Published online before print December 23, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The Granulocyte Research Laboratory, Department of Hematology, Rigshospitalet, University of Copenhagen, Denmark
1 Correspondence: The Granulocyte Research Laboratory, Rigshospitalet, Department of Hematology 93.2.2, 20 Juliane Mariesvej, DK-2100 Copenhagen, Denmark. E-mail: pklausen{at}rh.dk
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: granulopoiesis • cell cycle • bone marrow
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In this work, we present a scan of cell-cycle proteins regulating G1-to-S transition in myeloid precursor populations and terminally differentiated neutrophils isolated from healthy individuals.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell-cycle analysis
Cells/sample (105) were washed in ice-cold, phosphate-buffered saline and were lysed in 200 µL modified Vindeloev buffer [1% w/v sodium citrate, pH 7.4, 0.02 mg/mL RNase A, 0.3% v/v Nonidet P-40 (NP-40), 0.05 mg/mL propidium iodide (PI)] in the dark. The intensity of PI-labeled DNA was subsequently measured on a FACScan flow cytometer (Becton Dickinson).
RNA isolation and Northern blotting
Total RNA was isolated with Trizol (Invitrogen). The RNA was ethanol-precipitated and resuspended in 0.1 mM EDTA. For Northern blotting, 5 µg RNA was separated on a 1% agarose gel and transferred to a Hybond-N membrane (Amersham Biosciences) as described [19
]. Filters were prehybridized for 30 min at 42°C in 6 mL ULTRAhyb (Ambion, Austin, TX) and were hybridized overnight at 42°C after addition of further 4 mL containing the 32P-labeled probe and sheared salmon sperm DNA (10 µg/mL). The membranes were washed as described [19
] and developed by a Fuji BAS2500 Phosphor Imager (Fuji, Tokyo, Japan). Membranes were stripped by boiling in 0.1% sodium dodecyl sulfate (SDS) before rehybridization. Probes used for hybridization were radiolabeled with [
-32P]deoxycytidine 5'-triphosphate using the Random Primers DNA labeling system (Invitrogen). Sizes of the hybridizing bands were determined relative to 18S and 28S and were found to be in accordance with previous reports (Table 1
).
|
View this table: [in a new window] |
Table 1. Synthesis of cDNA Probes for Northern Blot Hybridization
|
Preparation of whole-cell lysates and Western blotting
Cells/mL (3x106–7.5x106) were preincubated in 0.7 µM diisopropyl fluorophosphate (DFP) for 5 min prior to lysis in 1 mL Trizol (Invitrogen)/107 cells. Lysates were added to 200 µL chloroform and were centrifuged for 15 min/12,000 g. DNA was removed by adding 300 µL ethanol to the organic phase and centrifuged for 5 min/7500 g. Isopropanol (1 mL) was added to the supernatant, and the lysate was incubated for 10 min at room temperature following centrifugation for 10 min/12,000 g. Precipitated protein was washed three times in 2 mL 0.3 M guanidine HCl/ethanol and once in 2 mL ethanol. Pellet was dissolved in 333 µL boiling 2x Laemmli buffer (0.125 M Tris-base, pH 6.8, 20% v/v glycerol, 4% w/v SDS, bromphenol blue, 2% v/v 2-mercaptoethanol). For Western blotting, 3 x 105 cells/population were run on SDS gels (between 7.5% and 14%, depending on the protein tested) and were transferred onto nitrocellulose membranes (Sartorius, Göttingen, Germany). Immediately after blotting, membranes were incubated in Ponceau S (Sigma-Aldrich, St. Louis, MO) for verification of successful blotting. The immune complexes were visualized by enhanced chemiluminescence (Amersham Biosciences), according to the manufacturer’s description. For the experiments using radio immunoprecipitation assay (RIPA) buffer, cells were lysed in ice-cold RIPA buffer (50 mM Tris-base, pH 7.4, 1% v/v NP-40, 0.25% v/v sodium deoxycholate, 150 mM NaCl, 1 mM EGTA, 1 mM Na3VO4) containing a broad cocktail of protease inhibitors (1836153; Roche, Mannheim, Germany) for 30 min on ice and subsequently centrifuged for 10 min/12,000 g. Supernatants were analyzed on Western blots as described above.
Antibodies and positive controls
Antibodies against CDK2 (sc-163), CDK4 (sc-260), CDK6 (sc-177), cyclin A (sc-751), cyclin D1 (sc-718), p16ink4a (sc-468), p18ink4c (sc-1064), p19ink4d (sc-1063), p57kip2 (sc-1040), p130 (sc-9963), signal transducer and activator of transcription (STAT)3 (sc-482; Santa Cruz Biotechnology, Santa Cruz, CA); -tubulin (T9026; Sigma-Aldrich); cyclin D2 (AHF0122; Biosource, Camarillo, CA); cyclin D3 (610279), p21cip1 (C24420), p27kip1 (K25020), pRb (14001A; BD Biosciences-PharMingen, San Jose, CA); cyclin E (RB-012), p15ink4b (MS-1053), p107 (RB1414; Neomarkers, Fremont, CA); and neutrophil gelatinase-associated lipocalin (NGAL) [22
] were used for Western blot analysis according to the manufacturers’ recommended concentrations. Horseradish peroxidase-linked rabbit anti-mouse immunoglobulin G (IgG; P0260, Dako, Glostrup, Denmark) or swine anti-rabbit IgG (P0217, Dako) was used as a secondary antibody. Whole-cell lysates from the following cell lines were used as positive controls for Western blot analysis: HepG2 (CDK2, CDK4, cyclins A, D1, E, p15ink4b, p27kip1, p57kip2), K562 (CDK6, p18ink4c, p19ink4d), HL60 (cyclin D2), Jurkat (cyclin D3), HeLa (p16ink4a, pRb), MCF-7 (p21cip1, p130), and NIH3T3 (p107).
Nitroblue tetrazolium (NBT) test
NBT test was performed as described previously [23
].
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We initially investigated the proportion of cells in G0/G1 and S/G2 + M phases. MBs/PMs, MCs/MMs, BCs/SCs, and PMNs (Fig. 1 A ) were lysed, and DNA was stained with PI. In the most immature cell fraction (MBs/PMs), 59% of the cells were in G0/G1, and 36% of the cells were in S/G2 + M phases (Fig. 1B and 1C) . The amount of cells in S/G2 + M was significantly reduced in the MC/MM fraction, only 12% of the cells were in S/G2 + M, and the amount of cells in G0/G1 was concordantly increased to 86% (Fig. 1B and 1C) . The two most mature cell populations, BCs/SCs and PMNs, were 99% G0/G1 cells (Fig. 1B and 1C) . None of the populations showed more than 9% apoptotic cells (Fig. 1B) .
 View larger version (39K): [in a new window] |
Figure 1. Morphology and DNA staining of the four neutrophil cell populations. Percoll-separated bone marrow populations: MBs/PMs, MCs/MMs, and BCs/SCs and PMNs from peripheral blood were purified using MACS columns. (A) Cytospin preparations of the four cell populations. Cells were stained with May-Grünwald-Giemsa. (B and C). Populations of 105 cells were lysed in modified Vindeloev buffer, and DNA was stained with PI. The extracts were analyzed by flow cytometry, and the percentages of cells in G0/G1 and S/G2 + M were calculated. The shown diagrams are representative of three independent experiments.
|
-tubulin (Fig. 2B)
. To ensure that the lower amount of protein in the RIPA and Laemmli lysates was not a result of unsuccessful lysis, we also measured the amount of the granule-derived protein NGAL, which is very resistant to degradation by granule-derived proteases [25
]. NGAL was fully recovered after lysis in RIPA, Laemmli, and Trizol buffers (Fig. 2A)
. The active proteases stored in neutrophil granulocytes are predominantly serine proteases [26
]. Thus, the effect of irreversible inhibition of serine proteases with DFP was therefore also tested. Preincubation of PMNs with DFP before lysis clearly improved the recovery of p27kip1, STAT3, and -tubulin by all three methods. As expected, the effect was less prominent in the Trizol/guanidine HCl lysate, as this had the least spontaneous proteolysis (Fig. 2A)
. Thus, for the following experiments, we decided to preincubate the cells with DFP before preparation of whole-cell lysates using Trizol/guanidine HCl.
 View larger version (16K): [in a new window] |
Figure 2. Comparison of three different lysis methods for recovery of proteins in PMNs (3x105), which were lysed in RIPA, 2x Laemmli, or Trizol buffer as described in Materials and Methods. Cells were preincubated with 0.7 µM DFP prior to lysis where indicated. (A) Western blots of STAT3 (a 91-kDa STAT3 form and a 90-kDa STAT3ß form), -tubulin, p27kip1, and NGAL expression. Equal amounts of cells (3x105) were loaded on each lane. RIPA lysate of HepG2 cells is shown as a positive control for successful RIPA lysis. (B) Long-time exposure of the -tubulin blot showing trace amounts of -tubulin in the RIPA lysates.
|
45-kDa form) was slightly up-regulated from the MC/MM stage and onward. Additionally, an 50-kDa form of cyclin E appeared in the more mature populations of neutrophil granulocytes: BCs/SCs and PMNs (Fig. 3A)
. The protein level of cyclin A was strongly decreased already at the MC/MM stage and was not present in BCs/SCs and PMNs (Fig. 3A)
, as expected from the cell-cycle arrest pattern seen in Figure 1
.
 View larger version (29K): [in a new window] |
Figure 3. Protein expression profiles of cyclins, CDKs, and pocket proteins in the four neutrophil populations. MACS-purified neutrophil cell populations (3x105) were pretreated with 0.7 µM DFP prior to lysis in Trizol buffer. Western blots showing protein levels of (A) cyclins D1, D2, D3, E, and A, (B) the G1 phase CDK2, -4, and -6, and (C) the pocket proteins pRb and P-pRb (hyper-phosphorylated), p107, and p130. We have previously observed a decrease of ß-actin and glyceraldehyde-3-phosphatase dehydrogenase in the more mature populations of neutrophils [19
]. Thus, we decided to use STAT3 expression as a loading control, as we found that the expression of this protein was concurrent with the cell number in all neutrophil populations. The shown results are representative of three independent experiments. See Materials and Methods for the positive controls used on the Western blots.
|
To confirm cell-cycle arrest and low activity of CDK2, -4, and -6 in the more mature stages of neutrophil granulocytes, we investigated the phosphorylation status of pRb. In concordance with the decreased fraction of cells in S/G2 + M phases (Fig. 1B and 1C) and the lack of expression of S-phase cyclin A (Fig. 3A) , no hyperphosphorylated pRb could be detected in BCs/SCs and PMNs (Fig. 3C) . Hyperphosphorylation of pRb was most prominent in MBs/PMs and already diminished in the MCs/MMs (Fig. 3C) . The protein level of nonphosphorylated pRb was decreased in BCs/SCs and PMNs, compared with the less mature populations (Fig. 3C) . We next investigated protein expression levels of the two pRb-related pocket proteins p107 and p130, which were present in MBs/PMs but were down-regulated at the MC/MM stage an onward. Although p130 already diminished at the MC/MM stages and was not present in the more mature populations, p107 was down-regulated at a slower rate, allowing the detection of minor amounts of p107 in BCs/SCs and PMNs (Fig. 3C) .
Protein expression profiles of CDKIs during differentiation of neutrophil granulocytes
Up-regulation of CDKIs might also play a role in induction of cell-cycle arrest during terminal differentiation. We therefore investigated the expression pattern of the specific inhibitors of CDK4 and –6, the INK4 CDKIs. Although we were unable to detect p15ink4b and p16ink4a proteins, the protein level of p18ink4c was constant through all stages of differentiating neutrophil granulocytes (Fig. 4 A
). Following prolonged exposure of the p19ink4d Western blot, we were also able to detect a constant level, comparable with the control lysate, of p19ink4d in all four populations (Fig. 4A)
. We next investigated the generic CIP/KIP CDKIs. The protein level of p57kip2 remained unchanged during granulopoiesis, and p21cip1 was undetectable. In sharp contrast, the protein level of p27kip1 clearly increased from the MC/MM stages and onward (Fig. 4B)
. These data are in contrast to those previously reported, where an induction of p21cip1 mRNA and protein was seen following differentiation of the promyelocytic APL cell line NB4 with all-trans retinoic acid (ATRA) [10
11
27
]. To determine whether we could reproduce the induction of p21cip1 under the same assay conditions, we incubated NB4 cells with 1 µM ATRA for 6 days, which induced differentiation of the cells into the granulocytic lineage as determined by morphology, NBT reduction, and down-regulation of CD49d expression (data not shown). Following MACS depletion of not fully differentiated cells still expressing CD49d [28
], protein expression of p27kip1 and p21cip1 was investigated. In concordance with the literature, we too observed an increase of p21cip1 and p27kip1 protein in the differentiated NB4 cells (Fig. 4C)
. This demonstrates differences in cell-cycle regulation between in vivo and in vitro systems. Whereas p21cip1 seems to be important in the NB4 cell system, it is apparently not involved in differentiation of neutrophils in vivo.
 View larger version (20K): [in a new window] |
Figure 4. Protein expression profiles of CDKIs in the four neutrophil populations. Cell lysates were prepared as in Figure 3
. Western blots showing protein levels of (A) the INK4 family of CDKIs, p15ink4b, p16ink4a, p18ink4c, and p19ink4d, and of (B) the CIP/KIP family of CDKIs, p21cip1, p27kip1, and p57kip2. (C) Protein expression levels of p21cip and p27kip1 were assessed on Western blots in undifferentiated NB4 cells (–) and in NB4 cells differentiated with 1 µM all-trans retinoic acid (ATRA) for 6 days (+). STAT3 expression was used as a loading control. The shown results are representative of three independent experiments. See Materials and Methods for the positive controls used on the Western blots.
|
 View larger version (61K): [in a new window] |
Figure 5. Transcript profiles of the CDKs, the pocket proteins, p27kip1, and p21cip1. Northern blots of total RNA from the four MACS purified neutrophil cell populations. The blots were hybridized with probes against (A) CDK2, -4, and -6, (B) pRb, p107, and p130, and (C) p21cip1 and p27kip1. 18S expression was used as a loading control. Total RNA extracted from lung, colon, and brain tissues, and A549, normal human bronchial epithelial (NHBE), normal human epidermal keratinocytes (NHEK), and HL60 cell lines were used as positive controls for p21cip1 mRNA expression.
|
 View larger version (47K): [in a new window] |
Figure 6. Protein from THP-1 cells is fully recovered in PMN lysates. THP-1 cells or PMNs were preincubated with 0.7 µM DFP prior to lysis in Trizol buffer. For the double lysate, equal amounts of THP-1 cells and PMNs were mixed immediately following DFP treatment and were subsequently lysed in Trizol buffer. Lysates from 3 x 105 cells were loaded on lanes 1 (THP-1) and 2 (PMN), and a lysate from 6 x 105 cells was loaded on lane 3 (THP-1+PMN) to ensure that the protein contribution from THP-1 cells was as in lane 1. The Western blots show the recovery of CDK2, -4, and -6, p107, and p130 proteins.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
50-kDa form of cyclin E emerged in the BC/SC and PMN populations. Of the CDKIs, only p27kip1 was extensively up-regulated from the MC/MM stage and onward. It is striking that all three G1-phase CDKs were strongly down-regulated from the MC/MM stages and onward on transcript and protein levels. Down-regulation of CDKs is a common theme in end-stage, differentiated cells [29
30
31
32
]. However, simultaneous down-regulation of all three G1-phase CDKs is not the general picture. Although CDK2 is down-regulated, protein levels of CDK4 and -6 are unchanged in differentiating mouse keratinocytes [32
] and during renal development [31
]. The expression pattern of G1-phase proteins during neuronal differentiation, conversely, resembles the one shown in granulocytic differentiation. Cell-cycle arrest and down-regulation of CDK4 and CDK2 [29
30
] accompanied differentiation of oligodendrocyte progenitor cells in vitro and a concomitant up-regulation of p27kip1 [30]. Combined, these data indicate that the cell-cycle arrest that accompanies end-stage differentiation of neutrophils is executed through two different mechanisms: by turning off the cell-cycle accelerators via down-regulation of all three G1 CDKs and by inducing a generic CDK brake via up-regulation of p27kip1. These events seem to be regulated by changes of transcript levels and possibly expression activity of the cognate genes. Regulation of p27kip1 has previously been shown to occur at the post-translational level [33
34
]. CDK2 complexed with cyclin E is known to down-regulate p27kip1 protein by phosphorylation, which tags p27kip1 for degradation via the ubiquitin proteasome pathway [35
36
37
]. However, during granulocytic differentiation, p27kip1 transcript is also up-regulated, rendering it unlikely that lack of CDK2 alone and thus merely stabilization of p27kip1 protein is causing the increase of p27kip1 protein in BCs/SCs and PMNs. Oligodendrocyte progenitor cells obtained from p27kip1-/- mice differentiate poorly [29
]. Likewise, p27kip1-/- mice produce a reduced number of mature neutrophils [16
]. These experiments indicate that p27kip1 deficiency alone is not sufficient to block cell-cycle arrest and differentiation. Speculatively, the concomitant down-regulation of the CDKs and up-regulation of p27kip1 might be two differently regulated mechanisms to induce cell-cycle stop. If one of these fails, the other will compensate to some degree, ensuring that the neutrophil is able to differentiate. This hypothesis has been tested in vitro, as inhibition of CDK2 restores the ability of Mad3/p27kip1 double-null granulocyte cell lines to differentiate in response to ATRA [16
]. The existence of such back-up systems might explain why a single mutation often is insufficient to cause the formation of a malignant myeloid clone [38
39
].
 View larger version (18K): [in a new window] |
Figure 7. Hypothetical distribution scheme of neutrophil cell-cycle proteins during granulopoiesis. Hypothetical distribution scheme of cell cycle proteins during granulopoiesis based on the data presented in Figures 3
and 4
.
|
When aligning the protein expression profiles of cell-cycle proteins during granulopoiesis (Fig. 7) , it becomes clear that the major alterations take place at the MC/MM stage. Also, this is the point where the proliferative myelocyte turns into a nonproliferative metamyelocyte [1 ]. Hypothetically, several events need to be coordinated in the myelocyte to propagate to the metamyelocyte stage. The transcription factor c-myc is a likely candidate to control these events, as c-myc is rapidly down-regulated during granulocytic differentiation and disappears at the MM/MC stage [16 ] (unpublished). Concordantly, it is well known that c-myc down-regulates p27kip1 by suppressing its transcription and stability [46 47 48 ] and up-regulates CDK4 and cyclin D2 [49 50 ]. Other transcription factors, which may be involved in cell-cycle regulation during granulopoiesis are members of the CCAAT/enhancer-binding protein (C/EBP) family. We have recently described an individual expression pattern of the six C/EBP members during granulopoiesis [19 ], and a role in cell-cycle regulation and granulocytic differentiation has been ascribed to many of these proteins [51 ]. The forkhead transcription factor Foxo3a has also been shown to regulate transcription of the p27kip1 gene and thereby induce cell-cycle arrest [52 53 54 ]. However, Kops et. al. [53] showed that in human colon carcinoma cells, conditional activation of Foxo3a increases p27kip1 and p130 protein expression. Our finding that p130 is down-regulated during granulopoiesis indicates that p130 is not regulated by Foxo3a in granulocytes, or Foxo3a is not involved in regulating cell-cycle arrest during granulopoiesis. Another likely candidate is STAT3, which also has been shown to regulate p27kip1 on a transcriptional level [55 56 ]. Granulocyte-colony stimulating factor is a well-known inducer of granulocytic differentiation, and this has been shown to be followed by a strong STAT3 activation and up-regulation of p27kip1 [55 ].
Cell-cycle regulation during myeloid differentiation has been studied extensively in vitro. However, two major problems exist when comparing the in vitro models and the presented in vivo data: the mechanisms by which cell-cycle arrest is executed using various agents and differences between model cell lines and neutrophil precursor cells. ATRA is a commonly used agent to differentiate granulocytes in vitro [11 27 ]. Furthermore, ATRA is used to exert end-stage differentiation of promyelocytes in APL patients [57 58 ]. No concise picture is evident regarding which cell-cycle proteins are influenced by ATRA in model cell systems. However, it seems to be that ATRA is acting on the brakes, especially the INK4s and p21cip1 [7 10 11 59 ] (Fig. 4C) and through down-regulation of cyclins [9 60 ] rather than influencing CDK levels [12 60 ]. This resembles more the pattern seen in the reversible withdrawal of cells from the cell cycle, which is regulated by cytokines [4 ], than the irreversible cell-cycle stop seen, e.g., in terminally differentiated neurons [61 ]. The second problem lies in the model cell systems used. As mentioned in Introduction, the human cell lines commonly used as models for granulocytic differentiation are unable to express matrix proteins of peroxidase-negative granules [17 18 ]. Expression of the genes encoding matrix proteins of specific granules (such as lactoferrin, NGAL, and hCAP18) is coordinately regulated and initiated in the myelocyte [62 ]. The observation that human myeloid cell lines such as NB4 and HL60 are unable to express these genes [17 18 62 ] indicates that a major transcriptional regulator(s), acting on the myelocyte stage, is missing or is nonfunctional in these cells. As transcription of the genes encoding specific granule matrix proteins occurs at the same differentiation stage as the up-regulation of p27kip1 mRNA and down-regulation of CDK2, -4, and -6 mRNA are observed in vivo, a concern is that the transcriptional regulation of the genes encoding the cell-cycle regulatory proteins also is corrupted in the model cell lines. One difference in the mechanisms by which cell-cycle arrest, e.g., of NB4 cells and human bone marrow neutrophils, is executed may be the role of p21cip1. In concordance with the literature, we detected up-regulation of p21cip1 protein in NB4 cells during differentiation, and neither p21cip1 mRNA nor protein was present in any of the neutrophil populations. Thus, ATRA-induced differentiation of human neutrophil cell lines does not faithfully reflect granulopoiesis in vivo.
In conclusion, cell-cycle arrest during granulocytic differentiation in vivo is mediated by a coordinated down-regulation of the G1 CDKs and pocket proteins and up-regulation of p27kip1. As regulation of p27kip1 and the CDKs appears to be controlled at the transcriptional level, delineation of the regulation of CDK and p27kip1 promoters in granulocytes is likely to give valuable information of how cell-cycle arrest is exerted during differentiation, which in turn might give rise to development of new therapeutic drugs.
Received October 13, 2003; revised November 14, 2003; accepted November 19, 2003.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
protein is required prior to inhibition of histone deacetylase for overcoming all-trans retinoic acid differentiation resistance in acute promyelocytic leukemia Blood 100,1008-1013 Mol. Cell Biol. Res. Commun. 1,125-131[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  C. D. Coldren, J. A. Nick, K. R. Poch, M. D. Woolum, B. W. Fouty, J. M. O'Brien, M. P. Gruber, M. R. Zamora, D. Svetkauskaite, D. A. Richter, et al. Functional and genomic changes induced by alveolar transmigration in human neutrophils Am J Physiol Lung Cell Mol Physiol, December 1, 2006; 291(6): L1267 - L1276. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Theilgaard-Monch, L. C. Jacobsen, M. J. Nielsen, T. Rasmussen, L. Udby, M. Gharib, P. D. Arkwright, A. F. Gombart, J. Calafat, S. K. Moestrup, et al. Haptoglobin is synthesized during granulocyte differentiation, stored in specific granules, and released by neutrophils in response to activation Blood, July 1, 2006; 108(1): 353 - 361. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
