Hematopoietic Cell Regulation by Rac1 and Rac2 Guanosine Triphosphatases

See allHide authors and affiliations

Science  17 Oct 2003:
Vol. 302, Issue 5644, pp. 445-449
DOI: 10.1126/science.1088485


The Rho guanosine triphosphatases (GTPases) Rac1 and Rac2 are critical signaling regulators in mammalian cells. The deletion of both Rac1 and Rac2 murine alleles leads to a massive egress of hematopoietic stem/progenitor cells (HSC/Ps) into the blood from the marrow, whereas Rac1–/– but not Rac2–/– HSC/Ps fail to engraft in the bone marrow of irradiated recipient mice. In contrast, Rac2, but not Rac1, regulates superoxide production and directed migration in neutrophils, and in each cell type, the two GTPases play distinct roles in actin organization, cell survival, and proliferation. Thus, Rac1 and Rac2 regulate unique aspects of hematopoietic development and function.

Rho GTPases, members of the Ras superfamily, are critical regulators of cellular function and signal transduction pathways in eukaryotic cells. In mammalian cells, the best-studied members— Rho, Rac, and Cdc42—play distinct roles in regulating actin assembly and motility (1). However, the roles of Rho GTPases in hematopoietic cell development and function have only recently begun to be elucidated. There are three Rac GT-Pases—Rac1, Rac2, and Rac3—and their high degree of homology suggests potential overlapping functions (2, 3). Rac1 is ubiquitously expressed, whereas the expression of Rac2 is restricted to cells of hematopoietic origin (2, 3) and Rac3 is expressed primarily in the brain (4).

We have previously reported the critical roles of Rac2 in a wide variety of primary hematopoietic cells, including the regulation of adhesion, migration, oxidase activity, and gene expression (510). These same functions have been attributed to Rac1 in nonhematopoietic cells in which Rac2 is not expressed (1114). However, the roles of Rac1 compared with the roles of Rac2 in blood cells, which express both GTPases, remain to be elucidated.

Because homozygous Rac1-deficient mice die at ∼E8 (embryonic day 8) in utero (15), we generated mice with a conditional Rac1 (flox) allele (Rac1flox/flox) (16) (fig. S1). These were compared with mice that were homozygous for both the Rac1 flox allele and a Rac2-null allele (Rac1flox/flox;Rac2–/–). Floxed Rac1 sequences were deleted by means of two methods, and the deletion of Rac1 sequences was confirmed with polymerase chain reactions (PCRs) and immunoblots (figs. S2 and S3) (17).

Most prominently, the absence of Rac1 led to a significant reduction compared with that of the wild type in the ability of hematopoietic stem/progenitor cells (HSC/Ps) to reconstitute hematopoiesis in a non-obese diabetic/severe combined-immunodeficiency (NOD/SCID) engraftment model (18) (Fig. 1A). Engraftment is a multistep process requiring proliferation and differentiation of stem cells after the movement of these cells into the bone marrow from the blood and adhesion in the hematopoietic microenvironment (19). Rac2–/– cells demonstrated normal short-term engraftment, indicating that defective engraftment was specific for Rac1 deficiency. Although modest movement (mobilization) of HSC/Ps out of the marrow cavity into the circulating blood was observed in Rac2–/– mice, as previously reported (6), the absence of both Rac1 and Rac2 resulted in a massive mobilization of progenitor colony-forming unit cells (CFU-C) into the peripheral circulation (Fig. 1B). These phenotypes occurred despite the normal expression of β1 integrin adhesion molecules (20) and the normal expression of CXCR4, the receptor for the stromal-derived factor-1 (SDF-1), both of which have previously been implicated in the engraftment and mobilization of stem cells (21) (Fig. 1C). However, whereas Rac1–/– HSC/Ps showed normal adhesion to fibronectin, Rac2–/– HSC/Ps and, more prominently, Rac1–/–;Rac2–/– cells displayed significantly decreased adhesion to fibronectin (Fig. 1D), strongly suggesting that Rac2 has a predominant but overlapping role with Rac1 in integrin-mediated stem cell adhesion. Moreover, mobilization of Rac1–/–;Rac2–/– HSC/Ps was associated with significantly increased expression of CXCR4 (Fig. 1C).

Fig. 1.

Defective phenotypes of Rac-deficient hematopoietic cells in vivo. (A) Reduced engraftment of Rac1-deficient HSC/Ps. Wild-type (WT) or Rac-deficient bone marrow nucleated cells (1 × 106 per mouse) were transplanted into NOD/SCID recipient mice, and the engraftment in peripheral blood was determined by flow cytometry using the antibody to CD45–allophycocyanin (APC) and the antibody to H2Kb-phycoerythrin at 6 weeks posttransplantation (16). Error bars show the mean ± SEM; 7 to 12 mice per group were analyzed. *, P < 0.05 for Rac1–/– and Rac1–/–;Rac2–/– cells compared with wild-type and Rac2–/– cells. (B) Enhanced mobilization of HSC/Ps from the bone marrow of Rac1–/–;Rac2–/– mice. CFU-C were enumerated in the peripheral blood of each genotype 2 days after the last of eight doses of polyinosine: polycytidylic. Error bars show the mean ± SEM; n = 3. *, P < 0.05 for Rac1–/–;Rac2–/– cells compared with wild-type cells. (C) Expression of CXCR4 on the HSC/P surface shown with the fluorescein isothiocyanate–conjugated antibody to CXCR4. The result represents the MFI analyzed by flow cytometry. Error bars show the mean ± SD; n = 3. *, P < 0.01 for Rac1–/–;Rac2–/– cells compared with wild-type cells. (D) Adhesion of HSC/Ps by means of α4β1 and α5β1 integrins to recombinant fibronectin fragments H296 and CH271, respectively (29). Error bars show the mean ± SD; n = 3. BSA, bovine serum albumin. The results shown in each panel are representatives of three independent experiments.

Rac1–/– HSC/Ps also displayed impaired growth factor–stimulated in vitro growth, as determined by progenitor colony formation (Fig. 2A and fig. S4) and expansion in liquid culture (fig. S5). Reduced Rac1–/– HSC/P growth was associated with significantly decreased thymidine incorporation (Fig. 2B). Rac1–/–;Rac2–/– cells had a more severe reduction in proliferation compared with that of the wild-type and Rac2–/– cells, and they formed profoundly abnormal colonies with no cellular halo, suggesting combined effects of impaired growth and migration. Indeed, this severe phenotype in Rac1–/–;Rac2–/– cells was associated with reduced proliferation (Fig. 2B), increased apoptosis associated with Rac2 deficiency (Fig. 2C and fig. S6), and profoundly decreased migration in response to SDF-1, as compared with that of wild-type cells (Fig. 2D).

Fig. 2.

Effect of Rac deficiency on the regulation of HSC/P proliferation, survival, and migration. HSC/Ps [lineage-depleted, c-kit–positive] (lin and c-kit+) of each genotype generated by in vivo Cre expression (16, 30) were used for these assays. Similar results were obtained using enhanced green fluorescent protein (EGFP+) cells after transduction in vitro with a bicistronic retrovirus vector expressing Cre and EGFP. (A) Progenitor colony-forming assay. HSC/Ps were plated in methylcellulose in the presence of SCF (100 ng/ml) and incubated for 7 days at 37°C. Colonies were enumerated and photographed using an inverted microscope and an attached charge-coupled device camera. (B) Cell proliferation. HSC/Ps were starved in RPMI 1640 and 1% serum for 8 hours and stimulated with SCF (100 ng/ml) for 48 hours. [3H]thymidine was added for 6 hours at 37°C. Cells were harvested on the filter and the β-emission was measured. Error bars show the mean ± SD; n = 6. CPM, counts per minute × 103. (C) Cell apoptosis. One hundred thousand cells were stained with APC-conjugated Annexin-V after SCF stimulation for 48 hours. The percentage of apoptosis (Annexin-V+) was determined by flow analysis. Error bars show the mean ± SD; n = 3. (D) Migration of HSC/Ps in a transwell chamber assay in response to SDF-1. The result represents the percentage of CFUs formed by cells that migrated in response to SDF-1 (100 ng/ml). Analysis by enumerating the number of the migrated cells showed similar results. Error bars show the mean ± SD; n = 3. The results shown in each panel are representatives of a minimum of three independent experiments.

To determine the mechanism of reduced HSC/P growth, we undertook additional analysis. Significantly fewer Rac1–/– HSC/Ps entered S and G2/M over 24 to 48 hours in response to stem cell factor (SCF), a growth factor for primitive hematopoietic cells, as compared with those in wild-type or Rac2–/– cells (Fig. 3A and fig. S7). We found that levels of cyclin D1, which is required for G1-S progression, were not detectable in Rac1–/– HSC/Ps after SCF stimulation (Fig. 3B). Rac1–/– but not Rac2–/– HSC/Ps also showed decreased extracellular signal–regulated kinase (ERK) (p42/p44) phosphorylation (Fig. 3C). We found these defects in cell cycle progression and signaling directly related to Rac1 deficiency, because expression of the Rac1 protein but not the Rac2 protein, induced with retrovirus-mediated gene transfer, restored cycle progression and ERK activation in Rac1–/– cells (Fig. 3A and fig. S8). Inhibitor studies in wild-type cells confirmed the role of ERK in cell cycle progression and cyclin D1 induction (figs. S9 and S10). In addition, SCF-induced reduction in expression of cyclin-dependent kinase (Cdk) inhibitor p27kip1 was absent in Rac1–/– cells (Fig. 3B).

Fig. 3.

Effect of Rac deficiency on cell cycle progression, signaling pathways, and actin polymerization in HSC/Ps. Wild-type, Rac2–/–, Rac1–/–, and Rac1–/–; Rac2–/– HSC/Ps (lin, c-kit+) that were generated by in vivo Cre expression (as in Fig. 2) were used for these assays. Similar results were obtained with cells transduced in vitro with a Cre-expressing retrovirus vector. (A) Cell cycle progression. Cells were starved and stimulated with SCF (100 ng/ml). Cycle progression into S and G2/M phases was determined by bromodeoxyuridine (BrdU) incorporation 48 hours after SCF stimulation, and cell cycle distribution was analyzed by flow cytometry using the antibody to BrdU and 7-amino actinomycin D staining. Error bars show the mean ± SD; n = 3. For genetic restoration studies, Rac1–/– and Rac2–/– HSC/Ps were transduced with a retrovirus, MIEG3–hemagglutinin antigen (HA)–Rac2 (expressing wild-type Rac2) or MIEG3-HA-Rac1 (expressing wild-type Rac1). Two days after transduction, EGFP+ and c-Kit+ cells were isolated by a fluorescence-activated cell sorter and used for cell cycle analysis. (B to D) Immunoblot analyses. HSC/Ps of each genotype were starved as above and stimulated with SCF (100 ng/ml) at the time points indicated. Cell lysates were prepared for immunoblot analyses using antibodies specific for (B) cyclin D1, p27kip1, and p21cip1, (C) phospho-p42/p44, and (D) phospho-Akt. Membranes were then stripped and blotted with antibodies for (B) β-actin, (C) total p42/p44, and (D) total Akt as loading controls. (E) F-actin subcellular localization. HSC/Ps were serum-depleted in Hanks' balanced salt solution and stimulated with the chemokine, SDF-1 (100 ng/ml), for 30 s before they were fixed with 2% paraformaldehyde. Cells were stained with rhodamine-labeled phalloidin on chamber slides. Fluorescence images were acquired on a Leica microscope equipped with a deconvolution system driven by OpenLab software (31). Images shown are representatives of more than 100 cells examined for each genotype. The results shown in each panel are representatives of three independent experiments.

In contrast, increased apoptosis of Rac2–/– HSC/Ps was associated with reduced Akt activation compared with that of wild-type cells after SCF stimulation (Fig. 3D). Apoptosis and Akt activation in Rac1–/– HSC/Ps were similar to those of the wild-type cells. Expression of Rac2 but not Rac1 in Rac2–/– HSC/Ps by means of retrovirus-mediated gene transfer led to a complete reversal of the apoptotic phenotype (fig. S11). Inhibitor studies in wild-type cells confirmed the roles of phosphoinositide 3-kinase and Akt in mediating survival (fig. S12). Thus, although some overlap exists in these pathways, Rac1 predominantly regulates HSC/P cell cycle progression. Defective proliferation in these cells likely contributes to their lack of engraftment in vivo. Rac2 predominantly regulates apoptosis, which likely contributes to the profound defect in growth in vitro seen in Rac1–/–;Rac2–/– HSC/Ps.

To further examine the effects of Rac1 and Rac2 on cytoskeleton changes that are important in engraftment and mobilization, SDF-1–induced actin polymerization and cell shape changes were studied. Rac2–/– but not Rac1–/– HSC/Ps demonstrated markedly impaired cortical F-actin assembly (Fig. 3E) (22). In timed-lapsed video microscopy images, Rac1–/–; Rac2–/– cells demonstrated significantly reduced cell spreading and actin-based membrane protrusion and essentially no coordinated migration in response to SDF-1. Although Rac1–/– HSC/Ps showed relatively normal migration in this assay, in some cases, Rac1 deficiency was associated with an apparent delay or defect in the retraction of the uropod trailing the cell (movie S1). Thus, Rac1 and Rac2 appear to play both unique and overlapping roles in regulating the cytoskeleton, which affects adhesion and mobilization of HSC/Ps from the medullary cavity into the blood circulation.

We next studied whether Rac1 and Rac2 play unique roles in the regulation of cytoskeleton in neutrophils that are derived from HSC/Ps. Consistent with HSC/Ps, agonist-induced assembly of F-actin was distinctly different in Rac1–/– and Rac2–/– neutrophils (Fig. 4A). Rac2–/–, but not Rac1–/–, neutrophils demonstrated impaired cortical F-actin assembly (23). Rac1–/– and Rac2–/– neutrophils displayed normal adhesion, but Rac1–/– cells notably showed increased cell spreading by means of β2 integrins compared with that of wild-type cells (Fig. 4, B to D). Rac2–/– cells showed decreased cell spreading and Rac1–/–;Rac2–/– neutrophils were comparable in circumference to Rac2–/– cells but showed decreased adhesion by means of β2 integrins (Fig. 4D).

Fig. 4.

Rac1 and Rac2 roles in neutrophil functions. After transduction in vitro of HSC/Ps with a Cre-expressing retrovirus vector, EGFP+ cells were sorted and then cultured in cytokines to induce neutrophil differentiation. No significant changes were seen in neutrophil differentiation between each genotype. (A) F-actin subcellular localization. Neutrophils were stained with rhodamine-labeled phalloidin 15 s after f MLP stimulation. Fluorescence images were acquired (as in Fig. 3). One representative cell of each genotype from three independent experiments is shown. (B) Representative images of neutrophils adherent to plates coated with the antibody to CD18. (C) Circumference of adherent cells, as determined using Openlab software (31). The result represents the median of the circumference compared with that of the wild type. Fifty cells of each genotype from two independent experiments were counted. (D) Integrin-mediated neutrophil adhesion. Cells were incubated on plates previously coated with monoclonal antibodies to CD11a, CD11b, and CD18 for 30 min at 37°C, and they were fixed and counted under the light microscope. Error bars show the mean ± SEM; n = 3. *, P < 0.05 for Rac1–/–;Rac2–/– cells compared with the wild-type, Rac2–/–, and Rac1–/– cells. (E) Chemotaxis evaluated using a Boyden chamber assay in response to 1μM of f MLP. Cells were loaded into the upper chamber and incubated for 45 min at 37°C. The results represent the number of migrated cells per field. Error bars show the mean ± SEM; n = 3. (F) Superoxide production by NBT test. Neutrophils were analyzed by reduction of NBT in a chamber slide in response to 10 μM of f MLP. The results represent the percentage of positive cells. Error bars show the mean ± SEM; n = 3. (G) Expression and activation of ERK proteins. Phosphorylated and total p42/p44 were evaluated on immunoblot in lysates of neutrophils from each genotype stimulated with 10 μM of f MLP.

We next examined the functional significance of these changes in neutrophil actin assembly. In contrast to Rac2–/– neutrophils, which showed impaired chemotaxis (Fig. 4E), no defect in migration was observed in Rac1-deficient neutrophils stimulated with N-formyl-Met-Leu-Phe (fMLP). Rac1–/–;Rac2–/– neutrophils demonstrated markedly reduced migration. Rac1–/– cells also displayed normal frequency of migration when observed by video microscopy (movie S2). However, as seen in HSC/Ps, Rac1 deficiency was also accompanied by an abnormal retraction of the uropod in some migrating neutrophils, suggesting a subtle defect in F-actin polymerization (24). Rac2–/– neutrophils and, more prominently, Rac1–/–;Rac2–/– neutrophils showed defects in cell polarization and migration.

Both Rac1 and Rac2 have been demonstrated to be essential for oxidase activity of the reduced form of nicotinamide adenine dinucleotide phosphate (NADPH) in cell free assays (25, 26). However, after stimulation with f MLP, the percentage of nitro blue tetrazolium (NBT+) cells, a measure of oxidase activity, was identical in Rac1–/– neutrophils and wild-type cells (Fig. 4F). Rac2-deficient neutrophils displayed ∼60% reduction in the number of NBT+ neutrophils compared with that in wild-type cells. Deficiency of both Rac1 and Rac2 was associated with a higher reduction in the percentage of NBT+ cells compared with that of Rac2–/– cells (Fig. 4F).

We next determined whether, in a manner similar to HSC/Ps, Rac proteins regulated ERK activation in neutrophils. A reduction in f MLP-induced ERK phosphorylation was seen in both Rac1–/– and Rac2–/– cells, but this reduction was reproducibly more pronounced in Rac2–/– neutrophils (Fig. 4G). Rac1–/–;Rac2–/– neutrophils demonstrated severely reduced ERK phosphorylation. In wild-type neutrophils' cortical F-actin assembly, cell migration but not superoxide production was inhibited in the presence of the Mek (ERK kinase) inhibitor U0126 (figs. S13 to S16). Thus, Rac2 appears to be a physiologically critical Rac GTPase in neutrophil migration and NADPH oxidase function, whereas Rac1 plays a role in controlling cell spreading.

Adhesion and migration of cells within the hematopoietic microenvironment are critical for blood formation and blood cell function (27). We show here that the Rho GTPases Rac1 and Rac2 are key regulators of these functions in hematopoietic cells. Despite the high degree of sequence identity, each GTPase plays unique physiological roles, particularly with respect to cell growth, survival signaling pathways, and distinct actin structures that mediate different cytoskeleton functions. Given that each Rac protein contains identical sequences mediating known effector interactions, the basis for the specificity of these functions is still to be determined. We propose that these differences likely reside in the subcellular localization of each protein (28).

Supporting Online Material

Materials and Methods

Figs. S1 to S16


Movies S1 and S2

References and Notes

Stay Connected to Science

Navigate This Article