Main

In normal hematopoiesis, cell differentiation is coupled to a temporary arrest in the cell cycle, and it is believed that only cells residing in the G0 or early G1 phase are susceptible to differentiation-inducing signals from the environment(1,2). Nearly a decade ago Pardee(3) described the restriction point in the G1 phase of the cell cycle as being that point beyond which a cell is committed to DNA replication and mitosis independent of further mitogenic stimuli. In recent years, some of the mechanisms controlling the major transitions in the cell cycle have been elucidated. In association with several cyclindependent kinases, it is presumed that the D- and E-type cyclins regulate phosphorylation of key substrates required for G1 progression and S phase entry(4–6). pRB, the product of the retinoblastoma tumor suppressor gene(7), is one of the main substrates that is phosphorylated by cyclin/cyclin-dependent kinase complexes. There is evidence that hypo-p110RB is associated with the nuclear matrix and can block progression through the G1 phase(8). Only during the G1 period can significant amounts of hypo-p110RB be detected, indicating that the underphosphorylated form of p110RB is active in growth suppression(9). In G0 and early G1 phase, hypo-p110RB is complexed to the E2F protein family of transcription factors. These complexes bind to E2F-responsive promotors and serve as transcriptional repressors(10). During an interval in G1, assumed to correspond to the restriction point(7,9), increased phosphorylation of pRB, which is continued during S phase, leads to liberation of E2F, which, now uncomplexed, induces transcriptional activity(11,12).

In a previous study we demonstrated that in childhood ALL, the component of resting G0 as well as that of cycling S phase cells was small (<10%) even though the growth fraction (percentage of cycling cells) was high (94%)(13). We and others have shown that the cell-cycling status of mobilized CD34+ hematopoietic cells was not different from that of ALL cells (i.e. low numbers of resting G0 and S phase cells and a high growth fraction)(13–15). Therefore, the majority of ALL and normal CD34+ cells should be in a long G1 phase. However, no data on the distribution of ALL cells within the G1 phase of the cell cycle are available. Phosphorylation of pRB is a crucial step in cell-cycle progression from G0/early G1 to late G1/S. We therefore compared the phosphorylation status of pRB in blast cells from children with untreated ALL with that of purified CD34+ cells before and after in vitro stimulation with cytokines by immunostaining and Western blotting. Further, in both the malignant ALL and the normal CD34+ immature cell population, the proportion of cycling cells and cells in G0 and S phase were compared.

METHODS

This study has been reviewed and was approved by the Institutional Review Board of the Department of Pediatrics, University Hospitals in Bern, as well as by the Institutional Review Board of the Medical Faculty of Bern.

Cell samples. Bone marrow aspirates were obtained from 16 children with untreated precursor B cell (n = 13) or T cell (n = 3) ALL. After Ficoll (Nycomed Pharma, Oslo, Norway) density gradient centrifugation, the smears contained >90% leukemic cells as assessed by morphology. Staining of cytocentrifuge preparations with the respective MAb was performed within 24 h. Hematopoietic progenitor cells were collected by leukapheresis from seven patients after combined mobilization treatment with chemotherapy and G-CSF(16). Enrichment of CD34+ cells was performed by the CEPRATE SC stem cell concentration system (CellPro Europe, Oppem, Belgium). The mean purity (± 1 SD) of enriched CD34+ cells was 95.3 ± 1.3%, as assessed by both flow cytometry and staining of cytocentrifuge smears with an anti-CD34 MAb. It is a well-known, however as yet unexplained, phenomenon that >95% of mobilized CD34+ cells reside in the G0/G1 phase of the cell cycle, independent of the mobilization regimen used(14,15,17,18). After the selection procedure, the cells were frozen in a controlled rate freezer and stored in liquid nitrogen until use.

In vitro incubation. CD34+ cells were thawed in a water bath at 37°C and slowly reconstituted with Iscove's modified Dulbecco's medium (IMDM) (GIBCO BRL, Basel, Switzerland) and divided into two aliquots. Immediately after thawing one aliquot was used for preparing cytocentrifuge smears, whereas without delay the second aliquot was incubated for 48 h in 5 mL IMDM culture medium supplemented with 20% fetal bovine serum (FBS) (GIBCO BRL), 5 × 10-5 M 2-mercaptoethanol (Fluka AG, Buchs, Switzerland), and recombinant human cytokines. The latter were added at the following final concentrations: SCF, 20 ng/mL; IL-3, 50 ng/mL; IL-6, 20 ng/mL; Epo, 6 U/mL (all from R & D Systems, Minneapolis, MN); and G-CSF, 100 ng/mL (Amgen/Roche, Basel, Switzerland). After 48 h of incubation at 37°C in a fully humidified 4.5% CO2 atmosphere, the cells were washed, centrifuged, and stained with the respective MAb. In all experiments, the viability was >95%, as determined by trypan blue exclusion.

Monoclonal antibodies. The following MAb were used: mouse anti-statin (S-44; kindly provided by Dr. E. Wang, Bloomfield Center for Research in Aging, Montreal, Canada) recognizing a 57-kD nuclear phosphoprotein that is expressed only in resting (G0) cells(19); mouse anti-hypo-p110RB (14441A, G99-549; Pharmingen, San Diego, CA) (early G1 phase); mouse anti-pRB (14001A, G3-245; Pharmingen) (control); anti-BrdU (BMC 9318; Boehringer Mannheim, Mannheim, Germany) (S phase); anti-PCNA (PCID; Dako, Glostrup, Denmark)(20); and anti-p120, which recognizes a proliferation-associated nucleolar protein of 120 kD (FB-2; Oncogene Science, Basel, Switzerland)(21). Cytocentrifuge smears were stained in duplicate or triplicate according to the alkaline phosphatase:anti-alkaline phosphatase (APAAP) method(22). The smears were evaluated by light microscopy and at least 1000 cells per smear were counted.

Western blot. For total cell extracts, ALL cells or CD34+ cells, before and after incubation with SCF, IL-3, IL-6, Epo, and G-CSF for 48 h, were lysed in NP-40 lysis buffer (150 mM NaCl 0.9%, 50 mM Tris-HCl, pH 8.0, 1% NP-40) containing protease inhibitors (1 µg/mL aprotinin, 10 µg/mL leupeptin, 10 µg/mL pepstatin, 1 mM phenylmethylsulfonyl fluoride). After 10 passages through a 27-gauge needle, the lysate was incubated for 30 min on ice, then centrifuged for 30 min at 15 000 rpm at 4°C. Protein extract (40 µg) was heated for 3 min at 80°C in sample buffer and loaded on a 7% SDS-gel.

For immuneprecipitation, 80 µL of protein extract (40 µg/µL) was precleared for 30 min at 4°C on Protein A-agarose (Boehringer Mannheim). Then 1 µg of antibody was added to 80 µL of protein extract (40 µg/µL) followed by incubation for 4 h at 4°C. IP were isolated on a 40-µL pellet of Protein A-agarose for 60 min at 4°C using a rotor. The Protein A-agarose-IP complexes were washed 4 times with 500 µL of cold 1 × PBS. To elute IP, 20 µL of sample buffer was added and heated for 3 min at 80°C. After centrifugation for 5 min at 13 000 rpm, the supernatant was loaded on a 7% SDS gel.

After 1.5 h electrophoresis at 200 V, the proteins were blotted on a nitrocellulose membrane (Schleicher & Schüll, Dassel, Germany) for 30 min at 60 V. The blot was blocked with 0.25% gelatin in 1 × Tris buffered saline (TBS) for 60 min at room temperature, washed 15 min and 2 times 5 min in 1 × TBS, 0.02% gelatin, 0.25% Triton X-100. Antibody incubation occurred for 60 min at room temperature at 1 µg/mL. The following antibodies were used: Sc-50, amino acids 914-928 of human pRB (Santa Cruz, Santa Cruz, CA) and 14441A, amino acids 514-610 of human pRB (G99-549; Pharmingen). Probing with Sc-50 reveals all forms of pRB(23), whereas 14441A reveals only hypo-p110RB(24). After a wash step, the blot was incubated in 1 × TBS, 0.02% gelatin, and 0.25% Triton X-100 containing recombinant protein G-horseradish peroxidase (Zymed, San Francisco, CA) 0.75 µg/5 mL for 60 min at room temperature. After an additional wash step, protein was detected by chemiluminescence (Super Signal substrate; Pierce, Rockford, IL) and film autoradiography.

Statistical analysis. Statistical evaluation was performed on a personal computer applying the Statgraphics Plus software (Manugistics, Inc., Rockville, MD). Because data distribution was not normal, the nonparametric Kruskal-Wallis test was applied. A p value <0.05 was considered statistically significant.

RESULTS

Comparison between ALL and CD34+ cells by immunostaining. We analyzed the proportion of early G1 cells (hypo-p110RB), resting G0 cells (statin), cycling cells (PCNA, p120), and S phase cells (BrdU) in ALL as well as in highly purified CD34+ cells before and after incubation for 48 h in the presence of SCF, IL-3, IL-6, Epo, and G-CSF. Because no difference between T cell (n = 3) and precursor B cell (n = 13) from ALL was observed, the results obtained in these two leukemic cell populations were pooled. As shown in Figure 1, the majority of unstimulated CD34+ cells contained the pRB in its active, hypophosphorylated form (hypo-p110RB), whereas in ALL and stimulated CD34+ cells the hypo-p110RB is rarely detectable. The percentage of hypo-p110RB-positive ALL cells was low and within a range comparable with that of stimulated CD34+ cells. The percentage of cells in S phase did not differ markedly between ALL and unstimulated CD34+ cells. However, after cytokine stimulation the proportion of CD34+ cells actively incorporating BrdU (S phase) increased dramatically (Fig. 2). Table 1 summarizes the results on the proportion of resting cells (statin), proliferating cells (PCNA, p120), and cells staining positive with a control MAb recognizing all forms of pRB (phosphorylated and hypophosphorylated pRB). Both populations of ALL and unstimulated CD34+ cells contained only a few resting cells whereas the percentage of cycling cells was high. To ensure that the freezing and thawing procedure did not influence the cell-cycle status, especially the phosphorylation status of the pRB, immunostaining was also performed with thawed ALL and freshly obtained CD34+ cells. Neither in ALL nor in the CD34+ cells were the results influenced by these manipulations.

Figure 1
figure 1

Expression of hypo-p110RB in childhood ALL (n = 16) and in highly purified, mobilized CD34+ progenitor cells from peripheral blood (CD34+) (n = 7) before (t = 0 h) and after in vitro stimulation for 48 h (t = 48 h) with SCF, IL-3, IL-6, Epo, and G-CSF. Columns represent the mean ± 1 SD. A nonparametric Kruskal-Wallis test was applied to compare the three groups.

Full size image
Figure 2
figure 2

Analysis of S phase cell number (BrdU-labeling index) in childhood ALL (n = 16) and in highly purified, mobilized CD34+ progenitor cells from peripheral blood (CD34+) (n = 7) before (t = 0 h) and after in vitro stimulation for 48 h (t = 48 h) with SCF, IL-3, IL-6, Epo, and G-CSF. Columns represent the mean ± 1 SD. A nonparametric Kruskal-Wallis test was applied to compare the three groups.

Full size image
Table 1 Cell cycle status of ALL and CD34+ cells before and after cytokine stimulation
Full size table

Comparison between ALL and CD34+ cells by Western blot. When cells are in a state of quiescence or in early G1, p110RB appears as a single band of about 110 kD. As cells progress through late G1 toward S, p110RB becomes hyperphosphorylated, and the protein appears as a discrete series of bands between 114 and 116 kD. Highly phosphorylated forms of pRB are typically recognized by their slower mobility on SDS-PAGE. After mitosis, reentry into G1 is accompanied by a rapid dephosphorylation of p110RB, producing the fast-migrating, hypophosphorylated form(25).

Two representative examples of each cell type analyzed are shown in Figure 3. Blot probing was done by using Sc-50, a polyclonal antibody recognizing all forms of pRB (Fig. 3A). Multiple forms of pRB are present in ALL but not in unstimulated CD34+ cells and can only be detected in the latter after stimulation with cytokines as a result of hyperphosphorylation. It is worth noting that a single band of 110 kD was never detected in ALL cells we have analyzed so far. Between two and five forms of pRB are detected in each sample, independently of the ALL type (unpublished observation). Figure 3B shows blot probing with a specific antibody (14441A) recognizing only hypo-p110RB. Only a single fast-migrating band of 110 kD was detected in unstimulated CD34+ cells but not in ALL and stimulated CD34+ cells. To further clarify the absence of the fast-migrating band in ALL and stimulated CD34+ cells, a more sensitive method was used. Figure 3C shows the results of immunoprecipitation using the 14441A antibody to detect p110RB in ALL and stimulated CD34+ cells. Only in ALL cells but not in stimulated CD34+ cells could a band of 110 kD be detected. Because in unstimulated CD34+ cells a fast-migrating 110-kD band was already detected (Fig. 3B), the more sensitive immunoprecipitation was not performed (nd in Fig. 3C). These results indicate that significant amounts of active, cell-cycle progression blocking hypo-p110RB can only be found in resting CD34+ cells, whereas in ALL and stimulated CD34+ cells the bulk of pRB is phosphorylated. Because hypo-p110RB was not detectable in protein extracts from stimulated CD34+ cells by immunoprecipitation, the concentration is presumably even lower than that in ALL cell extracts. As with immunostaining, freezing and thawing of ALL and CD34+ cells did not influence the respective results of Western blot.

Figure 3
figure 3

Detection of pRB in total cell extracts by Western blot. T-ALL (lane 1); pre-B cell ALL (lane 2); highly purified, mobilized CD34+ progenitor cells from peripheral blood of two individuals before (lanes 3 and 4) and after in vitro stimulation for 48 h with SCF, IL-3, IL-6, Epo, and G-CSF (lanes 5 and 6). A, Probing with Sc-50 detects all forms of pRB. B, Probing with 14441A detects only hypo-p110RB. C, Probing with 14441A after immunoprecipitation. The fast-migrating band corresponds to 110 kD as determined with a low molecular weight standard (161-0305, Bio-Rad, Hercules, CA) (not shown) (nd = not done; ndd = not detectable).

Full size image

DISCUSSION

The cell-proliferation pattern of childhood precursor B cell or T cell ALL blasts shares many common features with mobilized and enriched CD34+ progenitor cells from the peripheral blood(13,14). We reasoned that this could be a pattern typical for immature normal and leukemic hematopoietic cell populations. However, the failure of ALL cells to differentiate, as opposed to CD34+ cells, could be caused by differences in the cell-cycling status in these two immature cell populations, especially with respect to the G1 phase(13). We now show that only a minority of childhood ALL cells displayed hypo-p110RB as opposed to the majority of unstimulated CD34+ cells. This observation cannot be explained by loss of pRB expression in ALL cells, because the majority stained strongly positive with the MAb recognizing all forms of pRB. Furthermore, in all protein extracts from ALL, but not from unstimulated CD34+ cells, multiple retarded bands (phosphorylated pRB) were noted. These observations are in accordance with the fact that in childhood ALL, mutations of the RB1 gene are infrequent and, if mutations are present, no protein is made(26). Nevertheless, there was a higher variability of the number of ALL cells expressing pRB by immunostaining (1 SD, ±6.1%) compared with CD34+ cells before (1 SD, ±1.7%) and after (1 SD, ±2.3%) stimulation with cytokines, presumably accounting for the observed significant difference of total pRB (Table 1). Different levels of pRB expression have also been described in acute myelogenous leukemia cells(27).

After in vitro stimulation for 48 h with cytokines, the percentage of hypo-p110RB-positive CD34+ cells decreased dramatically and the proportion of S phase cells increased. Although slowly proliferating like resting CD34+ cells(13), ALL cells showed a distribution within the G1 phase that was similar to that of rapidly proliferating CD34+ cells after in vitro stimulation with cytokines. We found no difference between ALL and unstimulated CD34+ cells for the proportion of G0 and S phase cells, and only a minor one for the number of proliferating cells. One could consider altered rates of apoptosis to influence the distributions within the cell cycle. However, this possibility is very unlikely because only a very small percentage of leukemic bone marrow cells are apoptotic(28).

Our data obtained with different techniques point to the same evidence. ALL cells predominantly contain hyperphosphorylated pRB whereas in unstimulated CD34+ cells only the active, cell-cycle progression blocking, hypophosphorylated form of pRB was detected. Only after stimulation with cytokines dtd CD34+ cells undergo cell divisions at higher frequency, resulting in different phosphorylated forms of pRB. There is a very good accordance between the results of immunostaining and that obtained by Western blot.

It is known that cell-cycle progression through the G1 phase can be blocked by hypo-pRB(9). Thus, the present data could indicate that, in contrast to unstimulated CD34+ cells, the majority of ALL cells are beyond the restriction point and therefore irreversibly committed to DNA replication and mitosis. One may speculate that ALL cells either skip the early G1 phase and progress directly to late G1, or that in these cells there is significant shortening of the early G1 phase. Because only cells residing in G0 or early G1 phase are susceptible to external stimuli inducing differentiation(1,2), our findings are consistent with the observation that ALL blasts but not CD34+ cells are unresponsive to such signals. In preliminary experiments we performed cell-cycle analyses on childhood ALL cells incubated for 48 h with SCF, IL-3, IL-6, Epo, and G-CSF. Comparing these cells before and after cytokine stimulation, no differences in the BrdU-labeling index (S phase) or the percentage of statin-positive (G0), hypo-p110RB-positive (early G1), and p120-positive cells (cycling cells) could be detected, indicating that with this combination of cytokines, ALL cells remained beyond the restriction point and committed to mitosis (unpublished observations).

Recently, it became known that the MAb against hypo-p110RB used in the present study (14441A) recognizes a specific subfraction of pRB that is unphosphorylated at the serine 608 residue(24). Antibody reactivity is entirely lost by phosphorylation of serine 608. It remains to be seen whether additional MAb recognizing other pRB subfractions could help to understand in more detail the different distribution pattern of ALL and CD34+ cells within the G1 phase of the cell cycle.

In conclusion, we have demonstrated that childhood ALL cells predominantly contain phosphorylated pRB and therefore are beyond the restriction point, whereas in normal CD34+ progenitor cells from the peripheral blood the presence of hypo-p110RB indicates their residence in early G1 phase. Although slowly proliferating like resting CD34+ cells, ALL cells showed within the G1 phase a distribution pattern similar to that of rapidly proliferating CD34+ cells. The molecular mechanisms responsible for this phenomenon remain to be elucidated.