pCip Is Required for the Development of Monocytes and Their Response to Serum Transfer-induced Arthritis

【摘要】  One of the central functions of cyclin-dependent kinase inhibitors, such as p21, p27, or p16, is to prevent entry into the cell cycle. However, the question remains as to whether they have other functions in the cell. We previously demonstrated that overexpression of p21 in fibroblasts isolated from patients with rheumatoid arthritis decreases the production of pro-inflammatory molecules. Overexpression of p21 has been also shown to reduce the development of experimental arthritis in mice and rats. To explore the role of endogenous p21 in the development of arthritis, we induced arthritis in p21C/C mice using the K/BxN serum transfer model of arthritis. Mice deficient in p21 were more resistant to serum transfer-induced arthritis (K/BxN) than wild-type (wt) control mice. Fewer macrophages were detected in p21C/C as compared to wt joints following transfer of K/BxN serum. Chemotaxis assays of bone marrow-derived macrophages from p21C/C and wt mice revealed no difference in migration. However, there was a substantial decrease in inflammatory monocytes circulating in peripheral blood and in monocyte precursors in bone marrow of p21C/C mice as compared to wt mice. Adoptive transfer of wt bone marrow-derived macrophages into p21C/C mice restored the sensitivity to serum transfer-induced arthritis. These data suggest a novel role for p21 in regulating the development and/or differentiation of monocytic populations that are crucial for the induction of inflammatory arthritis.
--------------------------------------------------------------------------------
Proper regulation of the mammalian cell cycle is vital for cellular homeostasis, and alterations in the cell-cycle components may lead to pathogenesis. Progression through the cell cycle is dependent on the activities of cyclin-dependent kinases (cdks) bound to their cognate cyclins.1 Another level of cell-cycle regulation is imposed by the cdk inhibitors. cdk inhibitors bind to cdk or cdk-cyclin complexes and inhibit their kinase activity. The cdk inhibitors are grouped into two categories based on homology and preferential binding to cdk-cyclins (Inks = p15, p16, p18, and p19, and Cip/Kip = p21, p27, and p57). Overexpression of any of the cdk inhibitors will induce G1-cell cycle arrest.2 p21 may have additional roles in regulating transcription, kinase activity, and apoptosis dependent on and/or independent of its ability to inhibit the cell cycle. p21 has been shown to inhibit c-Myc,3 c-Jun NH2-terminal kinase,4 and signal transducer and activator of transcription-3.5 Furthermore, deficiency in p21 has been shown to be associated with enhanced or reduced oncogenesis, atherosclerosis, or lupus-like disease.6-10
Rheumatoid arthritis (RA) is a chronic inflammatory and destructive arthropathy of unknown etiology.11 During the pathogenesis of RA, highly activated monocytes/macrophages are directly involved in synovial inflammation and destruction of cartilage and bone, such that their number correlates with articular destruction.12,13 Further, macrophages are required for collagen-induced arthritis and interleukin-1 (IL-1)/methylated bovine serum albumin-induced arthritis.14,15 Although macrophages are unlikely to be the initiators of RA, the increase in the number of macrophages and the enhanced activation of macrophages in the joint indicate that monocytes/macrophages are one of the principal effector cell types in RA. Macrophages are one of the central producers of IL-1ß and tumor necrosis factor , two essential pro-inflammatory cytokines required for the progression of RA. IL-1ß and tumor necrosis factor , in turn, are capable of inducing other pro-inflammatory cytokines and activating matrix metalloproteinases in autocrine and paracrine fashions,16 leading to increased joint destruction. Inhibition of IL-1ß and tumor necrosis factor activity suppresses synovial inflammation and bone destruction in RA patients.17,18 Although macrophages are vital to the pathogenesis of RA, few studies have examined the factors that regulate their development during normal growth and during the induction and development of inflammatory arthritis.
Recently, the role of p21 in the pathogenesis of RA has been investigated. The expression of p21 is reduced in RA as compared to normal or osteoarthritis synovial tissue, particularly in the synovial fibroblast population.19 Overexpression of p21 inhibits the progression of the cell cycle and the constitutive and IL-1ß-induced production of cytokines, chemokines, and matrix metalloproteinases in synovial fibroblasts isolated from patients with RA.19-21 Moreover, articular injection of replication defective adenoviruses engineered to overexpress p21 prevents the development of experimental arthritis in mice and rats.20,21 These data suggest that p21 functions to suppress the development of arthritis.
Here, we demonstrate that, contrary to our prediction, p21 is required to sensitize mice to inflammatory arthritis following the transfer of K/BxN serum. p21C/C mice failed to display ankle swelling, which is a physical characteristic of inflammatory arthritis, and showed decreased histological scores of arthritis as compared to wild-type (wt) mice. In contrast, p27-deficient mice developed arthritis equivalent to wt mice. Fewer macrophages were detected in the synovium of p21C/C joints compared to wt mice following serum transfer. However, no differences were seen in the number of macrophages in the joints of p27C/C and wt mice. Additionally, a deficiency in recruitment of monocytes to the peritoneal cavity was also seen in p21C/C compared to wt mice following thioglycollate stimulation. Moreover, p21C/C but not p27C/C mice had a marked decrease in circulating inflammatory monocytes as compared to controls. The reduction of inflammatory monocytes in circulation was associated with fewer monocyte precursor cells in bone marrow. These data suggest that p21 is required for differentiation of monocytes in bone marrow and that these monocytes are necessary for the response to K/BxN serum.

【关键词】  required development monocytes response transfer-induced arthritis

Materials and Methods

Mice

B6;129S2-Cdkn1atm1Tyj/J (p21C/C), B6129SF2/J (congenic control for p21C/C mice), B6.129S4-Cdkn1btm1Mlf/J (p27C/C), and C57BL/6 (congenic control for p27C/C mice) mice were purchased from the Jackson Laboratory (Bar Harbor, ME). The non-obese diabetic mice were purchased from Taconic (Germantown, NY), and the homozygous KRN TCR transgenic mice (C57BL/6 background) were a kind gift from Drs. D. Mathis and C. Benoist (Harvard Medical School, Boston, MA, and the Institute de Gene-tique et de Biologie Moleculaire et Cellulaire, Strasbourg, France). All experiments on mice were approved by the Animal Care and Use Committee at St. Louis University.

Serum Transfer-Induced Arthritis

The K/BxN serum transfer model shares many features with human inflammatory arthritis, including leukocyte invasion, synovial lining hyperplasia, pannus formation, and cartilage and bone erosion.22,23 Progeny from mice transgenic for a TCR (KRN) that recognizes a specific peptide from bovine ribonuclease crossed with non-obese diabetic mice spontaneously develop arthritis.22 The KRN TCR also recognizes the self-peptide, glucose-6-phosphate isomerase (GPI) in the context of Ag7 major histocompatibility complex class II molecule.24 Transfer of serum or purified immunoglobulin (Ig)G from K/BxN mice leads to induction of a robust and reproducible acute disease in several mouse strains.23 In these studies 7-week-old progeny from KRN mice crossed with non-obese diabetic mice (K/BxN) were euthanized, peripheral blood was isolated via cardiac stick, and serum was collected via centrifugation and pooled. K/BxN serum (150 µl) was intraperitoneally injected on each flank of 6-week-old wt, p21C/C, or p27C/C mice. As a control, mice were treated with saline in place of serum (data not shown). At each time point and before euthanasia, the degree of arthritis, as indicated by joint swelling, was quantitated by measuring two perpendicular diameters of the ankles using a caliper (Lange Caliper, Cambridge Scientific Industries, Cambridge, MA). Joint circumference was calculated using the geometric formula of ellipse circumference (2 x (a2 + b2)/2), as previously described.25 Ankle joints were removed and either fixed in 10% neutral buffered formalin for 24 hours and then in ethylenediamine tetraacetic acid-decalcification buffer for 2 weeks, embedded in paraffin, and sectioned or placed in liquid nitrogen, ground into a fine powder by mortal and pestle, digested in protein lysis buffer, and homogenized on ice for 20 seconds.26

Immunophenotyping

Peripheral blood and bone marrow were isolated from 5- to 8-week-old mice following euthanasia. Peritoneal cells were isolated by peritoneal lavage from unstimulated mice and from mice 1 and 5 days after intraperitoneal injection of 4% aged thioglycollate. Total leukocyte numbers were determined using an automated hematology analyzer ABX Pentra 60 (Diamond Diagnostics, Inc., Holliston, MA). The red blood cells in peripheral blood were lysed, and the remaining cells were fixed with BD FACS lysing solution (BD Pharmingen, San Diego, CA) following incubation with antibodies. Cells were first incubated with Fc Block (BD Pharmingen) and then stained with fluorochrome-conjugated antibodies specific to CD45, CD11b, Gr-1, CD62L, CD16/32, Ly-6C, Ly-6G, or CD31 (BD Pharmingen) or isotype control antibodies for 30 minutes. Bone marrow cells were isolated by flushing Dulbecco??s modified Eagle??s medium (Gibco, Grand Island, NY) through the tibias, and red blood cells were lysed with BD PharM Lyse (BD Pharmingen) before incubation with antibodies. Cells were acquired on a BD FACSCalibur (BD Pharmingen) using CellQuest Software or BD FACSAria using Diva Software at the St. Louis University Core Flow Cytometry Facility. All analyses were performed using FlowJo software (Tree Star Inc., Ashland, OR).

Immunohistochemistry and Histopathological Scoring

Following euthanasia, mouse legs were isolated, decalcified with ethylenediamine tetraacetic acid (Sigma-Aldrich, St. Louis, MO) in 10% formalin, embedded in paraffin, and sectioned. To stain for F4/80, antigens were retrieved using the Dako target retrieval solution (Dako, Glostrup, Denmark). Following antigen retrieval, sections were blocked in hydrogen peroxide, incubated with anti-F4/80 antibody (Clone BM8; Caltag Laboratories, Burlingame, CA) or isotype control, and then incubated with secondary biotinylated rabbit anti-rat antibody (Dako). Sections were treated with streptavidin peroxidase conjugate (Dako), color was visualized with diaminobenzidine, and sections were counterstained with hematoxylin. All F4/80 staining was performed on a Dako autostainer. Ankle sections were also stained with hematoxylin and eosin (H&E) or Safranin O and methyl green. Histopathological scoring was performed as previously described in detail.27 A pathologist blinded to the study (GKH) evaluated ankle sections by examining at least three sections/ankle and three fields/section at magnification x400. H&E ankle sections were scored on a scale of 0 to 5 for inflammation, with 0 = normal, 1 = minimal infiltration, 2 = mild infiltration, 3 = moderate infiltration, 4 = marked infiltration, and 5 = severe infiltration. Bone erosion was scored on a scale of 0 to 5 by viewing H&E ankle sections, with 0 = no or normal bone resorption, 1 = small areas of resorption, 2 = more numerous areas of resorption, 3 = obvious resorption, 4 = full thickness defects in the bone without distortion of the profile, and 5 = full thickness defects in the bone with distortion of the profile. H&E ankle sections were scored on a scale of 0 to 5 for pannus formation, with 0 = no pannus formation, 1 = minimal pannus formation, 2 = mild pannus formation, 3 = moderate pannus formation, 4 = marked pannus formation, and 5 = severe pannus formation. H&E and Safranin O and methyl green sections were scored on a scale of 0 to 3 for cartilage damage, with 0 = no damage, 1 = superficial cartilage destruction, 2 = cartilage destruction to middle zone, and 3 = cartilage destruction to tide mark. Three fields of representative pannus and synovium stained with anti-F4/80 were viewed under oil emersion at magnification x1000, and the number of F4/80-positive cells was counted. The number of F4/80-positive cells from the six fields was averaged and compared to counts from untreated mice to determine -fold increase over the time course. Areas of pannus were defined as synovial-looking tissue adjacent to the area of bone erosion and proliferation of synovial-type tissue outside of joint space. Ankle joint sections that did not have areas of pannus were not counted and were not included in the average. Synovium was defined as synovial lining cells and subadjacent tissue, including granulation tissue outside the joint space. Histopathological scoring was conducted by using an Olympus BS40 microscope (Olympus, Tokyo, Japan). Photographs were taken on a Nikon microscope equipped with the Nikon digital camera DMX1200 (Nikon, Tokyo, Japan).

Enzyme-linked Immunosorbent Assay

For detection of GPI autoantibodies in the serum and ankles, sandwich enzyme-linked immunosorbent assays were performed as previously described.28 Plates (96-well) were coated with GPI type IV and XI (Sigma-Aldrich) at 10 µg/ml. Serum was diluted in phosphate-buffered saline and incubated overnight. Goat anti-mouse Ig-AP (SouthernBiotech, Birmingham, AL) was used as a detection antibody. Plates were developed with Alkaline Phosphate Yellow (Sigma-Aldrich), and absorption was read at 415 nm on a Microplate reader (Bio-Rad, Hercules, CA). Following euthanasia, serum was isolated from peripheral blood following cardiac stick, and ankles were isolated at various times from the mice following serum transfer. Ankles were snap frozen, ground with a mortar and pedestal, and then homogenized in lysis buffer (150 µmol/L NaCl, 0.5% Nonidet P-40, 50 mmol/L Tris, 2 mmol/L ethylenediamine tetraacetic acid, pH 8.0, with protease inhibitors and phosphatase inhibitors).26 All ankle data were normalized by total protein concentration (micrograms/microliter) of the ankle extracts.

Chemotaxis Assays

Chemotaxis assays were performed with minor modifications to a protocol previously optimized for RA synoviocyte chemotaxis.29 Bone marrow cells were isolated from p21-deficient and wt mice by tibia flush. Following tibia flush, bone marrow-derived macrophages (BMDMs) were grown in media containing 20% L-cell supernatant containing macrophage-colony stimulating factor, 10% fetal bovine serum (HyClone, Logan, UT), 5% equine serum (HyClone), 1% L-glutamine (200 mmol/L, Gibco), 1% penicillin-streptomycin (Gibco), and 1% sodium pyruvate (100 mmol/L, Cambrex, Walkersville, MD) for 7 days, then grown in the same media, but with only 10% L-cell supernatant. Cells were passage when they reached confluency and used between passages 3 and 6. Fluorescence-activated cell sorting (FACS) analysis revealed the cells were F4/80+, CD45+, CD11b+, and FC receptor+ (CD16/32). wt and p21C/C BMDMs were fed the night before the assay with full growth media for consistency. Chemoattractants or media without stimulants were added to the bottom wells of a 48-well microchemotaxis chamber (Neuroprobe, Gaithersburg, MD), and wells were covered with a 5-µm polycarbonate membrane. Chemoattractants included fractalkine (Fkn/CX3CL1, R&D Systems, Minneapolis, MN), leukotactin (Lkn/CCL15, PeproTech, Rocky Hill, NJ), the bacterial peptide formyl-methionyl-leucyl-phenlyalanine (fMLP, Sigma-Aldrich), and macrophage-colony stimulating factor (M-CSF, derived from dilutions of supernatant from L cells (LCS)). Subconfluent BMDMs were removed with Accutase (eBiosciences, San Diego, CA) and added to the top half of the well (40 µl at 5.0 x 105 cells/ml) in Dulbecco??s modified Eagle??s medium containing 0.1% equine and bovine growth serum. The chambers were incubated for 5 hours in a 5% CO2/95% air atmosphere at 37??C, allowing for cell migration. After migration, non-migrated cells were removed with a cotton swab, and cells that migrated to the opposite side of the membrane were fixed in methanol and stained with Diff-Quik (Dade Behring, Deerfield, IL). Each condition was assayed using eight replicates, and the sum of migrated cells was determined from three randomly selected high power fields per replicate.

Adoptive Transfer of BMDMs

BMDMs were isolated by tibia flush from wt mice and grown in culture conditions described above. 106 BMDMs were injected intravenously into tail veins of p21C/C mice 1 day before serum transfer, 1 day after serum transfer, and 3 days after serum transfer. K/BxN serum was administered to mice, and ankle swelling was measured as the increase in ankle circumference over a 7-day period following transfer of K/BxN serum.

Statistical Analysis

Results were expressed as the mean ?? SE. Differences between groups were analyzed using Student??s t-test.

Results

p21-Deficient Mice Are Less Sensitive to Serum Transfer-induced Arthritis

To determine the role of the cell-cycle inhibitory proteins p21 and p27 in the effector phase of arthritis, we used the serum-transfer model of inflammatory arthritis (K/BxN). p21C/C, p27C/C, and congenic control (wt) mice were injected with a single dose of K/BxN serum at a concentration similar to previous studies.30 p21-deficient mice failed to display physical inflammation as indicated by minimal ankle swelling (Figure 1A) . In contrast to p21C/C mice, p27C/C developed arthritis similar to wt (C57BL/6) mice (Figure 1B) . These data suggest that the inability of p21-deficient mice to develop inflammatory arthritis is unique to p21 and is not due to a general defect in Cip/Kip family members. wt mice demonstrated an increased ankle circumference as early as 3 days after serum transfer, peaking at day 6 and beginning to resolve by days 7 to 10. There was an eightfold increase (P < 0.001) at day 3, a fourfold increase (P < 0.0001) at day 5, a sixfold increase (P < 0.0001) at day 6, and an 82-fold increase (P < 0.0001) at day 7 in the circumference of the ankles from wt as compared to p21C/C mice (Figure 1A) . No statistical difference was observed in ankle circumferences of p21C/C mice between days 6 and 10 (data not shown).

Figure 1. p21-deficient mice do not develop ankle inflammation associated with experimental arthritis. A: Quantitative analysis of the differential ankle circumference in p21C/C mice. 300 µl of pooled serum from K/BxN mice was injected intraperitoneally into wt and p21C/C mice. Ankle joints were examined for arthritis by measuring two perpendicular diameters of both joints (anteroposterior; mediolateral) by calipers. The change in ankle circumference at each time point is defined as the difference between the ankle circumference and the measurement at day 0. wt (n = 50) and p21C/C (n = 60) mice have been examined in multiple independent experiments. Shown are the composite data from those experiments (p21C/C: day 3, n = 29 mice; day 5, n = 20 mice; day 6, n = 21 mice; day 7, n = 17 mice; wt: day 3, n = 20 mice; day 5, n = 22 mice; day 6, n = 17 mice; day 7, n = 23 mice). The values represent the mean ?? SE of ankles/time point, which were compared by Student??s t-test to wt mice under parallel conditions. **, P < 0.001 as compared to wt mice under parallel conditions. B: Quantitative analysis of the differential ankle circumference in p27C/C mice. wt and p27C/C mice were injected with K/BxN serum, and ankle joints were examined for arthritis as described above. Shown are the composite data from those experiments (p27C/C: day 3, n = 10 mice; day 5, n = 10 mice; day 7, n = 14 mice; wt: day 3, n = 10 mice; day 5, n = 10 mice; day 7, n = 16 mice). The values represent the mean ?? SE of ankles/time point, which were compared by Student??s t-test.

p21-Deficient Mice Demonstrate No Difference in Trafficking K/BxN Serum into Ankle Joints

Because increased clearance of GPI autoantibodies within peripheral blood31 or a failure to traffic the autoantibody to the joint would also explain resistance to serum transfer-induced arthritis in p21C/C mice,32 serum and ankles joints were isolated from wt and p21C/C mice following transfer of K/BxN serum. Quantitative analysis of GPI autoantibodies showed no difference in the levels of GPI autoantibodies in peripheral blood (Table 1) or in ankles (Table 1) between genotypes at any of the time points examined. These data indicate that the inability to develop arthritis in p21C/C mice is not mediated by a defect in trafficking of the autoantibody or a decrease in the half-life of the autoantibody.

Table 1. Similar Levels of GPI Autoantibodies Are Found in the Serum and Ankles of p21-Deficient Mice following Injection of K/BxN Serum

p21C/C Mice Display Reduced Histological Scores as Compared to wt Mice

To accurately assess the degree of inflammation and destruction of cartilage and bone, ankle sections were examined using a histopathological scoring system.27 Examination of H&E ankle sections revealed no statistical difference in pannus formation, cellular inflammation, average or median synovial lining, bone erosion, or cartilage destruction in untreated wt or p21C/C mice. In contrast, there was an increase in acute inflammation and soft tissue damage in wt ankles as compared to p21C/C ankles over the 10-day period following injection of K/BxN serum. At 4 days after transfer of serum there was a twofold, a threefold, and a threefold increase in pannus formation (P < 0.002), cellular inflammation (P < 0.001), and bone erosion (P < 0.001) in wt mice as compared to p21C/C mice. Similar results were seen at days 5 and 6; namely, a fivefold increase in pannus formation (P < 0.001), twofold increase in cellular inflammation (P < 0.002), and fourfold increase in bone erosion (P < 0.008) in wt as compared to p21C/C ankle sections (Figures 2 and 3) . However, the increase in bone destruction in wt mice was independent of the number of osteoclasts, because there were no differences in the number or location of TRAP-positive cells in wt and p21C/C mice (data not shown). An increase in cartilage destruction was also observed at days 4, 6 (21-fold increase, P < 0.003), and 10 after serum transfer. There was a marginal increase in the average number of cell layers in the synovial lining at days 4 (1.2-fold, P < 0.02) and 6 (1.3-fold, P < 0.01) in wt as compared to p21C/C mice (Figure 2E) . By days 7 to 10 there were no differences in histological scores between wt and p21C/C mice, which may be due to the fact that wt mice start to resolve the arthritis after day 6. In contrast to p21C/C mice, wt and p27-deficient mice displayed equivalent histological scores (supplemental data, see http://ajp.amjpathol.org). These data suggest that, contrary to the studies that showed overexpression of p21 prevents the development of arthritis,20,21,33,34 the endogenous level of p21 is essential for the full development of arthritis.

Figure 2. Increased histological scores of ankle sections in wt mice compared to p21C/C mice. Ankles isolated from p21C/C and wt mice as described in Figure 1 were fixed, embedded in paraffin, sectioned, and stained with H&E or Safranin O and methyl green. Ankle sections were evaluated and scored by a pathologist blinded to the study as described in Materials and Methods. Shown are the composite data from two experiments (p21C/C: day 0, n = 14 ankles; days 3, 5, and 7, n = 8 ankles; days 4, 6, and 10, n = 16 ankles; wt: days 0, 3, and 5, n = 8 ankles; days 4, 6, and 10, n = 12 ankles; day 7, n = 6 ankles). The values represent the mean ?? SE of ankles/time point, which were compared by Student??s t-test to wt mice under parallel conditions. *, P < 0.03 and **, P < 0.002 as compared to wt mice under parallel conditions.

p21C/C Mice Display Reduced Numbers of Infiltrating Macrophages in the Joint

The infiltrating cells in wt and p21C/C ankles consisted of mainly lymphocytes, granulocytes, and macrophages. However, p21C/C mice had less neutrophils and polymorphonuclear cells compared to wt mice (data not shown). The reduced numbers of infiltrating cells in the p21C/C joint were not mediated by enhanced apoptosis in the joint, because there was no difference in number or location of terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling-positive cells between wt and p21C/C mice (data not shown). Sections stained with F4/80, a marker of macrophages,35,36 revealed that p21-deficient mice had significantly fewer macrophages that were recruited to the joint following K/BxN serum transfer (Figure 4) . There were five- and threefold increase in macrophages found in wt mice at days 5 and 6 after serum transfer as compared to untreated joints. However in p21C/C ankle joints, the increase in recruited macrophages peaked at 1.5-fold over a 10-day time course. The recruitment of macrophages in wt joints correlated with the level of inflammation and ankle circumference. These data suggest that inflammation and joint destruction seen in p21C/C ankle joints following K/BxN serum transfer may be due to lack of macrophage recruitment.

Figure 4. Fewer macrophages are recruited to p21C/C joints following injection of K/BxN serum. A: Quantitative fold increase of macrophage recruitment in ankle sections. Ankles from mice as described in Figure 1 were harvested, sectioned, and stained with antibodies that recognize the F4/80 antigen (brown staining) or isotype control (data not shown) and counterstained with hematoxylin. Macrophage counts were determined by a pathologist blinded to the study as described in Materials and Methods. Shown are the composite data from two experiments (p21C/C: day 0, n = 14 ankles; days 3, 5, and 7, n = 8 ankles; days 4, 6, and 10, n = 16 ankles; wt: days 0, 3, and 5, n = 8 ankles; days 4, 6, and 10, n = 12 ankles; day 7, n = 6 ankles). The values represent the mean ?? SE of ankles/time point. The fold increase was defined as increase over untreated conditions. The data are representative of three independently performed experiments. *, P < 0.004; **, P < 0.0001 as compared to wt mice under parallel conditions. B: Photomicrographs (200x) of ankle sections from wt and p21C/C mice stained for F4/80 antigen. Ankles from mice as described in Figure 1 were harvested, sectioned, and stained with antibodies that recognize the F4/80 antigen (brown staining) or isotype control (data not shown) and counterstained with hematoxylin. SL, synovial lining; B, bone.

p21C/C Mice Have a Defect in Recruitment of Macrophages to the Inflamed Peritoneum

To determine whether the decrease in macrophages observed in joints of p21C/C mice following induction of arthritis occurs in other models of acute inflammation, we intraperitoneally injected wt and p21C/C mice with 4% aged thioglycollate. At 1 day following thioglycollate stimulation there was no difference in the fold increase in neutrophils or macrophages between wt and p21C/C mice (Figure 5, A and B) . However, p21C/C mice displayed a marked reduction in the fold increase in macrophages but not neutrophils as compared to wt mice at day three (Figure 5, A and B) . To determine whether the differences in the two models of acute inflammation were the result of a defect in macrophage migration, chemotaxis assays were performed using BMDMs from wt and p21C/C mice and several known chemoattractants, including fractalkine (CX3CL1, Fkn), leukotactin (CCL15, Lkn), fMLP, or M-CSF, which is contained in LCS. There was no significant differences in the migration of BMDMs derived from wt and p21C/C mice in response to Fkn (0.001 to 10 nmol/L), Lkn (1 nmol/L), fMLP (10 nmol/L), or M-CSF (LCS) (Figure 5C) . Taken together, these data suggest the reduction in recruitment of monocytes to the inflamed joint or to the peritoneum is not attributed to a defect in migration of macrophages.

Figure 5. Reduced migration of monocytes to the inflamed peritoneum of p21C/C mice. p21C/C and wt mice were given intraperitoneal injections of 4% thioglycollate. Peritoneal cells were isolated 0 (n > 10), 1 (n > 8), and 3 (n > 12) days after injection and examined by flow cytometry as described in Materials and Methods. A: The fold increase of neutrophils recruited. Mice were treated as described above, neutrophils (CD45+Gr-1+) were identified using FACS analysis. B: The fold increase of macrophages recruited. Mice were treated as described above. Macrophages (CD45+F4/80+) were identified using FACS analysis. The data were a combination of four independently performed experiments. The values represent the mean ?? SE of cells/time point, which were compared by Student??s t-test. The fold increase was defined as increase over untreated conditions. C: wt and p21C/C BMDMs respond similarly to chemotactic signals. Equal numbers of wt and p21C/C BMDMs were introduced to various chemoattractants to induce migration. Cell counts were expressed as -fold increase over background migration (no stimulus). Fkn, fractalkine/CX3CL1; Lkn, leukotactin/CCL15; fMLP, formyl-methionyl-leucyl-phenlyalanine; LCS, L-cell supernatant containing macrophage-colony stimulating factor.

p21C/C Mice Have a Significant Decrease in Inflammatory Monocytes in Circulation and in Monocyte Precursors in Bone Marrow

Because there is a severe reduction in the number of macrophages in p21C/C joints following induction of arthritis, we examined whether there is a deficiency in monocytes that circulate in blood. wt and p21C/C mice had equivalent number of leukocytes in peripheral blood (data not shown). As recently reported, monocyte populations may be subdivided into two categories: resident monocytes (CD45+CD11b++Gr-1CCD62LC) and inflammatory monocytes (CD45+CD11b++Gr-1+CD62L++).26,29,37 As compared to wt mice, p21-deficent mice displayed a marked reduction (3.4-fold; P < 0.001) in inflammatory monocytes (Figure 6A) . The reduction in inflammatory monocytes was not attributed to increased death, because there was no difference in the percent apoptotic (annexin V-positive) monocytes in peripheral blood (data not shown). The reduction in inflammatory monocytes was specific to p21, because no differences were observed in the monocytic or neutrophil populations in mice lacking p27 (supplemental data, see http://ajp.amjpathol.org). These data suggest that the lack of infiltrating monocytes into the inflamed joint of p21C/C mice may be mediated by a severe reduction in the circulating pool of inflammatory monocytes.

Figure 6. A: Immunophenotyping of peripheral blood reveals a deficiency in inflammatory monocytes in p21C/C mice. Peripheral blood was isolated by cardiac sticks from 5 to 8 week old wt (n = 23) and p21C/C (n = 16) mice. Cells were blocked for 10 minutes with Fc block and then stained with anti-CD45, anti-CD11b, anti-Gr-1, and anti-CD62L antibodies for 30 minutes. Following incubation with antibodies red blood cells were lysed and cells were fixed with BD Biosciences FACS lysing solution. Resident monocytes: CD45+CD11b++Gr-1CCD62LC; inflammatory monocytes: CD45+CD11b++Gr-1+CD62L++; neutrophils: CD45+CD11b++Gr-1++CD62L++. Values represent the mean ?? SE, which were compared by Student??s t-test. *, P < 0.001 as compared to control under parallel conditions. B: Immunophenotyping of bone marrow reveals a deficiency in monocyte precursor populations in p21C/C mice. Bone marrow cells were isolated by tibia flush from 5 to 8 week old wt (n = 18) and p21C/C (n = 13) mice. Red blood cells were lysed, and the remaining cells were blocked for 10 minutes with Fc block and then stained with anti-CD45, anti-CD11b, anti-Gr-1, anti-CD31, anti-Ly6C, and anti-Ly-6G antibodies. Values represent the mean ?? SE, which were compared by Student??s t-test. *, P < 0.002 as compared to wt under parallel conditions.

To determine whether the reduction of monocytes in p21C/C mice is a result of a deficiency in bone marrow, flow cytometric analyses were performed on isolated bone marrow from wt and p21C/C mice. Based on previous studies, the myeloid populations in bone marrow may be separated into different subpopulations based on surface expression of CD45, CD31 (ER-MP12 and PECAM), CD11b, Ly6C (ER-MP20), and Ly-6G.38,39 Further, cell-sorting analyses followed by culturing in media containing M-CSF showed that the CD11bCCD31+Ly-6C+Ly-6GC and CD11b+CD31+Ly-6C+Ly-6GC populations were responsible for the development of macrophages in culture (data not shown) as indicated by positive staining for F4/80, CD11b, and FcR (CD16/CD32) of adherent cells. There was a twofold decrease (P < 0.002) in the CD11bCCD31+Ly-6C+Ly-6GC and CD11b+CD31+Ly-6C+Ly-6GC populations in p21C/C mice as compared to wt mice (Figure 6B) . Although there was also a reduction (twofold, P < 0.002) in the CD11b+CD31+Ly-6C+Ly-6G+ population in p21C/C as compared to wt mice (data not shown), this population failed to develop into macrophages in culture and Giemsa-Wright staining of the sorted cells, demonstrating that the vast majority of these cells were mature neutrophils (data not shown). Collectively, these data suggest that p21 is required for the differentiation of monocyte progenitors and/or precursors independent of its inhibitory role in apoptosis.

Adoptive Transfer of wt BMDMs Restores the Development of Arthritis in p21C/C Mice

Because p21-deficient mice have a deficiency in the inflammatory monocyte population in peripheral blood, we attempted to reconstitute this population by adoptively transferring cultured wt BMDMs into p21C/C mice and then administering K/BxN serum. There was a ninefold increase (P < 0.02) in ankle swelling in p21C/C mice that received wt BMDMs as compared to uninjected p21C/C mice at 7 days after serum transfer (Figure 7) . These data suggest that reconstitution of the macrophage population in p21C/C mice partially eliminates their resistance to serum transfer-induced arthritis.

Figure 7. Adoptive transfer of wt BMDMs restores sensitivity to serum transfer in p21C/C mice. wt BMDMs were grown in culture as described in Materials and Methods and intravenously injected via tail vein into p21C/C mice at 1 day before serum transfer, 1 day after serum transfer, and 3 days after serum transfer. Ankle circumference was measured as described in Figure 1A . The values represent the mean ?? SE of ankles/time point, which were compared by Student??s t-test to mice not receiving BMDMs under parallel conditions. (p21C/C: day 2, n = 12 mice; day 7, n = 20 mice; wt: day 2, n = 16 mice; day 7, n = 23 mice).

Discussion

To date all studies that have examined the function of p21 in RA have specifically focused on the synovial fibroblast population. We demonstrated a twofold reduction in the number of synovial fibroblasts expressing p21 in synovial tissue isolated from RA patients as compared to control osteoarthritis patients.19 Articular injection of replication-defective adenoviruses engineered to overexpress p21 prevents development of experimental arthritis.20,21,33,34 Based on these data we anticipated that p21C/C mice would have a marked increase in the development of experimental arthritis. However, our data reveal a surprising role for endogenous p21 in the effector phase of arthritis. The p21-deficient mice are resistant to K/BxN serum transfer-induced arthritis, a model system for the effector phase of arthritis.30 Although these data suggest that p21 may be pro-arthrogenic, the paradoxical role for p21 may be explained by our observation that p21 is required for the development of inflammatory monocytes. Our data also indicate that the level of inflammatory monocytes influences the sensitivity of arthritis following transfer of K/BxN serum. In support of this concept, Wipke et al40 showed that the addition of anti-Gr-1 antibody (Clone RB6C8C5), which binds to neutrophils and to inflammatory monocytes,37 results in their depletion (data not shown), and prevents serum transfer-induced arthritis.40 Thus, these data suggest that both neutrophils and inflammatory monocytes are central contributors to the pathogenesis of inflammatory arthritis following transfer of K/BxN serum. Because there are no differences in the number of neutrophils in p21C/C mice as compared to wt mice, we can infer from these data that in this model the inflammatory monocytes may be necessary for arthritis development. Recently, a study showed that depletion of macrophages by clodronate liposome treatment leaves mice completely resistant to K/BxN serum transfer-induced arthritis.41 Further, adoptive transfer of BMDMs from wt mice restores sensitivity to K/BxN serum transfer-induced arthritis. Taken together, these data suggest that failure to recruit macrophages following an inflammatory response is one of the primary defects in p21C/C mice that may be responsible for the lack of development of inflammatory arthritis.

The lack of inflammatory monocytes suggests a mechanism for decreased sensitivity to serum transfer-induced arthritis in p21C/C mice. These data also indicate that a potential function of inflammatory monocytes may be to traffic the GPI-anti-GPI antibody immune complex to the articular surface. A reduction in anti-GPI antibody deposition on the articular cartilage surface is observed in mice deficient in Fc receptors or in mice depleted of Gr-1-positive cells.30 Because inflammatory monocytes have Fc receptors on their cell surface and because they express Gr-1, these data suggest that one of the functions of inflammatory monocytes may be to traffic the GPI-anti-GPI-immune complex. Thus, it is possible that the lack of inflammatory monocytes contributes to a reduction in the total levels of Fc receptor. However, this is unlikely because all of the monocytes and neutrophils in p21C/C and control mice were positive for Fc receptor, and there were no differences in the mean fluorescent intensity of the receptors on the cell surface (data not shown). These data indicate that a similar concentration of Fc receptors are present on myeloid cells of p21C/C and control mice. Alternatively, there might be an increase in the clearance of the anti-GPI autoantibody in serum of p21C/C mice. Mice lacking FcRn, which is required for stabilization of IgGs, display increased clearance of the anti-GPI autoantibody, and excessive amounts of serum lead to the induction of arthritis.42 Thus, the lack of inflammatory monocytes or the lack of p21 in other hemopoietic cells may affect the stabilization of the anti-GPI antibody. However, we observed similar levels of GPI autoantibodies over time in the joints isolated from wt and p21C/C mice. These data suggest that the resistance to arthritis observed in p21C/C mice is not due to a defect in the binding to Fc receptors or in the trafficking of the anti-GPI autoantibody to the joint.

Similar to RA, the inflammatory response is a major contributor to the pathogenesis of atherosclerosis.43 Monocytes and macrophages are required for the development of experimental atherosclerosis in mice.44 Recently, p21-deficient mice have been shown to be resistant to experimental atherosclerosis.10 On the other hand, p27C/C mice show increased arterial inflammation, macrophage accumulation, and lesion size when crossed with ApoEC/C mice.45 We show that p21C/C but not p27C/C mice are resistant to K/BxN serum transfer-induced arthritis. Thus, these data demonstrate a differential role for the cdk inhibitor family members during the development of experimental atherosclerosis and arthritis.

p21 is known to be essential for differentiation of skeletal muscle cells,46 oligodendrocytes,47 keratinocytes,48 and neuronal cells.49 p21 has also been shown to be required for differentiation of monocytic cells and for maintaining their survival in culture.50,51 Moreover, in human monocytes isolated from peripheral blood, antisense oligonucleotides to p21 block differentiation into macrophages.52 These data suggest that p21 is required for differentiation and the survival of monocyte cells that remain in the cell cycle. The effect of p21 on BMDMs is somewhat different from monocytic cell lines and primary monocytes, because p21 is not required for the differentiation of BMDMs. Additionally, in BMDMs, which also proliferate, p21 is essential for prevention of apoptosis but not for arresting the cell cycle in response to interferon or decorin.53-55 However, p21 has no effect on survival or growth arrest in BMDMs deprived of M-CSF or treated with lipopolysaccharide.55 Here we show that p21 is also necessary for the development of monocytes in vivo, independent of its role in apoptosis. Although there are substantially less inflammatory monocytes in circulation and myeloid progenitors or precursors in bone marrow in p21C/C mice, the apoptotic indices are not increased in bone marrow and blood of p21C/C mice as compared to control mice (data not shown). Because we observe a decrease in myeloid precursors and because there is a similar complement of neutrophils in p21C/C and control mice, these data suggest that p21 may be required for differentiation of monocytes in bone marrow at a developmental stage that is distinct from granulocytes. Further, these monocytes are necessary for the effector phase of arthritis.

Figure 3. Increased inflammation and destruction in wt mice as compared to p21C/C mice following transfer of serum. At 4, 6, and 10 days following serum transfer, ankles as described in Figure 1 were harvested, sectioned, and stained with hematoxylin (blue) and eosin (pink). Shown are representative photomicrographs (40x or 200x) of the joint from wt and p21C/C mice. Arrows denotes invading pannus. SL, synovial lining; B, bone; P, pannus.

Acknowledgements

We thank Dr. Maripat Corr for her assistance and consult with the anti-GPI enzyme-linked immunosorbent assays. We also thank Joy Eslick for her assistance with the flow cytometry and Jacqueline Spencer for her assistance with tail vein injections.

【参考文献】
  Sherr CJ: Cancer cell cycles. Science 1996, 274:1672-1677

Gao CY, Zelenka PS: Cyclins, cyclin-dependent kinases and differentiation. BioEssays 1997, 19:307-315

Kitaura H, Shinshi M, Uchikoshi Y, Ono T, Iguchi-Ariga SM, Ariga H: Reciprocal regulation via protein-protein interaction between c-Myc and p21(cip1/waf1/sdi1) in DNA replication and transcription. J Biol Chem 2000, 275:10477-10483

Shim J, Lee H, Park J, Kim H, Choi EJ: A non-enzymatic p21 protein inhibitor of stress-activated protein kinases. Nature 1996, 381:804-806

Coqueret O, Gascan H: Functional interaction of STAT3 transcription factor with the cell cycle inhibitor p21WAF1/CIP1/SDI1. J Biol Chem 2000, 275:18794-18800

Martin-Caballero J, Flores JM, Garcia-Palencia P, Serrano M: Tumor susceptibility of p21(Waf1/Cip1)-deficient mice. Cancer Res 2001, 61:6234-6238

Lawson BR, Baccala R, Song J, Croft M, Kono DH, Theofilopoulos AN: Deficiency of the cyclin kinase inhibitor p21(WAF-1/CIP-1) promotes apoptosis of activated/memory T cells and inhibits spontaneous systemic autoimmunity. J Exp Med 2004, 199:547-557

Santiago-Raber ML, Lawson BR, Dummer W, Barnhouse M, Koundouris S, Wilson CB, Kono DH, Theofilopoulos AN: Role of cyclin kinase inhibitor p21 in systemic autoimmunity. J Immunol 2001, 167:4067-4074

Balomenos D, Martin-Caballero J, Garcia MI, Prieto I, Flores JM, Serrano M, Martinez AC: The cell cycle inhibitor p21 controls T-cell proliferation and sex-linked lupus development. Nat Med 2000, 6:171-176

Merched AJ, Chan L: Absence of p21Waf1/Cip1/Sdi1 modulates macrophage differentiation and inflammatory response and protects against atherosclerosis. Circulation 2004, 110:3830-3841

Pope RM, Perlman H: Rheumatoid arthritis. Tsokos GC eds. Current Molecular Medicine: Principles of Molecular Rheumatology. 2000:pp 325-361 Humana Press Inc, Totowa, NJ

van den Berg WB, van Lent PLEM: The role of macrophages in chronic arthritis. Immunobiology 1996, 195:614-623

Choy EH, Panayi GS: Cytokine pathways and joint inflammation in rheumatoid arthritis. N Engl J Med 2001, 344:907-916

Campbell IK, Rich MJ, Bischof RJ, Hamilton JA: The colony-stimulating factors and collagen-induced arthritis: exacerbation of disease by M-CSF and G-CSF and requirement for endogenous M-CSF. J Leukoc Biol 2000, 68:144-150

Yang YH, Hamilton JA: Dependence of interleukin-1-induced arthritis on granulocyte-macrophage colony-stimulating factor. Arthritis Rheum 2001, 44:111-119

Feldmann M, Brennan FM, Maini RN: Role of cytokines in rheumatoid arthritis. Annu Rev Immunol 1996, 14:397-440

Nuki G, Bresnihan B, Bear MB, McCabe D: Long-term safety and maintenance of clinical improvement following treatment with anakinra (recombinant human interleukin-1 receptor antagonist) in patients with rheumatoid arthritis: extension phase of a randomized, double-blind, placebo-controlled trial. Arthritis Rheum 2002, 46:2838-2846

Genovese MC, Bathon JM, Martin RW, Fleischmann RM, Tesser JR, Schiff MH, Keystone EC, Wasko MC, Moreland LW, Weaver AL, Markenson J, Cannon GW, Spencer-Green G, Finck BK: Etanercept versus methotrexate in patients with early rheumatoid arthritis: two-year radiographic and clinical outcomes. Arthritis Rheum 2002, 46:1443-1450

Perlman H, Bradley K, Hongtao L, Cole S, Shamiyeh E, Smith RC, Walsh K, Fiore S, Koch AE, Firestein GS, Haines GK, Pope RM: IL-6 and MMP-1 are regulated by the cyclin-dependent kinase inhibitor p21 in synovial fibroblasts. J Immunol 2003, 170:838-845

Nasu K, Kohsaka H, Nonomura Y, Terada Y, Ito H, Hirokawa K, Miyasaka N: Adenoviral transfer of cyclin-dependent kinase inhibitor genes suppresses collagen-induced arthritis in mice. J Immunol 2000, 165:7246-7252

Nonomura Y, Kohsaka H, Nasu K, Terada Y, Ikeda M, Miyasaka N: Suppression of arthritis by forced expression of cyclin-dependent kinase inhibitor p21(Cip1) gene into the joints. Int Immunol 2001, 13:723-731

Kouskoff V, Korganow AS, Duchatelle V, Degott C, Benoist C, Mathis D: Organ-specific disease provoked by systemic autoimmunity. Cell 1996, 87:811-822

Korganow AS, Ji H, Mangialaio S, Duchatelle V, Pelanda R, Martin T, Degott C, Kikutani H, Rajewsky K, Pasquali JL, Benoist C, Mathis D: From systemic T cell self-reactivity to organ-specific autoimmune disease via immunoglobulins. Immunity 1999, 10:451-461

Ditzel HJ: The K/BxN mouse: a model of human inflammatory arthritis. Trends Mol Med 2004, 10:40-45

Perlman H, Liu H, Georganas C, Koch AE, Shamiyeh E, Haines GK, Pope RM: Differential expression pattern of the anti-apoptotic proteins, Bcl-2 and Flip in experimental arthritis. Arthritis Rheum 2001, 44:2899-2908

Brown NJ, Hutcheson J, Bickel E, Scatizzi JC, Albee LD, Haines GK, 3rd, Eslick J, Bradley K, Taricone E, Perlman H: Fas death receptor signaling represses monocyte numbers and macrophage activation in vivo. J Immunol 2004, 173:7584-7593

Pettit AR, Ji H, von Stechow D, Muller R, Goldring SR, Choi Y, Benoist C, Gravallese EM: TRANCE/RANKL knockout mice are protected from bone erosion in a serum transfer model of arthritis. Am J Pathol 2001, 159:1689-1699

Corr M, Crain B: The role of FcR signaling in the K/BxN serum transfer model of arthritis. J Immunol 2002, 169:6604-6609

el-Deiry WS, Harper JW, O??Connor PM, Velculescu VE, Canman CE, Jackman J, Pietenpol JA, Burrell M, Hill DE, Wang Y: WAF1/CIP1 is induced in p53-mediated G1 arrest and apoptosis. Cancer Res 1994, 54:1169-1174

Kyburz D, Corr M: The KRN mouse model of inflammatory arthritis. Springer Semin Immunopathol 2003, 25:79-90

van Lent P, Nabbe KC, Boross P, Blom AB, Roth J, Holthuysen A, Sloetjes A, Verbeek S, van den Berg W: The inhibitory receptor FcRII reduces joint inflammation and destruction in experimental immune complex-mediated arthritides not only by inhibition of FcRI/III but also by efficient clearance and endocytosis of immune complexes. Am J Pathol 2003, 163:1839-1848

Wipke BT, Wang Z, Nagengast W, Reichert DE, Allen PM: Staging the initiation of autoantibody-induced arthritis: a critical role for immune complexes. J Immunol 2004, 172:7694-7702

Nonomura Y, Kohsaka H, Nagasaka K, Miyasaka N: Gene transfer of a cell cycle modulator exerts anti-inflammatory effects in the treatment of arthritis. J Immunol 2003, 171:4913-4919

Taniguchi K, Kohsaka H, Inoue N, Terada Y, Ito H, Hirokawa K, Miyasaka N: Induction of the p16INK4a senescence gene as a new therapeutic strategy for the treatment of rheumatoid arthritis. Nat Med 1999, 5:760-767

Leenen PJ, de Bruijn MF, Voerman JS, Campbell PA, van Ewijk W: Markers of mouse macrophage development detected by monoclonal antibodies. J Immunol Methods 1994, 174:5-19

Schaller E, Macfarlane AJ, Rupec RA, Gordon S, McKnight AJ, Pfeffer K: Inactivation of the F4/80 glycoprotein in the mouse germ line. Mol Cell Biol 2002, 22:8035-8043

Geissmann F, Jung S, Littman DR: Blood monocytes consist of two principal subsets with distinct migratory properties. Immunity 2003, 19:71-82

Soper BW, Duffy TM, Lessard MD, Jude CD, Schuldt AJ, Vogler CA, Levy B, Barker JE: Transplanted ER-MP12hi20C58med/hi myeloid progenitors produce resident macrophages from marrow that are therapeutic for lysosomal storage disease. Blood Cells Mol Dis 2004, 32:199-213

McCormack JM, Leenen PJ, Walker WS: Macrophage progenitors from mouse bone marrow and spleen differ in their expression of the Ly-6C differentiation antigen. J Immunol 1993, 151:6389-6398

Wipke BT, Allen PM: Essential role of neutrophils in the initiation and progression of a murine model of rheumatoid arthritis. J Immunol 2001, 167:1601-1608

Solomon S, Rajasekaran N, Jeisy-Walder E, Snapper SB, Illges H: A crucial role for macrophages in the pathology of K/BxN serum-induced arthritis. Eur J Immunol 2005, 35:3064-3073

Akilesh S, Petkova S, Sproule TJ, Shaffer DJ, Christianson GJ, Roopenian D: The major histocompatibility complex class I-like Fc receptor promotes humorally mediated autoimmune disease. J Clin Invest 2004, 113:1328-1333

Libby P: Inflammation in atherosclerosis. Nature 2002, 420:868-874

Smith JD, Trogan E, Ginsberg M, Grigaux C, Tian J, Miyata M: Decreased atherosclerosis in mice deficient in both macrophage colony- stimulating factor (op) and apolipoprotein E. Proc Natl Acad Sci USA 1995, 92:8264-8268

Diez-Juan A, Andres V: The growth suppressor p27(Kip1) protects against diet-induced atherosclerosis. FASEB J 2001, 15:1989-1995

Wang J, Walsh K: Resistance to apoptosis conferred by Cdk inhibitors during myocyte differentiation. Science 1996, 273:359-361

Zezula J, Casaccia-Bonnefil P, Ezhevsky SA, Osterhout DJ, Levine JM, Dowdy SF, Chao MV, Koff A: p21cip1 is required for the differentiation of oligodendrocytes independently of cell cycle withdrawal. EMBO Rep 2001, 2:27-34

Missero C, Calautti E, Eckner R, Chin J, Tsai LH, Livingston DM, Dotto GP: Involvement of the cell-cycle inhibitor Cip1/WAF1 and the E1A-associated p300 chain protein in terminal differentiation. Proc Natl Acad Sci USA 1995, 92:5451-5455

Poluha W, Poluha DK, Chang B, Crosbie NE, Schonhoff CM, Kilpatrick DL, Ross AH: The cyclin-dependent kinase inhibitor p21WAF1 is required for survival of differentiating neuroblastoma cells. Mol Cell Biol 1996, 16:1335-1341

Asada M, Yamada T, Fukumuro K, Mizutani S: p21Cip1/WAF1 is important for differentiation and survival of U937 cells. Leukemia 1998, 12:1944-1950

Liu X, Lee M-H, Cohen M, Bommakanti M, Freedman LP: Transcriptional activation of the Cdk inhibitor p21 by vitamin D3 leads to the induced differentiation of the myelomonocytic cell line U937. Genes & Dev 1996, 10:142-153

Kramer JL, Baltathakis I, Alcantara OS, Boldt DH: Differentiation of functional dendritic cells and macrophages from human peripheral blood monocyte precursors is dependent on expression of p21 (WAF1/CIP1) and requires iron. Br J Haematol 2002, 117:727-734

Comalada M, Xaus J, Sanchez E, Valledor AF, Celada A: Macrophage colony-stimulating factor-, granulocyte-macrophage colony-stimulating factor-, or IL-3-dependent survival of macrophages, but not proliferation, require the expression of p21(Waf1) through the phosphatidylinositol 3-kinase/Akt pathway. Eur J Immunol 2004, 34:2257-2267

Xaus J, Comalada M, Cardo M, Valledor AF, Celada A: Decorin inhibits macrophage colony-stimulating factor proliferation of macrophages and enhances cell survival through induction of p27(Kip1) and p21(Waf1). Blood 2001, 98:2124-2133

Xaus J, Cardo M, Valledor AF, Soler C, Lloberas J, Celada A: Interferon gamma induces the expression of p21waf-1 and arrests macrophage cell cycle, preventing induction of apoptosis. Immunity 1999, 11:103-113

作者单位：From the Department of Molecular Microbiology and Immunology* and the Department of Medicine, Division of Rheumatology, School of Medicine, St. Louis University, St. Louis, Missouri; the Department of Microbiology and Immunology, Chicago College of Osteopathic Medicine, Midwestern University, Downer