A point mutation in the murine Hem1 gene reveals an essential role for Hematopoietic Protein 1 in lymphopoiesis and innate immunity

In the course of the ENU program, a flow cytometry screen identified a pedigree (denoted NTB.1 for no T or B cells) in which the mice presented with an increased percentage of CD11b⁺ myeloid cells and a decreased percentage of B220⁺ B lymphocytes in peripheral blood samples (Fig. 1 A). To identify the gene responsible for the NTB.1 phenotype, G1 and G2 NTB.1 carriers were crossed to C3HeB/FeJ mice, and the resulting F1 animals were intercrossed to generate F2 and N2 animals for mapping (see Materials and methods). The levels of B220⁺ B cells in peripheral blood were assessed in the resultant generations. A genome scan of affected and unaffected animals indicated that the NTB.1 mutation mapped to chromosome 15 between 59.8 and 60.1 cM. Candidate genes were sequenced and a single-nucleotide polymorphism was identified in exon 13 of the Hem1 gene, which changed a CAG (Gln) to TAG (stop), resulting in premature termination at residue 445 (Fig. 1 B). Immunoblotting of whole cell lysates from normal littermate and Hem1 mutant thymocytes, splenocytes, and BM cells with an anti-Hem1 antibody (25) revealed the absence of Hem1 protein in mice homozygous for the NTB.1 mutation (Fig. 1 C; and see Fig. 7 F).

A previous Northern blot study using cell lines derived from various tissues suggested that Hem1 expression is restricted to hematopoietic tissues (20). We performed a more extensive survey of Hem1 expression using real-time PCR for Hem1 message in cDNA derived from 25 different mouse tissues. We find that although Hem- is predominantly expressed in hematopoietic tissues, it is also expressed in urogenital tissues including testis (Fig. S1, available). In contrast, Hem2 is not expressed in hematopoietic tissues but is widely expressed in many other tissues (Fig. S1). Indeed, mice deficient in Hem2 die around embryonic day 8.5–10.5 because of morphological abnormalities ranging from severe neurulation defects to complete resorption (26).

We next characterized the expression pattern of Hem1 in different hematopoietic cell lineages and progenitor cells. We find that Hem1 is expressed in lineage-negative (Lin⁻) Sca1⁺cKit⁺ side population (SP)⁺ cells, which are highly enriched for HSCs, as well as Lin⁻Sca1⁺cKit⁺ SP⁻ cells and Lin⁻Sca1⁻cKit⁻ cells, which contain early progenitors (Fig. S1). Hem1 is also expressed in B and T cell progenitors during all stages of maturation and in purified macrophages and neutrophils (Fig. S1). The results collectively indicate that Hem1 is predominantly expressed in developing and mature hematopoietic cells, whereas Hem2 is expressed in nonhematopoietic cells.

We performed complete blood counts (CBCs) and differential cell counts to determine the representation of mature hematopoietic cell types in peripheral blood from Hem1^−/− and normal WT littermate mice. We find that loss of Hem1 results in a moderate increase in total white blood cells (WBCs) per microliter of blood, which corresponds to a 25-fold increase in the number of neutrophils and a 3.5-fold decrease in the number of lymphocytes per microliter of blood in Hem1^−/− mice compared with WT mice (Figs. 2, A–C). Hem1^−/− mice also exhibit altered RBC indices, such as decreased hematocrit and mean corpuscular volume (MCV; Fig. 2 A), with reticulocytosis (increased RBC progenitors; Fig. 2 D), anisocytosis (variation in RBC size), echinocytosis (membrane spickles), and poikilocytosis (abnormally shaped RBC), which is characterized by keratocytes (horned-shaped cells), dacryocytes (tear-shaped cells), schistocytes (RBC fragmentation), and acanthocytes (thorny cells; Fig. 2 E and not depicted). Hem1^−/− RBCs exhibit increased fragility relative to WT RBCs, as measured by an increased percentage of lysis in response to hypotonic NaCl solutions (Fig. 2 F). Collectively, these RBC indices are suggestive of a microcytic hypochromic hemolytic anemia in Hem1^−/− mice, which parallels the anemia phenotype of Rac-deficient mice (27).

Gross and histological examination of Hem1^−/− mice reveal a spectrum of lesions, including dense amyloid accumulations containing foci of mineral at the margins of the liver (Fig. 2 G), and occasional diffuse inflammation in the liver, epididymis, mesentery, thoracic pleura, heart, and lungs (not depicted). Membranoproliferative glomerulopathy (Fig. 2 H) and diffuse lymphoid hypoplasia are also common lesions, whereas myocardial fibrosis, amyloidosis, and pancreatic atrophy occur less frequently (not depicted). Hem1^−/− spleens are substantially enlarged, even though lymphoid cells are depleted, because of marked proliferation of myeloid and erythroid progenitors (Fig. S2, available), which is indicative of extramedullary hematopoiesis. Hem1^−/− thymi are hypoplastic, and lack the normal cortical/medullary architecture evident in WT thymi (unpublished data).

We next examined the potential of clonogenic hematopoietic progenitor cells in the BM, spleen, and blood from Hem1^−/− and WT mice. We performed in vitro quantitation of myeloid (CFU-GM), erythroid (burst forming units-erythroid [BFU-E]), and multi-lineage (CFU-granulocyte/erythroid/macrophage/monocyte [CFU-GEMM]) assays using standard techniques. We find a significant reduction in CFU-GM, BFU-E, and CFU-GEMM in BM from Hem1^−/− mice compared with WT BM (Fig. S2). In contrast, there is a dramatic increase in CFU-GM, BFU-E, CFU-GEMM colonies in spleens and peripheral blood from Hem1^−/− mice compared with WT mice. This correlates with a fourfold increase in the percentage of Lin⁻Sca1⁺cKit⁻ cells in the spleens of Hem1^−/− mice compared with WT mice, which is consistent with increased mobilization of HSC and progenitors (HSC/Ps) from the BM to peripheral blood and spleen (Fig. S2).

Hem1^−/− mice exhibit an increase in the percentage (Fig. 3, A and B) and total number (not depicted) of cells bearing Mac1 and Gr1 (Fig. 3, A and B [top]), and CD61 (Fig. 3, A and B [bottom]) in the BM and spleen relative to WT mice, suggesting that myeloid cell homeostasis is altered by loss of Hem1 protein. We next determined whether loss of Hem1 affects neutrophil and macrophage functions. Purified neutrophils from Hem1^−/− and WT mice were loaded into trans-well plates, and the ability of cells to migrate through a filter containing different concentrations of MLP (a tripeptide chemoattractant) was measured. Hem1^−/− neutrophils fail to migrate efficiently across the filter in response to different concentrations of fMLP (Fig. 3 C). Hem1^−/− neutrophils are also deficient in their ability to phagocytose fluorescent-conjugated beads (Fig. 3 D) and bacteria (not depicted) in response to MLP stimulation, as measured by flow cytometry. Likewise, Hem1^−/− macrophages are deficient in their ability to phagocytose fluorescent-labeled Escherichia coli after 2 and 4 h of stimulation, relative to WT macrophages (Fig. 3 E). However, Hem1^−/− macrophages produce normal or increased levels of the cytokines IL1β, TNF-α, Mip2 (macrophage-inflammatory protein 2), and cytokine-induced neutrophil chemoattractant (KC) after stimulation with E. coli or LPS, and the transcription of the corresponding genes, as well as IL-23, are increased (Fig. S3, available).

The polymerization of F-actin is critical for forming phagosomes, as well as lamellipodia and filopodia, which are fingerlike projections responsible for polarizing the cytoskeletal machinery during cell migration. As shown in Fig. 3 F, >25% of WT neutrophils polymerize F-actin after 15 s of MLP treatment, as determined by increased phalloidin staining, whereas <5% of Hem1^−/− cells polymerize F-actin even after 120 s of MLP stimulation. Deconvolution images reveal that polymerized F-actin in Hem1^−/− cells is distributed diffusely in the cytoplasm, whereas F-actin in WT cells is clearly polarized toward the leading edge (Fig. 3 G). Transfection of Hem1 cDNA into BM-derived neutrophils (28) from WT or Hem1^−/− mice resulted in restoration of F-actin polymerization in response to MLP stimulation (Fig. S4, available). The results collectively indicate that Hem1 is required in innate immune cells for a variety of functions dependent on F-actin polymerization and cytoskeletal reorganization, including phagocytosis and migration.

Because Hem1-deficient mice are lymphopenic, we examined whether T and/or B cell development are impaired. The development of lymphocytes can be broken down into many distinct stages based on the presence or absence of particular intracellular and surface proteins, as determined by flow cytometry (29, 30). As shown in Fig. 4 A, Hem1^−/− mice have a 30-fold mean reduction in thymocyte number with an impairment in αβ T cell development within the CD4⁻CD8⁻ double-negative (DN) to CD4⁺CD8⁺ double-positive (DP) cell transition at the CD25^midCD44⁻ (DN3) to CD25⁻CD44⁻(DN4) cell stage, which is mediated by formation of the preTCR. The number of γδ thymocytes is also decreased, albeit to a lesser extent than αβ T cells (Fig. 4, A and B). The total numbers of peripheral CD4, CD8, and γδ T cells (Fig. 4 B) and B lymphocytes (not depicted) are reduced by at least 75%. The ratio of Foxp3⁺CD4⁺ T regulatory cells (T_reg) to Foxp3⁻CD4⁺ T cells is significantly higher in both thymus (Fig. 4 A) and spleen (Fig. 4 B), suggesting that Hem1 regulates both lymphocyte development and homeostasis.

To determine if the defects in T and B cell development are autonomous to hematopoietic cells or are caused by a defective microenvironment, we transferred marked Lin⁻ (lineage negative; Fig. 5) or total BM cells (Fig. S5, available) from WT (CD45.1⁺ marker), Hem1^−/− (CD45.2⁺), or mixed WT and Hem1^−/− Lin⁻ cells into unirradiated (Fig. 5), or lethally irradiated (Fig. S5), Rag2^−/−γc^−/− (recombinase activating gene 2/common γ chain double-null) hosts (CD45.2⁺), and lymphocyte development was analyzed 8–20 wk after transplant. As shown in Fig. 5 A, Rag2^−/−γc^−/− mice are completely blocked in T cell development at the CD4⁻CD8⁻ DN stage (columns 1 and 2). Injection of WT Lin⁻ cells into Rag2^−/−γc^−/− mice completely reconstitutes T cell development in the thymus, leading to normal representation of CD4 and CD8 cells in the spleen (Fig. 5 A, column 3). In contrast, Hem1^−/− Lin⁻ cells poorly reconstitute the thymi of Rag2^−/−γc^−/− mice, with the majority of cells residing at the DN stage (Fig. 5 A, column 4). Hem1^−/− CD4 and CD8 SP cells can mature and expand in Rag2^−/−γc^−/− hosts, as determined by the presence of CD4⁺CD45.2⁺ and CD8⁺CD45.2⁺ cells in the spleen of transplanted mice. However, when Hem1^−/− Lin⁻ cells are mixed with WT Lin⁻ cells at 1:1, 10:1, or 20:1 (Hem1^−/−/WT) ratios, they fail to reconstitute the thymus or spleen of transplanted mice, as determined by the presence of CD4 and CD8 cells bearing the CD45.1 (WT) but not CD45.2 markers (Fig. 5 A column 5; Fig. S5; and not depicted). The inability of Hem1^−/− hematopoietic cells to compete with WT hematopoietic cells in BM chimeras is not caused by inadequate space in the BM niche because we obtain similar results after transplantation of WT, Hem1^−/−, or mixed 1:10 WT:Hem1^−/− Lin⁻ or total BM cells into lethally irradiated Rag2^−/−γc^−/− hosts (Fig. S5 and not depicted). Hem1^−/− BM cells can contribute to the generation of GR-1⁺ myeloid cells when mixed 1:10 WT/Hem1^−/− BM cells and transplanted into lethally irradiated Rag2^−/−γc^−/− hosts, indicating that HSC/Ps are able to home to the BM niche in mixed BM chimeras (Fig. S5). In addition, Hem1 heterozygous (+/−) BM cells compete normally with WT BM cells to produce CD4⁺, CD8⁺, GR1⁺, and B220⁺ when mixed 1:10 WT/Hem1^+/−, indicating that the inability of Hem1^−/− cells to contribute is not because of allorejection by WT cells (Fig. S5).

We next determined the ability of Hem1-deficient progenitor cells to reconstitute the B cell lineage in either unirradiated (Fig. 5 B) or lethally irradiated Rag2^−/−γc^−/− hosts (Fig. S5). As shown in Fig. 5 B, Rag2^−/−γc^−/− mice are completely blocked in B cell development at the B220⁺IgM⁻ pro–B cell stage (columns 1 and 2). Lin⁻CD45.1⁺ WT progenitors reconstitute B220⁺IgM⁺ cells in the BM and spleen of Rag2^−/−γc^−/− hosts (Fig. 5 B, column 3). In contrast, Lin⁻CD45.2⁺ Hem1^−/− progenitors fail to generate mature B cells in unirradiated Rag2^−/−γc^−/− hosts (Fig. 5 B, column 4), despite partial reconstitution of mature T cells (Fig. 5 A). However, Hem1^−/− progenitors can generate a small population of mature B cells when transplanted into lethally irradiated Rag2^−/−γc^−/− hosts (Fig. S5). WT, but not Hem1^−/− progenitors, develop normally into B cells when mixed 1:1 or 1:10 with Hem1-deficient progenitors before transplantation into Rag2^−/−γc^−/− hosts (Fig. 5 B, columns 5 and 6; and Fig. S5). These results suggest that Hem1 plays an important role in the development and/or homeostasis of lymphocytes.

We next examined whether loss of Hem1 affects T cell activation and T helper differentiation. We find that Hem1 deficiency results in impaired T cell activation based on reduced CD69 expression (Fig. 6 A and Fig. S6, available), decreased cell division as assessed by limited CFSE dye dilution of live-gated cells (Fig. 6 B and Fig. S6), and reduced tritiated thymidine incorporation (Fig. S6) after TCR stimulation. However, IL-2 and IFN-γ are produced normally by Hem1^−/− T cells after TCR stimulation (Fig. 6, C and E), and exogenous IL-2 does not rescue proliferation of Hem1^−/− T cells after α–CD3-ε/CD28 stimulation (Fig. 6 B). These results suggest that the inability of Hem1^−/− T cells to divide is not because of the inability to make IL-2. Surprisingly, Hem1^−/− T cells produce 15-fold more IL-17 and twofold more IL-6 and TNF-α after α-CD3/CD28 stimulation, compared with their WT counterparts (Fig. 6 C), and contain increased serum IL-17 and IL-6 in vivo (Fig. S6). Purified CD4 T cells also contain a greater percentage of Th17 effector cells after in vitro stimulation, as assessed by intracellular staining with α–IL-17 and α–IL-2 antibodies (Fig. S7). To assess whether Hem1 might modulate Th17 differentation, we purified naive WT and Hem1^−/− CD4 T cells and subjected them to Th1 or Th17 polarizing conditions in the presence of irradiated APCs. We find that Th1 differentiation occurs normally in the absence of Hem1 based on an equivalent percentage of IFN-γ–producing cells (Fig. S7) and equivalent IFN-γ production (Fig. 6 E) from Hem1^−/− and WT CD4 T cells after Th1 culture conditions. In contrast, the generation of Th17 cells is enhanced after IL-17 polarizing conditions with the addition of TGF-β (Figs. 6 D). It is unlikely that the skewing toward Th17 cells is caused by secondary infections because we only occasionally find histological evidence of infection in Hem1^−/− mice, and Hem1^−/− mice treated with antibiotics throughout their lives still produce increased IL-17 (Fig. S7).

To determine how loss of Hem1 results in defective T cell activation and proliferation, we examined whether F-actin polymerization is altered in Hem1^−/− T cells. Thymocytes and peripheral T cells were stimulated with α–CD3-ε/CD28 for 24 h, followed by PMA/Ionomycin for 15 min, and then stained with Alexa Fluor 488–conjugated phalloidin. Both Hem1^−/− thymocytes (Fig. 7 A) and T cells (not depicted) fail to polymerize F-actin after α–CD3-ε/α-CD28 stimulation, relative to WT thymocytes and T cells. Hem1^−/− T cells also fail to form synapse-like actin cap structures after stimulation with α–CD3-ε-coated beads (Fig. 7 B). Thymocytes and T cells from Hem1^−/− mice exhibit reduced integrin-mediated cell adhesion to fibronectin after TCR stimulation, relative to WT cells (Fig. 7 C). These results suggest an essential role for Hem1 in cytoskeletal reorganization and cell adhesion at the immune synapse.

We next asked at what point Hem1 deficiency impinges on the signaling pathways leading to F-actin polymerization in lymphocytes and neutrophils. Hem1^−/− thymocytes and T cells express normal levels of the membrane-proximal signaling molecules CD3-ε (not depicted), Lck, Zap70, and Rac1 (Fig. 7 D, right). In addition, we observe equivalent levels of total phosphotyrosine, active (phospho) extracellular regulated mitogen-activated protein kinase (Erk), and active Rac1 (Pak1 binding) in WT and Hem1^−/− thymocytes, T cells stimulated with α–CD3-ε (Fig. 7 D), and neutrophils stimulated with MLP (Fig. 7 E), suggesting that membrane proximal signaling is intact. Hem1^−/− CD4 T cells also exhibit normal nuclear translocation of NF-κB (active NF-κB) after TCR stimulation and slightly increased basal NF-κB translocation before stimulation relative to WT CD4 T cells, as measured using ImageStream quantitative flow cytometry (Fig. S7) (31, 32). However, analysis of whole cell lysates from WT and Hem1^−/− thymocytes, T cells, and neutrophils indicates that loss of Hem1 results in the near complete absence of the WAVE complex proteins Wave2, Abi1, Abi2, and Sra1 in thymocytes and purified T cells (Fig. 7 F, left and middle), whereas WAVE members are reduced in Hem1^−/− neutrophils (Fig. 7 F, right). The reduction in WAVE complex proteins is not caused by decreased transcription because Abi1, Abi2, Sra1, and Wave2 mRNA are expressed normally in Hem1^−/− T cells relative to WT T cells (Fig. S8). In addition, acute small inhibitory RNA (siRNA) knockdown (33) of Hem1 in WT T cells in vitro results in decreased amounts of Wave2, Abi2, and Sra1 (Fig. S8). These results suggest that Hem1 is required for stabilization and/or translation of the WAVE complex proteins in hematopoietic cells (Fig. 8), and they are consistent with several recent studies indicating that siRNA knockdown of Hem1 in HL-60 cells (25) or Jurkat T cells (34) results in degradation of WAVE complex proteins, which are essential mediators of F-actin polymerization.

Reorganization of the actin cytoskeleton is implicated in many active cellular processes in hematopoietic cells including cell adhesion, migration, homing, phagocytosis, and proliferation. However, the specific signaling molecules and pathways that regulate the actin cytoskeleton in hematopoietic cells in vivo are unclear. In this study, we used a phenotype-driven ENU mutagenesis strategy in mice to identify a novel loss-of-function mutation in Hem1. Our studies reveal that Hem1 is an essential hematopoietic cell-specific regulator of the actin cytoskeleton, where it controls cell migration, adhesion, phagocytosis, and proliferation. Importantly, this study supports the utility of using ENU mutagenesis to generate immunodeficiency models in mice caused by single-point mutations, which more closely mimic the genetic alterations that result in human genetic diseases (35, 36). The relative hematopoietic cell-specific expression of Hem1, and the multilineage immunodeficiency phenotype of Hem1-deficient mice, suggests that a subset of immunodeficiency diseases in humans may include loss-of-function mutations in Hem1.

Analysis of Hem1^−/− mice reveals a severe defect in the development and expansion of αβ T cells, predominantly at the DN-to-DP cell transition. Further examination of gated DN populations by flow cytometry indicates that Hem1-deficient thymocytes are inhibited at the CD25⁺CD44⁻ DN3 to CD25⁻CD44⁻ DN4 cell stage, mediated by formation of the pre-TCR complex. Although Rac activation appears normal in Hem1^−/− thymocytes, F-actin polymerization and T cell capping are completely lost, and components of the WAVE complex disappear in the absence of Hem1. These results suggest that Hem1 has an essential role in T cell development and F-actin polymerization, perhaps by stabilizing the WAVE complex downstream of Rho family members. Our results extend a previous series of elegant studies suggesting that the Rac GTP exchange factor Vav1 (37, 38), as well as Rac (14) and Rho (16, 39, 40), are essential for T cell development at the DN-to-DP cell transition by regulating cell survival. Indeed, we also find an increased incidence of apoptosis in Hem1^−/− thymocytes and mature T cells (Fig. S6). The generation of T cell–specific Hem1-deficient mice will be required to more clearly delineate the involvement of Hem1 in pre-TCR signaling and to separate specific and potentially nonspecific contributions of Hem1 deficiency on T cell development.

Previous studies have also suggested that the actin-regulatory proteins Vav (see (41) for review), Rho (15), Rac (3), HS1 (42), Dynamin (43), Abi1/2 (44), and WAVE2 (34) have essential roles in T cell activation by regulating actin reorganization, immune synapse formation, and calcium signaling. In addition, Rac2 has been shown to regulate Th1 cell differentiation by inducing the IFN-γ promoter through the cooperative activation of the NF-κB and p38 mitogen-activated protein kinase pathways (4). We find that loss of Hem1 impairs T cell activation based on decreased CD69 up-regulation and decreased cell division of live-gated cells after α-CD3/α-CD28 stimulation. In addition, we find a defect in F-actin polymerization, actin capping, and cell adhesion in Hem1-deficient thymocytes and T cells. However, in contrast to loss of Vav1 (45), WASp (46), and Rac2 (3), we find that loss of Hem1 does not affect TCR-induced calcium signaling (unpublished data) and IL-2 production, nor is cell proliferation rescued by the addition of IL-2. In addition, we find that IFNγ, IL-6, and TNF-α production occurs normally in Hem1^−/− T cells. The differences between loss of Vav1 and Rac2 versus loss of Hem1 likely reflect a bifurcation of signals from Vav1 and Rac2 upstream of Hem1. Consistent with this notion, we find that loss of Hem1 specifically disrupts processes that are dependent on actin reorganization, such as cell migration, adhesion, and immune synapse formation, whereas NF-κB–dependent transcription of IL-2, IFN-γ, and proinflammatory cytokines proceed normally.

Loss of Vav1 (47) and Rac (17) have also been shown to be essential for normal B cell development and signaling by regulating cell survival. In addition, high Hem1 expression has recently been associated with a poor clinical outcome in B cell chronic lymphocytic leukemia, and down-regulation of Hem1 by antisense treatment increased susceptibility of chronic lymphocytic leukemia cells to fludarabine-mediated killing (48). Our results show that Hem1 is essential for both B and T cell development, and they are consistent with Hem1 acting downstream of Vav and Rac during lymphocyte development. These results collectively suggest that Hem1-deficient mice provide an important resource to examine the importance of Hem1 in transformation and metastasis of hematopoietic cancers.

Innate immune responses are dependent on the ability of phagocytic cells to sense invading microbes through Toll-like receptors and to respond to chemoattractant molecules by migrating to inflammatory sites, where they produce proinflammatory cytokines and phagocytose and kill invading pathogens. Our analyses of blood and BM from Hem1^−/− mice reveal severe but distinct alterations in innate immune responses associated with Hem1 deficiency. Purified neutrophils and macrophages from Hem1^−/− mice fail to phagocytose bacteria, and mutant neutrophils fail to migrate across a membrane and polymerize F-actin in response to stimulation. Neutrophil numbers are increased up to 25-fold in peripheral blood from Hem1^−/− mice, which is consistent with decreased migration in vivo. The neutrophilic response in Hem1^−/− mice also correlates with increased numbers of proinflammatory Th17 cells, increased serum IL-17 and IL-6, and increased production of IL-1β, Mip-2, and KC from in vitro–stimulated macrophages. Because IL-17 potently stimulates the expansion and chemotaxis of neutrophils in a G-CSF–dependent manner (49), the neutrophilic phenotype seen in Hem1^−/− mice may result from increased Th17 effector cells. Likewise, the increased serum amyloid A from liver cells (which is made in response to IL-1, IL-6, and TNF-α) is likely the source of the amyloid deposits seen in Hem1^−/− liver, kidney, and heart (Fig. 2 G and Fig. S6). Although the exact mechanism is unclear, the increase in Th17 effector cells in Hem1^−/− mice may be caused by the following: impaired neutrophil migration into tissues combined with defective phagocytosis of tissue neutrophils by macrophages, which could result in increased production of IL-23 from tissue macrophages (Fig. S3) (50); increased Th17 differentiation from naive T cells (Fig. 6 and Fig. S7) in response to increased IL-6 (Fig. 6 and Fig. S6) and ICOS (Fig. S7); and/or increased production of IL-17–producing memory T cells in response to lymphopenia and impaired innate immunity.

Neutrophils from Vav1/2-deficient mice also exhibit defective adhesion, migration, and phagocytosis (51). Similarly, neutrophils from Rac2-deficient mice (7, 52), as well as a human patient with a dominant-negative RAC2 mutation (8), exhibit impaired neutrophil migration and phagocytosis. Although Hem1^−/− neutrophils exhibit normal Rac activation, components of the Rac-effector WAVE complex are severely reduced after loss of Hem1. These results are consistent with Hem1 acting downstream of Rac to stimulate actin polymerization in neutrophils, perhaps by stabilizing the WAVE complex. However, although macrophages from Hem1^−/− mice are also defective in their ability to phagocytose bacteria, they produce normal or increased amounts of the proinflammatory cytokines and chemokines in response to toll-like receptor stimulation. Hence, as we observed in T cells, our results suggest that NF-κB–mediated transcription of proinflammatory cytokines proceeds independently of Hem-1 in phagocytes. The imbalance between a functional proinflammatory cytokine arm and dysfunctional effector arm in both lymphocytes and phagocytes from Hem1^−/− mice may explain the unique pathology after loss of Hem1. Altogether, our results suggest that Hem1 is essential for innate immune cell function and homeostasis by regulating F-actin polymerization.

Rac1 and Rac2 also regulate many aspects of HSC biology, including adhesion, migration, mobilization, and engraftment. Indeed, we also find that Hem1 deficiency results in decreased colony-forming activity in the BM and increased colony-forming activity in the spleen and peripheral blood, and Hem1^−/− Lin⁻ cells compete poorly with WT Lin⁻ cells after BM transplantation. These results are consistent with decreased adhesion of HSC/Ps in the BM niche, resulting in their release (mobilization) into peripheral blood and spleen of Hem1^−/− mice (10), as is seen in Rac-deficient mice (11, 53). Additional experimentation will be required to definitively determine whether Hem1 also regulates HSC survival and proliferation.

Hem family proteins are highly conserved across Phyla. Disruption of specific Hem family orthologues in D. discoideum (24), D. melanogaster (54), and C. elegans (22) reveal a common theme of defective development, failed migration of embryonic tissues, and disrupted morphogenesis, all of which correlate with impaired actin reorganization. Similarly, the recent disruption of Nap1 (Hem2) in mice results in early embryonic lethality caused by impaired neuronal differentiation and outgrowth, whereas premature retroviral expression of Nap1 promotes postmigratory differentiation (26). Our results indicate that Hem1 is the key Hem family member in hematopoietic cells, where it controls lymphocyte development and activation, as well as phagocyte function. The finding that Hem1 is relatively restricted to hematopoietic cell types, and that it plays a key role in actin reorganization (55) and survival of both normal and transformed cells, suggests that Hem1 might prove to be a valuable drug target in inflammatory diseases or cancer.

Mutagenesis with ENU was performed on C57BL/6J male mice according to the protocol used by Justice et al. (56). Two mating schemes were used. G2 NTB.1 carriers were crossed to C3HeB/FeJ and the resulting F1 animals were intercrossed to generate F2 animals for mapping. In addition, the G1 NTB.1 carrier was mated to C3HeB/FeJ mice and the resulting F1 animals were backcrossed to the G1 carrier to generate N2 mice for mapping. F2 and N2 animals were screened for the NTB.1 trait and DNA was collected. DNA was analyzed with 85 microsatellite markers spaced a mean of 15 cM apart. Linkage analysis was performed using MapManager QTX13b software (K. Manly, Roswell Park, Buffalo, NY). The ∼360-kb NTB.1 candidate interval was narrowed between 59.8 and 60.1 cM on chromosome 15. Computational analysis identified 11 genes, whose exons were PCR amplified from genomic DNA and sequenced.

Hem-1 mutant mice were housed under specific pathogen-free conditions. Mouse procedures were approved by the Institutional Animal Care and Use Committees of Celltech R&D, Inc., and the University of Washington. The NTB.1 line was screened and maintained by DNA sequencing after amplification with Hem1 forward (5′-TGCAGAGCTTCTGTTCCTGTTGGA-3′) and reverse (5′-AGAAGGAACTGCTGGCTACCCAAA-3′) primers and by flow cytometry comparing the ratio of B220^low to Gr-1^high cells. The majority of the phenotypic analysis of Hem1^−/− mice was performed on intercrosses between mice backcrossed three generations onto C57BL6J (from the original C57BL6J × C3H hybrids).

Single-cell suspensions of BM cells were cultured in 6-well culture dishes at 2 × 10⁶ cells/ml. Cells were grown with DMEM/20% FBS medium in the presence of IL-3/GM-CSF (28). After 3 d, nonadherent cells were removed, and adherent cells were continuously grown with IL-3/GM-CSF. After an additional 10 d, nonadherent cells were used in the experiments after initiation of the culture. Nonadherent cells were stained with Gr-1/Mac1 and analyzed by flow cytometry.

Blood was obtained from WT and Hem1^−/− mice and diluted 1:1,000 in PBS containing EDTA for the cell counting. 20 × 10⁶ erythrocytes in 100 μl (0.9% NaCl) were dropped into solutions containing either 0.9% saline or a hypotonic solutions to generate final NaCl concentrations of 0.86, 0.74, 0.69, 0.47, or 0.26% NaCl. In brief, erythrocyte samples were slowly dropped (without mixing) into 250-μl NaCl solutions, incubated for 10 min, and then centrifuged for 1 min at 1,300 rpm. The supernatant was analyzed for released hemoglobin by measuring the absorbance at 540 nm. Samples were normalized to a standard curve generated from complete lysis (100% lysis) and zero lysis (0.9% NaCl) for each of the mice (57).

Rag2^−/−γc^−/− mice (Taconic) were reconstituted with Hem1^−/− (KO; CD45.2) or C57BL/6J WT (CD45.1; Jackson ImmunoResearch Laboratories) Lin⁻ cells or mixed populations of Lin⁻ or total BM cells (WT:Hem1^−/−, 1:1, 1:10, or 1:20). Lineage⁺ cells were depleted using the Mouse lineage panel kit (BD) according to manufacturer's instructions. Lin⁻ cells (2 × 10⁴ cells/each) were injected i.v. into nonirradiated or 950 rad–irradiated Rag2^−/−γc^−/− recipients. In other experiments, total BM cells depleted of CD3⁺ cells were used for BM chimera studies. Thymus, BM, and spleen cells were collected from the recipient mice 6–20 wk after reconstitution.

The EZ-Detect Rac1 activation kit (Thermo Fisher Scientific) was used to pull down active Rac1. The PAK1-bound active Rac1 was detected by Western blot analysis (52) using α-Rac1 antibody (BD) and visualized using the ODYSSEY infrared imaging system (LI-COR Biosciences) located at the Fred Hutchinson Cancer Research Center (Seattle, WA).

Neutrophils were purified from WT and Hem1^−/− BM using Histopaque 1119 and 1077 gradients (Sigma-Aldrich). Chemotaxis assays were conducted according to a protocol provided by Neuro Probe, Inc. The number of migrated cells per 400× field were counted for three random spots and mean values were displayed on the y axis (6).

Neutrophils or peritoneal macrophages (10⁵ cells) were incubated with fluorescent beads (Polysciences, Inc.) or fluorescent E. coli (Invitrogen) for 0, 2, and 4 h. The intensity of phagocytosed beads or E. coli was then analyzed by flow cytometry.

Cytokines (IL-2, IFN-γ, IL-17, IL-6, IL-1β, and TNF-α) and chemokine (MIP2 and KC) levels were quantified with capture/detection antibody-based ELISA systems (eBioscience; R&D Systems), according to the manufacturer's instructions. IL-17 in sera from WT and Hem1^−/− mice was assessed using a Biosource (Invitrogen) mouse IL-17 antibody bead kit according to the manufacturer's recommendations. In brief, 100 μl of sera were incubated with fluorescent beads conjugated to the capture antibody. After addition of the detection antibody mix and PE-conjugated streptavidin, plates were read on a LUMINEX machine using xMAP technology (QIAGEN).

Flat-bottom Maxisorb 96-well plates were coated with 0.5 μg/ml of human fibronectin (BD) and bovine serum albumin (Sigma-Aldrich). Cells were added to the wells, stimulated with α–CD3-ε, washed, and counted under an inverted microscope as previously described (11).

Thymocytes from WT and Hem1^−/− mice were stimulated with α-CD3/α-CD28 for 24 h, followed by PMA for 15 min. Cells were fixed, permeabilized, and stained with Alexa Fluor 488 phalloidin. F-actin polymerization and actin capping formation were determined by flow cytometry and confocal microscopy after phalloidin–Alexa Fluor 488 and DAPI staining, as previously described (3, 7). Hamster IgG– or α–CD3-ε-coated 5-μm sulfate white microspheres were prepared according to the manufacturer's protocol (Invitrogen).

CD4⁺CD44^lowCD25⁻CD62L⁺ T cells (4 × 10⁴ cells) sorted using FACS Vantage or FACS Aria (BD) were activated with plate-bound 1 μg/ml α-CD3 and irradiated APC (10⁵ cells, 3,000 rads) in the presence of 10 ng/ml IL-12 (R&D Systems), 10 μg/ml α–IL-4 antibody (11B11; eBioscience), and 5 ng/ml IL-2 (R&D Systems) for Th1 differentiation condition, as previously described (4). For Th17 differentiation conditions, different doses of TGF-β (5, 1, 0.2, and 0 ng/ml; R&D Systems), IL-6 (50 ng/ml, R&D Systems), α–IL-4 (10 μg/ml), and α–IFN-γ (10 μg/ml) antibodies were added to culture media (58).

Western blot analyses were performed as previously described (59) using rabbit polyclonal antibodies specific for Hem1 (gift from O.D. Weiner, University of California, San Francisco, San Francisco, CA), β-actin (Santa Cruz Biotechnology, Inc.), Wave2 (Santa Cruz Biotechnology, Inc.), Abi1 (Santa Cruz Biotechnology, Inc.), Abi2, and Sra1 (gifts from A.M. Pendergast [Duke University, Durham, NC] and J.D. Scott [University of Washington, Seattle, WA]), Lck, Zap70 (Millipore), and donkey α–goat IgG HRP (Santa Cruz Biotechnology, Inc.), goat α–rabbit IgG HRP (Bio-Rad Laboratories), goat α–mouse IgG HRP (Bio-Rad Laboratories), 4G10 antibody (Millipore), Phospho-Erk, and total Erk (Cell Signaling Technology) (38).

Tissues were embedded in paraffin and stained with hematoxylin and eosin (Histology Consultation Services, Inc.). CBCs were performed by Phoenix Central Laboratory (Everett, WA).

T cells (10⁶ cells/ml) were labeled with CFSE dye (10 μM; Invitrogen) and stimulated with α-CD3/α-CD28 in the presence of 10 ng/ml IL-2 (R&D Systems) for 4 d. Cell divisions were determined in 1 nM of live TO-PRO-3 (Invitrogen) and negative cells were determined by flow cytometry, as previously described (59). In addition, cells were labeled with ³H thymidine for 18 h, and cell proliferation was determined as previously described (60).

Hem1^−/− and WT tissues (BM, thymus, spleen, and LN) were stained with fluorescent-conjugated antibodies specific for CD8, GR1, B220 (Invitrogen), CD62L, Sca1, cKit, Foxp3, IL-2, IL-17, IFN-γ (eBioscience), CD4, CD3-ε, CD69, CD43, ICOS, IgM, MAC1, GR1, Ter119, CD61, CD45.1, CD45.2, Annexin V, CD44, CD25, β-TCR, γδ-TCR, BP1, and heat stable antigen (BD) and analyzed by flow cytometry as previously described (59, 60). To purify enriched HSC/Ps, Lin⁻ cells were then sorted based on expression of ckit (CD117), Sca1 (Ly-6A/E), and Hoechst 33342 (Invitrogen) to define SP cells (61).

Data were analyzed using the Student's one-tailed t test. P-values < 0.05 were considered as significant values.

Fig. S1 shows that the Hem1 gene is preferentially expressed in lymphoid tissues and HSC. Fig. S2 shows decreased clonogenic hematopoietic cell potential in the BM and increased clonogenic hematopoietic cell potential in the blood and spleen in Hem1^−/− mice. Fig S3 shows that Hem1^−/− peritoneal macrophages produce normal or increased levels of proinflammatory cytokines in response to toll-like receptor signaling. Fig. S4 shows that transfection of Hem1 plasmid rescues F-actin formation in Hem1^−/− granulocytes. Fig. S5 shows that Hem1^−/− BM cells compete poorly with WT BM cells after transfer into irradiated Rag2^−/−/γc^−/− mice. Fig. S6 shows that loss of Hem1 impairs T cell proliferation and survival, whereas proinflammatory cytokine production is increased. Fig. S7 shows that Hem1^−/− mice produce increased IL-17 and exhibit normal or increased NF-κB activation. Fig. S8 shows that components of the WAVE complex are transcribed normally in Hem1^−/− mice, and siRNA knockdown of Hem1 in WT T cells results in decreased amounts of WAVE complex proteins.

We thank Leslie Wilson for assistance with mouse genotyping and colony management; Orion Weiner, Ann Marie Pendergast, and J.D. Scott for generously providing antibodies; and Yansong Gu for assistance with confocal microscopy.

This study was supported in part by National Institutes of Health grant (RO1AI0535468 to B.M.I.) and by funding from Celltech R&D, Inc, and the Department of Comparative Medicine (University of Washington, Seattle, WA).

A potential conflict of interest is that S. Friend is employed by Amnis, Inc., the company which makes the ImageStream instrument used to generate Fig. S7 E. The authors have no further competing financial interests.