Development of Macrophages with Altered Actin Organization in the Absence of MafB

Mice.

We described previously the generation of mafB-deficient mice on a 129Sv-C57BL/6 background and their genotyping by PCR with primers for mafB and gfp, replacing mafB in the knockout allele (4). Congenic C57BL/6SJL (Ly5.1 allele) mice were purchased from the Charles River Laboratory (L'Arbresle, France). All experiments were performed in accordance with institutional guidelines using mice maintained under specific-pathogen-free conditions.

Cell preparations and FACS analysis.

Bone marrow cells were flushed from femurs and tibias in Iscove modified Dulbecco medium (IMDM)-20% fetal calf serum (FCS), and splenocytes and fetal liver cells were obtained by mechanical dissociation of organs. Peripheral blood was obtained by heparinized microcapillary aspiration from decapitated embryos. Cell preparations containing red cells were treated with lysis buffer (0.15 M NH₄Cl, 17 mM Tris, pH 7.2) prior to fluorescence-activated cell sorting (FACS) analysis. For antibody staining cells were resuspended in FACS medium (0.2% bovine serum albumin [BSA] and 0.1% NaN₃ in phosphate-buffered saline [PBS]) at a concentration of 1 × 10⁶ to 1 × 10⁷ cells/ml followed by incubation at 4°C for 20 min with properly diluted fluorochrome- or biotin-labeled monoclonal antibodies, followed by secondary incubation with fluorochrome-coupled streptavidin where required. Flow cytometry was performed on a FACScalibur cytometer (Becton Dickinson, San Jose, CA) with CellQuest (Becton Dickinson) software.

Bone marrow-reconstituted mice.

Fetal liver cells (1 × 10⁶) from E14.5 control and mafB^−/− Ly5.2 embryos were injected into the tail vein of lethally irradiated (900- to 1,000-rad) age- and sex-matched Ly5.1 recipient mice. Irradiation was done at least 4 h before cell transfer, and mice were kept on antibiotics in the drinking water for 4 weeks posttransplantation.

Colony and plasma clot assays.

Fetal liver cells (1.5 × 10⁴) were seeded in 1 ml of IMDM containing 1% methylcellulose supplemented with 10% FCS, 450 μM monothioglycerol (Sigma), 10 μg/ml insulin (Sigma), 50 U/ml penicillin, 50 mg/ml streptomycin, and 2 mM glutamine (all from Invitrogen); 2.5% interleukin 3 (IL-3)-containing conditioned medium; 5% granulocyte-macrophage colony-stimulating factor (GM-CSF)-containing conditioned medium; and 5% M-CSF-containing conditioned medium (CA medium). At day 11, differentiated cells were washed out of methylcellulose and stained as described below for cytospins.

In plasma clot assays the morphology of the colony can be preserved during fixation and staining. Cells (10⁴ to 10⁵) from mouse fetal spleen were seeded directly in plasma clot media: CA medium and 10% bovine citrated plasma (Sigma) clotted by the addition of 10 U/ml of thrombin (Sigma). Assays were analyzed at day 6, and clotted colonies were compressed onto slides using Whatman paper and stained as described below for cytospins.

Cytospins, blood smears, and morphological staining and immunostaining.

Cytocentrifugation of cells was carried out using 10 to 100,000 cells per well in a Shandon cytocentrifuge. Samples were air dried and fixed in 100% methanol for 4 min, followed by 2 min of incubation in neutral benzidine (1% O-dianisidine in methanol) to stain erythrocytes. Slides were then rinsed in hydrogen peroxide (0.5% in 50% ethanol) for 1.5 min followed by a 30-s rinse in water. Leukocytes were stained with Diff-Quick (Baxter), 3 min in eosin G in phosphate buffer, pH 6.6, and 2 min in thiazine dye in phosphate buffer, pH 6.6.

Peripheral blood smears were fixed in 100% methanol for 5 min and stained with Giemsa stain (Sigma Diagnostics; GS-5000) for 30 min at room temperature. Differential leukocyte counting was done on a Leitz DMRBE microscope and images acquired with Act-1 software. Nucleated red cells were counted per 100 leukocytes and used for correcting total leukocyte count. Whole blood was diluted (1:40) in Turk solution (Merck, catalog no. 9277), and total nucleated cells were counted in a Neubauer chamber.

For immunostaining, 18.5-day-postcoitum spleens were fixed in 3% paraformaldehyde for 120 min followed by O.C.T. (Tissue-Tek, The Netherlands) embedding. Cells were stained with biotin-conjugated anti-mouse antibody against F4/80 (eBioscience) and counterstained with the nuclear dye DAPI (4′,6′-diamidino-2-phenylindole; 5 μg/ml for 45 min; Sigma-Aldrich). Photomicrographs were taken with a multifluorescence Zeiss Axioplan 2 microscope and acquired with SmartCapture 2 software.

In vitro-derived primary macrophages.

Control or mafB^−/− fetal liver cells were seeded at 2 × 10⁵ cells/ml on two 15-cm bacterial dishes in complete medium (IMDM-10% FCS; 450 μM monothioglycerol σ; 10 μg/ml insulin σ; 50 U/ml penicillin, 50 mg/ml streptomycin, and 2 mM glutamine [all from Invitrogen]; 2.5% IL-3-containing conditioned medium; 5% GM-CSF-containing conditioned medium; 5% M-CSF-containing conditioned medium). From day 6, medium was changed to M-CSF medium (MM: complete medium without IL-3 and with GM-CSF and M-CSF only). At day 10, adherent cells were scraped off and harvested for RNA or protein extraction or split into new dishes for functional analysis.

Microarray and quantitative real-time reverse transcription-PCR (RT-PCR) analysis.

Total RNA was extracted using Trizol (Invitrogen) reagent from equal numbers of in vitro M-CSF-derived macrophages and digested with DNase I (QIAGEN). RNA was pooled from two MafB^+/+ control and two MafB^−/− macrophage samples derived from independent individual fetal livers and processed by a commercial service to generate probes for Affymetrix HG_U95Av2 chip hybridization with Microarray Suite software. Results from two independent arrays each containing duplicate data points were analyzed with Gene Chip EASI expression analysis software. For single gene expression analysis 1- to 2-μg RNA samples of macrophages derived from individual mafB^+/+ or mafB^−/− fetal livers were reverse transcribed with SuperScript II (Invitrogen) and subjected to quantitative real-time PCR using the SYBR green PCR master mix (PE Applied Biosystems, Foster City, CA), following the manufacturer's instructions. Reactions were performed in a 7500 Fast quantitative real-time PCR detection system (PE-Applied Biosystems, Foster City, CA). Tested expression levels were normalized to hypoxanthine phosphoribosyltransferase (HPRT) expression. Primers used were as follows: Fr1, plus strand, CCGACCGGAAACAGACGTT, and minus strand, AACCCGTGAGCTGTGGGTACT; adenylate cyclase-associated protein (CAP), plus strand, GAAAAGTGCCAACCATTTCCA, and minus strand, CACAGTCCAGGGAGTTCTTGCT; cytoplasmic β-actin, plus strand, GACGGCCAAGTCATCACTATTG, and minus strand, CAAGAAGGAAGGCTGGAAAAGA; fascin 1, plus strand, TCAGTCCTCCTGTTATCCTTACTCATC, and minus strand, CCGTTTTCTCTTGGGTTTCCA; α-actinin-1, plus strand, AGCCAGGAACAGATGAACGAA, and minus strand, CCAACGTGCCGGAGTGAT; c-Maf, plus strand, GGATGGCTTCAGAACTGGCA, and minus strand, AACATATTCCATGGCCAGGG; HPRT, plus strand, AGCCCTCTGTGTGCTAAGG, and minus strand, CTGATAAAATCTACAGTCCATAGGAATGGA.

Western blot analysis.

Proteins were extracted from in vitro-differentiated macrophages. Western analysis was performed as described previously (20). Primary antibodies were rabbit anti-mouse c-Maf (Santa Cruz Biotechnology; sc-7866; 1:500), mouse anti-β-actin (Sigma; A5316; 1:20,000), and mouse anti-β-tubulin 1 (Sigma; T7816; 1:20,000). Secondary antibodies were horseradish peroxidase-coupled goat anti-rabbit immunoglobulin G (Santa Cruz Biotechnology; sc-2054; 1/1,000) and anti-mouse immunoglobulin G (Santa Cruz Biotechnology; sc-2055; 1/1,000).

Macrophage stimulation and assays for NO.

Day 12 in vitro-differentiated macrophages at 10⁶ cells/ml were incubated for 24 h with 100 ng/ml lipopolysaccharide (LPS) (Sigma) and 50 U/ml of gamma interferon (IFN-γ) (mouse recombinant; Sigma). Nitric oxide (NO) production was determined by measurement of nitrite, which forms in a reaction of NO with the culture medium. Cell supernatant was cleared of cell debris by centrifugation, and 200 μl was combined with 600 μl of Griess-Ilovays reagent (Merck) and incubated at room temperature for 10 min as recommended elsewhere (1). Absorbance was measured at 546 nm and compared to a standard curve prepared with sodium nitrite.

Phagocytosis assay.

Fluorescent beads (Molecular Probes; 1 μM; F-8851) were washed once in sterile PBS, resuspended in Dulbecco modified Eagle medium-10% FCS, and sonicated for 10 intervals at 10 seconds each. Twenty-five microliters of bead solution was incubated with day 12 differentiated macrophages in 24-well plates for 2 h. Cells were then extensively washed with PBS, fixed with 1% paraformaldehyde, and analyzed by flow cytometry on a FACScalibur cytometer (Becton Dickinson).

M-CSF stimulation and actin filament staining.

Day 10 in vitro-derived macrophages were plated in MM at 4 × 10⁵ cells/ml on alcian blue-treated coverslips. After 24 h cells were cultured in M-CSF-free MM for 12 h and then restimulated with 5% M-CSF-conditioned medium for 5 min. Cells were fixed in 4% paraformaldehyde-PBS and stained with tetramethyl rhodamine isocyanate-conjugated phalloidin (Sigma; 25 ng/ml) in PBS-0.1% saponin-2% BSA. Peritoneal exudate cells were harvested from adult reconstituted mice. Cells (2 × 10⁶) were plated for 2 h in IMDM-10% FCS, 50 U/ml penicillin, and 50 mg/ml streptomycin on alcian blue-treated coverslips. Coverslips were then washed once in PBS and stimulated with 5% M-CSF-conditioned medium for 5 min. Cells were fixed and stained with tetramethyl rhodamine isocyanate-conjugated phalloidin as above and labeled with biotin anti-F4/80 (eBiosciences; 1/500) in PBS-0.1% saponin-2% BSA and fluorescein isothiocyanate-coupled streptavidin (eBiosciences; 1/500). Photomicrographs were taken with a multifluorescence Zeiss Axioplan 2 microscope and acquired with SmartCapture 2 software.

Listeria infection.

Day 12 in vitro-differentiated macrophages were prestimulated by treatment with 100 ng/ml LPS overnight (16 h). Listeria monocytogenes, strain 10403S cells were cultured overnight in brain heart infusion medium (DIFCO) at room temperature until they were in the exponential growth phase. A serial dilution of bacteria covering a range of 1,000 bacteria per cell to 10 bacteria per cell was added to macrophages in Dulbecco modified Eagle medium-10% FCS (without antibiotics) for 35 min. The wells were rinsed three times with medium, and remaining bacteria were killed by the addition of gentamicin at 5 μg/ml. Actin filaments were stained with phalloidin as described above, and bacterial and cellular nuclei were stained with 10 μg/ml DAPI (Roche) for 10 min. Photomicrographs were taken with an Axiophot fluorescence microscope.

Normal numbers of macrophages in MafB^−/− fetal liver.

We have described previously the generation of MafB-deficient mice by homologous recombination and verified the absence of MafB mRNA and protein expression in these mice (4). Because gene expression data and gain-of-function and dominant-negative experiments (3, 10, 12, 13, 20, 34) indicated a critical role of MafB in macrophage differentiation, we investigated the effect of MafB deletion on macrophage development. Due to the neonatal death of MafB^−/− mice from central respiratory defects, we initially analyzed the fetal liver of E14.5 MafB^−/− embryos as the major site of hematopoiesis at this stage of development. Surprisingly we could detect macrophages in cytocentrifuged preparations of MafB^−/− fetal liver cell suspensions (Fig. 1A and B). Furthermore, we also observed robust development of mature macrophages of typical morphology (Fig. 1C and D) by in vitro differentiation of MafB^−/− fetal liver cells in IL-3, GM-CSF, and M-CSF, cytokines that support myeloid progenitor proliferation and macrophage differentiation. To investigate whether MafB deficiency had quantitative effects on macrophage development or on the relative proportions of different lineages in the hematopoietic system, we analyzed E14.5 fetal liver of MafB^−/− mice by FACS. As shown in Fig. 2A and B, we did not observe any differences in the relative abundance of Mac-1⁺/F4/80⁺ macrophages or other myeloid and lymphoid cell populations between MafB^−/− and wild-type (wt) control embryos of the same litter. The analysis of a larger number of MafB^−/− and wt control embryos from several litters also did not reveal any significant difference in the number of Mac-1⁺/F4/80⁺ macrophages or other myeloid cell populations between MafB^−/− (n = 7) and wt control (n = 7) E14.5 fetal livers (Fig. 2C). Together these observations indicated that normal numbers of fetal macrophages could develop in MafB-deficient embryos.

Normal numbers of macrophages in MafB^−/− spleen and blood at birth.

During late gestation hematopoiesis shifts sequentially from the fetal liver to the spleen and eventually to the bone marrow, which coincides with the generation of adult-type macrophages (7). To test whether macrophage defects might become apparent in MafB^−/− mice only at this stage of development, we analyzed the presence of macrophages in the spleen at E18.5 just prior to birth. As shown in Fig. 3A and C histological staining of cytospin preparations from cell suspensions of E18.5 spleen did not reveal any apparent differences between MafB^−/− and wt control samples and the presence of monocytic and granulocytic cells for both genotypes. Since mature macrophages can survive for a long time in tissues, it could not be formally excluded that the observed monocytic cells were at least in part of embryonic origin. We therefore performed plasma clot assays with E18.5 spleen cells to analyze the presence of myeloid progenitors that could give rise to adult-type macrophages. Both macrophages (Fig. 3B and D) and granulocytes (not shown) developed from MafB-deficient spleen cells after 6 days in GM-CSF, which supports the growth of both cell types. Furthermore immunofluorescence staining with anti-F4/80 antibody on frozen sections of E18.5 spleen revealed normal numbers and tissue distribution of macrophages in MafB-deficient samples (Fig. 3E and F). To further analyze whether quantitative differences between macrophage populations might exist, we analyzed MafB^−/− and wt control E18.5 fetal spleen by FACS. As shown in Fig. 3G, Mac-1/F4/80 profiles did not reveal any differences between MafB^−/− and wt control samples and the quantification of both Mac-1⁺/F4/80⁺ and Mac-1⁺/F4/80⁻ populations did not show any statistically significant differences between MafB^−/− (n = 5) and wt control (n = 11) samples (Fig. 3I).

We also analyzed the blood from newborn mice and found normal monocytes, granulocytes, and lymphocytes in blood smears from MafB-deficient samples (Fig. 3K) as well as similar total leukocyte counts in wt control (5,925 ± 990/μl) and MafB^−/− samples (6,380 ± 930/μl). Furthermore, FACS staining also did not reveal any statistically significant alterations in the relative frequency of both Mac-1⁺/F480⁺ and Mac-1⁺/F4/80⁻ populations between MafB^−/− (n = 5) and wt control (n = 9) samples (Fig. 3H and J).

Together these results indicated that normal numbers of adult-type monocytes and macrophages could develop in MafB-deficient embryos.

Normal numbers of MafB^−/− macrophages in reconstituted adult mice.

Despite the apparently normal development of adult-type macrophages in MafB-deficient late gestational embryos, it could not be excluded that defects may become evident only later during adult life or in specific macrophage populations that could not be analyzed in the embryo. To test this possibility, we reconstituted lethally irradiated mice with MafB^−/− and wt control fetal liver as a source of hematopoietic cells and analyzed macrophage populations in these mice after 8 to 12 weeks, when a complete hematopoietic system had been regenerated. Only mice with quantitative donor reconstitution and less than 2% remaining host contribution to the hematopoietic system were used for analysis. As shown in Fig. 4, we observed normal Mac-1/Gr-1 and Mac-1/F4/80 profiles of myelomonocytic populations in the bone marrow and spleen of MafB^−/− reconstituted mice as well as elicited macrophages in the peritoneum 96 h after thioglycolate injection. No statistically significant differences could be detected between the relative proportions of MafB^−/− (n = 4) and wt control (n = 4) myelomonocytic and macrophage populations in the analyzed organs (Fig. 4C). Together this indicated that also in adult mice MafB deficiency did not result in compromised macrophage differentiation.

Preserved macrophage functions in MafB^−/− macrophages.

To further analyze whether MafB-deficient macrophages might have defects in characteristic macrophage functions, we derived macrophages by in vitro differentiation from fetal liver and subjected them to functional assays. A defining characteristic of macrophages is the ability to phagocytose pathogens and cellular debris. To investigate whether this function might be affected by MafB deficiency, we infected MafB^−/− and control macrophages with L. monocytogenes and stained the cells with phalloidin for actin and with DAPI for cellular and bacterial DNA. As shown in Fig. 5A and B both MafB^−/− and control macrophages had phagocytosed bacteria that had no actin tail (blue arrows), indicative of killed bacteria in the lysosomes, as well as actin-tailed bacteria (red arrows), indicative of bacteria that escaped the phagolysosome and used the cellular actin machinery to move through the host cell, a typical characteristic of L. monocytogenes infection (9). To further quantitatively analyze phagocytic ability, we incubated MafB^−/− and wt control macrophages with fluorescent latex beads and subjected them to FACS analysis. As shown in Fig. 5C and D, no differences were found in the percentage of phagocytic cells or the average of phagocytosed beads per cell. Together this indicated that MafB is not required for the general phagocytic machinery of macrophages. Upon activation by IFN-γ and bacterial cell wall components such as LPS, macrophages secrete large quantities of NO, which is thought to contribute to bacterial killing. As shown in Fig. 5E, we observed the same level of NO production after LPS-IFN-γ stimulation for MafB^−/− and wt control macrophages, indicating that MafB is not required for the signaling and effector mechanisms of this pathway.

Increased expression of c-Maf in MafB^−/− macrophages.

Given the surprising absence of a macrophage differentiation phenotype in MafB-deficient hematopoietic cells, we pursued the hypothesis that MafB function in this process might be compensated for by other, closely related members of the Maf family such as c-Maf, which similarly to MafB can induce monocytic differentiation in human myeloid cell lines (14). As shown in Fig. 6, we indeed observed increased c-Maf expression on both the message (Fig. 6A) and the protein (Fig. 6B) level in MafB-deficient macrophages, suggesting that upregulation of c-Maf might rescue MafB function in macrophage differentiation of MafB^−/− cells.

Increased expression of actin organizing factors in MafB^−/− macrophages.

To reveal potential unsuspected functions of MafB in macrophages, we decided to monitor global changes in gene expression resulting from MafB deficiency that might group directly and indirectly controlled genes into functionally related classes. Therefore, we isolated mRNAs from in vitro-differentiated fetal liver-derived MafB^−/− and wt control macrophages and subjected them to hybridization with Affymetrix microarray DNA chips. Detailed analysis of the expression data revealed the coregulation and increased expression of a group of genes involved in actin organization (Fig. 7A). To verify the altered expression of these genes in MafB-deficient macrophages, we used quantitative real-time RT-PCR with specific primers for β-actin, CAP, formin-related protein 1 (Fr1), fascin, and α-actinin. As shown in Fig. 7B, a significantly increased expression, ranging from 2- to over 12-fold, was observed in MafB-deficient macrophages for all of the analyzed genes, thus confirming the gene array data. Increased expression of β-actin in MafB^−/− macrophages was further confirmed on the protein level by Western blotting (Fig. 7C).

Altered morphology and actin organization in MafB^−/− macrophages.

To identify potential phenotypic consequences from the altered expression of actin-modulating genes, we investigated processes in macrophages known to involve actin metabolism.

When Listeria monocytogenes cells are phagocytosed by macrophages, some bacteria can escape the phagolysosome and rearrange cellular actin to propel themselves through the cell and infect neighboring cells (9). Our studies with L. monocytogenes shown in Fig. 5A and B demonstrated that both MafB^−/− and control macrophages contained actin-tailed bacteria that used the cellular actin machinery to move through the host cell. The altered expression of actin-modulating genes in MafB^−/− macrophages thus did not appear to prevent the typical actin rearrangements induced by L. monocytogenes infection.

Macrophages are also highly motile cells and extend filopodia from the cell body during their movements, a process that equally involves extensive actin rearrangement. At least three of the genes that we found to be upregulated in MafB^−/− macrophages, the formin, fascin, and α-actinin genes, have been shown to participate in the formation of filopodia (30, 31, 33, 35, 36). We therefore carefully analyzed cellular protrusions of phalloidin-stained control and MafB^−/− macrophages at high magnification under the microscope. In contrast to MafB^+/+ control macrophages (Fig. 8A), we observed multiple and frequently branched protrusions on a high proportion of MafB-deficient cells (Fig. 8B). Although control macrophages also had protrusions, they were less prominent, less frequent, and rarely branched. This difference was confirmed by counting straight and branched protrusions per 100 cells on control and MafB^−/− macrophages (Fig. 8C).

Macrophages respond to M-CSF by massive morphological changes involving actin reorganization (29). To test whether this response might be involved in the observed morphological differences of control and MafB^−/− macrophages, we M-CSF starved macrophage cultures overnight (Fig. 8D and E) and analyzed their morphology and actin organization by phalloidin staining after restimulation for 5 min with M-CSF. As shown in Fig. 8F and G, stimulation of both control and MafB^−/− macrophages resulted in dramatic morphological reorganization and spreading on the culture dish, but whereas MafB^+/+ control macrophage were spreading mainly by lamellopodial extensions (Fig. 8F), MafB^−/− macrophages showed multiple filopodial and frequently branched protrusions as observed before (Fig. 8G). Even though MafB^+/+ control cells eventually also developed protrusions at later time points, they were less prominent, less frequent, and rarely branched.

To analyze whether this phenotype was also observed in macrophage populations directly isolated from tissues in vivo, we prepared peritoneal exudate cells from wt control or MafB^−/− reconstituted mice, cultured them for 5 min in M-CSF, and stained them with antibody for F4/80 to detect macrophages and phalloidin to reveal the actin skeleton. As shown in Fig. 8H and I, actin⁺ F4/80⁻ lymphocytes with a typical round morphology were detected in both control and MafB^−/− samples. By contrast, macrophages still had a round morphology in the control samples (Fig. 8H) but had undergone dramatic shape changes in the MafB^−/− samples (Fig. 8I). The macrophages had spread out and formed multiple extensive and often branched protrusions, similar to those observed before with in vitro-differentiated macrophages.

Together these results suggested that the increased expression of actin-modulating genes in MafB^−/− macrophages caused changes in actin organization that altered the rapidity and extent of the formation of cellular, possibly filopodial, protrusions in response to M-CSF.