygen results can be viewed as consistent with our findings, considering that in tissues, higher pH values coincide with higher pO2 values [10,, 11]. Our findings are also analogous to those reported for E differentiation, whereby higher pH promotes differentiation while lower pH enhances BFU-E proliferation . Unlike Mk and E cells, proliferation and differentiation of G cells are enhanced at lower pH , which is consistent with the topology of differentiation in the BMHC: G cells differentiate and mature away from the BM sinus, while both Mk and E cells mature adjacent to the BM sinus walls.
Our finding that higher pH promotes Mk apoptosis is consistent with a previous report  that high pH promotes T cell apoptosis. These effects of pH on apoptosis are somehow unexpected in view of the previously established relationship between apoptosis and low intracellular pH (pHi) [29–, 31] and the generally accepted relationship between pHi and culture pH . Indeed, it was reported that protection against apoptosis by overexpression of antiapoptotic proteins, such as Bcl-2 and v-Abl, in hematopoietic cells was accompanied by prevention of a decrease in pHi [30,, 31]. Artificial reduction of pHi was reported in some cases to induce apoptosis , while in other cases, it was reported that intracellular alkalinization blocked apoptosis . It appears then that apoptosis may be induced by either low or high pH through, at the present, unknown mechanisms.
Although the exact mechanisms of apoptosis involvement in Mk maturation are still unclear, several reports indicate a close relationship among apoptosis, Mk maturation, and platelet release. It has been shown by Zauli et al.  that the onset of apoptosis coincides with the maximum in the high-ploidy Mk-cell fraction. They also found that the number of platelets released in the culture was well correlated with the fraction of apoptotic Mk cells . Ogilvy et al.  and Bouillet et al.  have independently confirmed that such a correlation exists also in vivo by deleting the proapoptotic gene Bim. In Bim knockout mice, the Mk cell count was normal, but platelets were half the normal amount, thus suggesting that apoptosis mediated by the caspases may trigger or regulate platelet release. Recently, Sanz et al.  reported that the antiapoptotic protein Bcl-xL was expressed in immature and mature Mk cells and platelets but not in senescent Mk cells. Their interpretation was that senescent Mk cells, which have already released platelets, may undergo apoptosis due to the lack of antiapoptotic protection. In this study, we observed that the onset of apoptosis coincided with the peak day of high-ploidy Mk-cell fraction in Mk cultures at different pH values (Fig. 2, Table 1), which is consistent with the observations of Zauli et al. . We have also shown that the addition of zVAD delayed the onset of apoptosis and simultaneously delayed and enhanced the generation of high-ploidy Mk cells (Fig. 2). These findings support the hypothesis that apoptosis triggers platelet release rather than that platelet release leads to apoptosis. If senescent Mk cells underwent apoptosis due to the lack of antiapoptotic protection as a result of platelet release, an effective antiapoptotic agent would prevent Mk-cell apoptosis but not a decrease in the number of high-ploidy cells as a result of platelet release. Our results (Fig. 2) also indicate that more Mk cells with high ploidy can be obtained using an effective antiapoptotic agent such as zVAD. Finally, our results show that pH can be used as an additional (to cytokine concentrations and pO2 levels) parameter for controlling ex vivo Mk differentiation for potential transplantation therapies .
This work was supported by NIH Grant R01 HL 48276. We thank Genentech for TPO donation. We are grateful to Northwestern Memorial Hospital (Dr. J. Winter, R.W. Guo) for providing apheresis products. We thank C. Chen and R. Kaliney for laboratory assistance.