Toll-Like Receptor 4/Stem Cell Antigen 1 Signaling Promotes Hematopoietic Precursor Cell Commitment to Granulocyte Development during the Granulopoietic Response to Escherichia coli Bacteremia

Abstract

In response to severe bacterial infection, bone marrow hematopoietic activity shifts toward promoting granulopoiesis. The underlying cell signaling mechanisms remain obscure. To study the role of Toll-like receptor 4 (TLR4)/stem cell antigen-1 (Sca-1) signaling in this process, bacteremia was induced in mice by intravenous injection of Escherichia coli. A subgroup of animals also received intravenous 5-bromo-2-deoxyuridine (BrdU). In a separate set of experiments, bone marrow lineage-negative (lin⁻) stem cell growth factor receptor-positive (c-kit⁺) Sca-1⁻ cells containing primarily common myeloid progenitors were cultured in vitro without or with E. coli lipopolysaccharide (LPS). In genotypic background control mice, bacteremia significantly upregulated Sca-1 expression by lin⁻ c-kit⁺ cells, as reflected by a marked increase in BrdU-negative lin⁻ c-kit⁺ Sca-1⁺ cells in the bone marrow. In mice with the TLR4 gene deletion, this bacteremia-evoked Sca-1 response was blocked. In vitro, LPS induced a dose-dependent increase in Sca-1 expression by cultured marrow lin⁻ c-kit⁺ Sca-1⁻ cells. LPS-induced upregulation of Sca-1 expression was regulated at the transcriptional level. Inhibition of c-Jun N-terminal kinase/stress-activated protein kinase (JNK) activity with the specific inhibitor SP600125 suppressed LPS-induced upregulation of Sca-1 expression by marrow lin⁻ c-kit⁺ Sca-1⁻ cells. Engagement of Sca-1 with anti-Sca-1 antibodies enhanced the expression of Sfpi1 spleen focus-forming virus (SFFV) proviral integration 1 (PU.1) in marrow lin⁻ c-kit⁺ Sca-1⁻ cells cultured with LPS. Sca-1 null mice failed to maintain the marrow pool of granulopoietic cells following bacteremia. These results demonstrate that TLR4/Sca-1 signaling plays an important role in the regulation of hematopoietic precursor cell programming and their enhancement of granulocyte lineage commitment in response to E. coli bacteremia.

INTRODUCTION

Representing the largest population of phagocytes in the circulation, neutrophilic granulocytes (neutrophils or polymorphonuclear leukocytes [PMNs]) constitute a major component of the innate immune system (1, 2). Like all other types of leukocytes, granulocytes are derived from hematopoietic stem cells (HSCs) (1). Under normal circumstances, the commitment of HSCs to each lineage (lin) development in the bone marrow is tightly controlled to maintain the homeostasis of blood cell production (3, 4). During bacterial infection, the equilibrium of hematopoiesis is altered, whereby granulocyte production becomes predominant, along with inhibition of other lineage (lymphoid and erythroid) development (5, 6). Early investigations have demonstrated that in response to bacterial infection, bone marrow generation of granulocytes from their precursors is accelerated (7). The transit time of PMNs through the marrow mitotic (or proliferative) and postmitotic (maturation-storage) pools to blood is substantially shortened during bacterial infections in both experimental animals and hospitalized patients (8, 9).

Our recent investigations have revealed that, in response to severe bacterial infection, the expression of stem cell antigen-1 (Sca-1, or Ly-6A/Ly-6E) by bone marrow cells, particularly primitive hematopoietic precursor cells and granulopoietic progenitors, is markedly enhanced in mice (10–12). This upregulation of Sca-1 expression correlates with expansion of the marrow primitive hematopoietic precursor cell pool and enhancement of granulocyte lineage development. Disruption of Sca-1 upregulation impairs the granulopoietic response to severe bacterial infection (11–13). At this time, however, the cell signaling mechanisms underlying the programming of primitive hematopoietic precursor cells for their enhancement of granulocyte lineage commitment during the process of host defense against bacterial infection remain unclear.

HSCs and progenitors express a repertoire of different Toll-like receptors (TLRs) and coreceptors (14). TLR4, along with its coreceptors, lymphocyte antigen 96 (MD-2) and CD14, serves as the essential apparatus for cells to sense lipopolysaccharide (LPS), a major cell wall component of Gram-negative bacteria. LPS is a potent stimulant that evokes various inflammatory reactions in different hosts (15, 16). Recent studies by our group and others have shown that primitive hematopoietic precursor cells respond to different humoral factors, including LPS and proinflammatory cytokines (10, 17). During systemic infection with Gram-negative bacteria, however, LPS is the most proximal mediator that induces alteration of cell activities prior to the release of humoral factors by the responding host cells. Our current investigation traced a TLR4/c-Jun N-terminal kinase/stress-activated protein kinase (JNK)/Sca-1/Sfpi1 spleen focus-forming virus (SFFV) proviral integration 1 (PU.1) cascade that promotes primitive hematopoietic precursor cell commitment to myeloid lineage development during the granulopoietic response to Escherichia coli bacteremia. This pathway may represent an important target for basic investigation and, potentially, therapeutic intervention.

MATERIALS AND METHODS

Animals.

Male mice (BALB/c or C57BL/6 strain, 6 to 10 weeks old) were obtained from Charles River Laboratories (Wilmington, MA). Male mice with the TLR4 gene deletion (C57BL/10ScNJ, 7 to 10 weeks) and gender-matched background control mice (C57BL/10ScSnJ, 7 to 10 weeks) were purchased from The Jackson Laboratory (Bar Harbor, ME). Male Sca-1^−/− mice (7 to 10 weeks), considered congenic on the C57BL/6 background, were bred under specific-pathogen-free conditions in the Animal Care Facility of Michigan State University. All animals were housed in specific-pathogen-free facilities with a 12-h light/dark cycle. Approvals from the Institutional Animal Care and Use Committees of Michigan State University and Louisiana State University Health Sciences Center in adherence with National Institutes of Health guidelines were obtained prior to initiation of experiments.

Bacteremia was induced in mice by intravenous (i.v., via penile vein) injection of live E. coli (∼1 × 10⁶ or ∼1 × 10⁸ CFU of E11775 from the American Type Culture Collection, Rockville, MD, in 100 μl of saline/mouse for inducing different severities of infection) under isoflurane anesthesia. Controls were injected with an equal volume of pyrogen-free saline. In a subset of experiments, animals received an i.v. injection of 5-bromo-2-deoxyuridine (BrdU; 1 mg in 100 μl of phosphate-buffered saline [PBS]/mouse) (BD Biosciences, San Diego, CA) at either 24 h or 4 h before termination of the experiment. Animals were sacrificed at different time points as indicated in the figure legends. Upon sacrifice, a heparinized blood sample was obtained by cardiac puncture. White blood cells (WBCs) were quantified under a light microscope with a hemocytometer. Blood smears were prepared on slides. Wright-Giemsa stain was used to perform differential WBC counts. Plasma was separated and stored at −80°C. Peripheral blood mononuclear cells (PBMCs) were isolated using Lympholyte-mammal density separation medium (Cedarlane, Burlington, NC) with protocols provided by the manufacturer. Femurs and tibias were collected, and bone marrow cells (BMCs) were flushed with a total volume of 2 ml of RPMI 1640 medium (Life Technologies, Grand Island, NY) containing 2% bovine serum albumin (BSA; HyClone Laboratories, Logan, UT) through a 23-gauge needle. Bone marrow cells were filtered through a 70-μm nylon mesh (Sefar America, Inc., Kansas City, MO). Contaminating erythrocytes in bone marrow cell samples and isolated PBMC samples were lysed with red blood cell (RBC) lysis solution (Qiagen Sciences, Germantown, MD). After being washed twice with RPMI 1640 medium containing 2% BSA, the remaining nucleated cells were quantified under a light microscope with a hemocytometer.

Preparation of bacteria.

For each experiment, a frozen stock culture of E. coli was added to tryptic soy broth and incubated for 18 h at 37°C in an orbital shaker. Bacteria were collected and washed twice with PBS. A suspension of bacteria in PBS at a concentration of 1 × 10⁹ CFU/ml was prepared based on its optical density at 600 nm. Actual numbers of viable bacteria were verified by standard plate counts of the bacterial suspensions on MacConkey agar plates following overnight incubation at 37°C.

Flow cytometric analysis.

Nucleated bone marrow cells or isolated PBMCs suspended in RPMI 1640 containing 2% BSA (1 × 10⁶ cells in 100 μl medium) were added with a mixed panel of biotinylated anti-mouse lineage marker antibodies (10 μg/ml each of antibody against CD3e [clone 145-2C11], CD45R/B220 [clone RA3-6B2], CD11b [Mac-1, clone M1/70], granulocyte differentiation antigen 1 [Gr1, or Ly-6G/Ly-6C; clone RB6-8C5], or TER-119 [clone TER-119]) or isotype control antibodies (clones A19-3, R35-95, and A95-1) (BD Biosciences). Following incubation for 15 min at 4°C, phycoerythrin (PE)-conjugated streptavidin (10 μg/ml) and 10 μg/ml of each fluorochrome-conjugated anti-mouse antibody for stem cell growth factor receptor (c-kit, or CD117; clone 2B8), Sca-1 (Ly-6A/E; clone D7), and Gr1 (Ly-6G/C; clone RB6-8C5) (BD Biosciences) or the matched isotype control antibodies (clones A95-1 and R35-95) were added into the incubation system. The samples were further incubated in the dark for 15 min at 4°C. The cells were then washed with cold PBS containing 2% BSA. For measuring BrdU incorporation, the cells were further processed using a BD BrdU flow kit (BD Biosciences). At the end of the staining procedure, cells were suspended in 0.5 ml of PBS containing 1% paraformaldehyde. Analysis of cell phenotypes and BrdU incorporation was performed on a FACSAria II or an LSR II flow cytometer with FACSDiva software (Becton, Dickinson, San Jose, CA). Depending on the cell types being analyzed, the number of cells acquired in each sample was in the range of 5,000 to 500,000.

Sorting and culture of bone marrow lin⁻ c-kit⁺ Sca-1⁻ cells.

Pooled nucleated bone marrow cells were suspended in StemSpan serum-free expansion medium (SSSFEM; StemCell Technologies, Vancouver, BC, Canada). The staining procedure for cell surface markers was similar to that described above, except that the amount of antibodies used for each sample was increased proportionally. Sorting of marrow lin⁻ c-kit⁺ Sca-1⁻ cells was performed on the FACSAria flow cytometer with FACSDiva software. The purity of the sorted cell population was 97 to 100%.

For phenotypic analysis, sorted marrow lin⁻ c-kit⁺ Sca-1⁻ cells from normal mice were plated into a 96-well tissue culture plate at 5 × 10⁴ cells per well in a total volume of 100 μl of SSSFEM containing 10% mouse plasma. For determination of specific gene expression, sorted marrow lin⁻ c-kit⁺ Sca-1⁻ cells from normal mice were plated into a 24-well tissue culture plate at 2 × 10⁵ cells per well in a total volume of 0.5 ml of SSSFEM containing 10% mouse plasma. The cells were cultured in the absence or presence of different stimulants, including lipopolysaccharide (E. coli 0111:B4, 1 to 100 ng/ml; Sigma-Aldrich Co. LLC, St. Louis, MO), anti-murine Sca-1 monoclonal antibodies (combination of clones D7 and E13, each of which has a distinct binding site on Sca-1, 5 μg/ml each; BD Biosciences), or isotype control antibodies (BD Biosciences) and the specific JNK inhibitor SP600125 (20 μM; Sigma-Aldrich Co. LLC). The cells were cultured at 37°C in an atmosphere of 5% CO₂ for 24 h. In a subset of cell cultures, BrdU (100 μg/ml) was added into the culture medium 4 h prior to the termination of cell culture. At the end of culture, cells were stained for surface markers and BrdU incorporation using specific fluorochrome-conjugated antibodies as previously described. Flow cytometric analysis of live (propidium iodide-negative) cells was conducted on a FACSAria II or LSR II flow cytometer. For determination of specific gene expression, total RNA was extracted using an RNeasy plus minikit (Qiagen, Valencia, CA) with the procedures recommended by the manufacturer.

Preparation of RNA standard and quantitative real-time RT-PCR determination.

The RNA standards for mouse PU.1 and granulocyte colony-stimulating factor receptor (G-CSFR) were prepared using previously described methods (18, 19) with some modifications. Briefly, total RNA was extracted from nucleated bone marrow cells of naive mice by using TRIzol reagent (Life Technologies, Inc., Grand Island, NY). cDNA was prepared by reverse transcription (RT)-PCR using the enhanced avian reverse transcriptase RT-PCR kit (Sigma-Aldrich Co. LLC) with specific primers for each gene designed to amplify the entire mRNA gene sequence. The forward and reverse primer sequences used for PU.1 were 5′-GGG ATC TGA CCA ACC TGG AGC-3′ and 5′-TAG AGT CCT GGT GGG AGG CAA A-3′, respectively. The forward and reverse primer sequences used for G-CSFR were 5′-TAACTCATCCAAGTTCACCAGGCA-3′ and 5′-CTCCTCTGAATAGGGCAGGGT-3′, respectively. The PU.1 and G-CSFR cDNA products were determined by gel electrophoresis (1% agarose gel containing ethidium bromide) and then extracted from the gel using the QIAex II DNA extraction kit (Qiagen) with the procedures recommended by the manufacturer. Purified cDNA was ligated into the pCR2.1 vector using the original TA cloning kit, K2000-J10 (Life Technologies). Each vector was then transformed into INVaF′ One Shot competent E. coli cells. Selected colonies for each gene were grown overnight in 2 ml of LB broth containing 100 μg/ml ampicillin and kanamycin at 37°C. Plasmid purification was performed using the Wizard plus SV miniprep DNA purification system (Promega Co., Madison, WI) with the protocols recommended by the manufacturer. The plasmid DNA isolated for each gene was analyzed for the presence of the gene in the correct orientation by restriction enzyme digestion and sequence analysis. The identified colonies were cultured overnight in 200 ml of LB broth containing 100 μg/ml ampicillin and kanamycin. Plasmid DNA was isolated using the Qiagen plasmid maxikit (Qiagen) with the protocols recommended by the manufacturer. To obtain the plasmid Sca-1 gene DNA, a loopful of cloned E. coli (catalog no. MGC-6188; ATCC, Manassas, VA) carrying the murine Sca-1 gene in the pCMV-SPORT6 vector was cultured in 100 ml of LB broth containing ampicillin (10 mg/100 ml) at 37°C on an orbital shaker for 18 h. Plasmid DNA was purified using a Qiagen plasmid midikit (Qiagen) with procedures provided by the manufacturer. Sca-1, PU.1, and G-CSFR RNA standards were prepared by in vitro transcription of purified plasmid DNA for each gene using a T7 RiboMAX large-scale RNA production system kit (Promega Co). Purification was performed using the RNase-free Bio-Rad micro Bio-Spin columns with Bio-Gel P-30 in Tris buffer (Bio-Rad Laboratories, Hercules, CA). The standard RNA was quantified spectrophotometrically at 260 nm and tested for DNA contamination using real-time RT-PCR in the absence of reverse transcriptase. The residual DNA copies in the prepared RNA standard were less than 1/50,000. The standard RNA was stored in aliquots at −80°C until assay. For each subsequent real-time quantitative RT-PCR assay, Sca-1, PU.1, G-CSFR, and 18S rRNA (Life Technologies, Grand Island, NY) standard curves (ranging from 10² to 10⁹ copies of Sca-1, PU.1, and G-CSFR RNA standards per reaction mixture volume and 10⁻⁵ to 10¹ ng of 18S rRNA per reaction mixture volume) were generated by serial dilution of stock standard RNA aliquots.

Real-time RT-PCR determination of Sca-1, PU.1, and G-CSFR mRNA expression was performed using a protocol described previously (11, 18, 19). The amplification primers and 6-carboxyfluorescein (FAM)-labeled probes used for determinations were as follows: for Sca-1, forward primer, 5′-GTTTGCTGATTCTTCTTGTGGCCC, reverse primer, 5′-ACTGCTGCCTCCTGAGTAACAC, and probe, 5′-AGCTCAGGGACTGGAGTGTTACCAGTGCT; for PU.1, forward primer, 5′-AAGATTCGCCTGTACCAGTTCC, reverse primer, 5′-TGTCCTTGTCCACCCACCAG, and probe, 5′-TGCTGTCCTTCATGTCGCCGCTGC; and for G-CSFR, forward primer, 5′-CCAGTCTACACCCTACAGATGC, reverse primer, 5′-GATCTAGTTGCTTCTTCTGACACC, and probe, 5′-CATTCGCTCATCTCTGCCTGGATTCTGGA.

These sets of primers and probes were designed using Primer Express software (Life Technologies). The primers and probe for detection of 18S rRNA were purchased from Life Technologies. The Sca-1, PU.1, and G-CSFR mRNA quantities in each sample were determined by comparing their cycle threshold (C_T) numbers with those of each RNA standard curve. The Sca-1, PU.1, and G-CSFR mRNA values were then normalized to the content of 18S rRNA in each sample. The results are expressed as mRNA copies/ng 18S rRNA.

Statistical analysis.

Data are presented as means ± standard errors of the means. The sample size is indicated in each figure legend. Statistical analysis was conducted using the unpaired Student's t test (for comparison between two groups) and one-way analysis of variance (ANOVA) followed by the Student-Newman-Keuls test (for comparisons among multiple groups). Data sets that were not normally distributed were analyzed using nonparametric statistical analysis (Kruskal-Wallis test with a Dunn's multiple comparison test for differences among the groups). Data analyzed with nonparametric statistical analysis are presented as the median values (the center lines in boxes that present the lower and upper quartiles of the data range), with whiskers representing the lower and upper ranges of data. Statistical difference is accepted at a P value of <0.05.

RESULTS

TLR4 in the Sca-1 response to E. coli bacteremia.

To examine the role of TLR4 signaling in the primitive hematopoietic precursor cell response to E. coli bacteremia, bone marrow cells obtained from mice without or with TLR4 gene deletion were analyzed by flow cytometry based on their phenotypic surface markers. As shown in Fig. 1, the lin⁻ c-kit⁺ Sca-1⁺ cell populations (enriched hematopoietic stem and progenitor cell [HSPC] population) in the bone marrow of saline-treated mice were very small (0.19% ± 0.03% in genotypic background control mice and 0.15% ± 0.02% in mice with TLR4 gene deletion, respectively). This observation was similar to previously reported values in adult C57BL/6 mice (20). However, the number of lin⁻ c-kit⁺ Sca-1⁺ cells in the bone marrow increased markedly following E. coli bacteremia. At 24 h after i.v. injection of 1 × 10⁶ CFU E. coli, the lin⁻ c-kit⁺ Sca-1⁺ cell population increased 3.4-fold in the bone marrow of genotypic background control mice. This bacteremia-induced increase in the marrow lin⁻ c-kit⁺ Sca-1⁺ cell population was not observed in mice with TLR4 deletion. Our previous investigations have shown that both an increase in lin⁻ c-kit⁺ Sca-1⁺ cell proliferation and phenotypic conversion of lin⁻ c-kit⁺ Sca-1⁻ cells to lin⁻ c-kit⁺ Sca-1⁺ cells by reexpression of Sca-1 contribute to the expansion of the marrow lin⁻ c-kit⁺ Sca-1⁺ cell pool during host response to severe bacterial infections (10, 11). To further define the role of TLR4 signaling in mediating Sca-1 expression by primitive hematopoietic precursors in response to bacteremia, we used an in vivo BrdU incorporation technique. At 24 h after E. coli infection, the percentage of BrdU-negative cells in the marrow lin⁻ c-kit⁺ Sca-1⁺ cell population of genotypic background control mice was significantly increased compared to that in saline-treated animals. This increase in BrdU-negative lin⁻ c-kit⁺ Sca-1⁺ cells primarily represented the activation of lin⁻ c-kit⁺ Sca-1⁻ cell phenotypic conversion to lin⁻ c-kit⁺ Sca-1⁺ cells by expression of Sca-1. In mice with TLR4 gene deletion, no increase in BrdU-negative lin⁻ c-kit⁺ Sca-1⁺ cells was observed in the bone marrow following bacteremia. In genotypic background control mice, bacteremia also caused a significant increase in BrdU-positive lin⁻ c-kit⁺ Sca-1⁺ cells in the bone marrow. TLR4 gene deletion impaired this proliferative response of marrow lin⁻ c-kit⁺ Sca-1⁺ cells to systemic E. coli infection.

Fig 1.

(A) The effects of TLR4 gene deletion on the lin⁻ c-kit⁺ Sca-1⁺ cell response to 24 h of bacteremia (i.v. challenge with 1 × 10⁶ CFU E. coli/mouse). Control, C57BL/10ScSnJ background control mice; TLR4 KO, C57BL/10ScNJ TLR4 knockout (KO) mice; Saline, i.v. injection of saline; E. coli, i.v. challenge with E. coli. n = 4 mice per group. #, P < 0.05 compared to other groups. (B) Representative flow cytometric plots. I and II, gates for bone marrow cells and lineage-negative bone marrow cells (8.5% of total nucleated bone marrow cells), respectively. (C) Representative flow cytometric plots of changes in lin⁻ c-kit⁺ Sca-1⁺ cells in bone marrow cells. I, cells from C57BL/10ScSnJ background control mice that received i.v. injection of saline; II, cells from C57BL/10ScSnJ background control mice that received i.v. injection of E. coli; III, cells from C57BL/10ScNJ TLR4 KO mice that received i.v. injection of saline; IV, cells from C57BL/10ScNJ TLR4 KO mice that received i.v. injection of E. coli. (D) Representative flow cytometric plots of changes in BrdU incorporation into lin⁻ c-kit⁺ Sca-1⁺ cells. I, cells from C57BL/10ScSnJ background control mice that received i.v. injection of saline; II, cells from C57BL/10ScSnJ background control mice that received i.v. injection of E. coli; III, cells from C57BL/10ScNJ TLR4 KO mice that received i.v. injection of saline; IV, cells from C57BL/10ScNJ TLR4 KO mice that received i.v. injection of E. coli.

In vitro Sca-1 expression by lin⁻ c-kit⁺ Sca-1⁻ cells in response to LPS.

LPS, the major cell wall component of Gram-negative bacteria, is a natural ligand of TLR4. To further verify the role of TLR4 signaling in mediating the Sca-1 response in primitive hematopoietic precursor cells, marrow lin⁻ c-kit⁺ Sca-1⁻ cells were sorted and cultured with different concentrations of LPS for 24 h. As shown in Fig. 2, LPS induced a dose-dependent increase in conversion of lin⁻ c-kit⁺ Sca-1⁻ cells to c-kit⁺ Sca-1⁺ cells. This conversion of lin⁻ c-kit⁺ Sca-1⁻ cells to c-kit⁺ Sca-1⁺ cells was driven by LPS-induced upregulation of Sca-1 expression, as reflected by the corresponding increase in the mean channel fluorescence (MCF) of Sca-1 staining in cultured cells.

Fig 2.

(A) Dose response of LPS-induced conversion of lin⁻ c-kit⁺ Sca-1⁻ cells to c-kit⁺ Sca-1⁺ cells. MCF, mean channel fluorescence. n = 5 replicates per group. *, P < 0.05 compared to groups of cells stimulated with 0, 1, and 5 ng/ml of LPS; †, P < 0.05 compared to groups of cells stimulated with 0, 1, 5, and 10 ng/ml of LPS; #, P < 0.05 compared to other groups. (B) Representative flow cytometric plots of LPS-induced conversion of lin⁻ c-kit⁺ Sca-1⁻ cells to c-kit⁺ Sca-1⁺ cells.

Sca-1, PU.1, and G-CSFR gene expression.

To delineate the regulatory mechanism underlying the TLR4-mediated Sca-1 response and the correlation of Sca-1 upregulation with the enhancement of granulocyte lineage development in the bone marrow during host response to systemic bacterial infection, Sca-1 gene expression by bone marrow cells was determined using quantitative real-time RT-PCR analysis. PU.1 is a master granulopoietic transcription factor (21–23). It promotes the expression of many gene products that are critical for granulocyte development and maturation. As shown in Fig. 3A, bacteremia caused a marked increase in Sca-1 mRNA expression by nucleated bone marrow cells, indicating that the Sca-1 response to bacteremia was regulated at the level of gene transcription. In association with the upregulation of Sca-1 gene expression, the expression of PU.1 and G-CSFR mRNA by bone marrow cells was significantly enhanced (Fig. 3B and C). In vitro culture of marrow lin⁻ c-kit⁺ Sca-1⁻ cells with LPS significantly increased Sca-1 mRNA expression by these precursor cells. This provides direct evidence that LPS-TLR4 signaling regulates the Sca-1 response in marrow hematopoietic precursors. Interestingly, LPS-induced upregulation of Sca-1 mRNA expression by cultured lin⁻ c-kit⁺ Sca-1⁻ cells was markedly enhanced by antibodies against Sca-1 (Fig. 3D). Furthermore, Sca-1 engagement with anti-Sca-1 antibodies markedly enhanced PU.1 gene expression by cultured marrow lin⁻ c-kit⁺ Sca-1⁻ cells in response to LPS stimulation (Fig. 3E). These data indicate that Sca-1 facilitates TLR4 signaling in mediating granulopoietic transcription factor expression by primitive hematopoietic precursors in response to infectious stimuli.

Fig 3.

(A, B, and C) Changes in Sca-1, PU.1, and G-CSFR mRNA expression by bone marrow cells following 24 h of bacteremia (i.v. challenge with 10⁸ CFU E. coli/mouse). n = 4 mice per group. *, P < 0.05 compared to control group. (D and E) Sca-1 and PU.1 mRNA expression by marrow lin⁻ c-kit⁺ Sca-1⁻ cells following in vitro stimulation with LPS (100 ng/ml) and anti-Sca-1 antibodies (clones D7 and E13, each at 5 μg/ml). n = 3 replicates per group. *, P < 0.05 compared to control group; #, P < 0.05 compared to other groups.

JNK signaling in LPS-induced Sca-1 response.

Analyzing the promoter region of the Sca-1 gene, we found 20 AP-1 binding sites with an optimization (OPT) value between 0.87 to 0.94 in the sequence from −1 to −13743 bp upstream from the Sca-1 gene transcriptional starting site. c-Jun is the most potent transcriptional activator in the AP-1 family (24). The engagement of LPS with TLR4 activates JNK and consequently enhances the transcriptional activity of c-Jun by phosphorylation of its N-terminal activation domain. We therefore determined whether TLR4-JNK signaling plays an important role in the LPS-induced Sca-1 response in primitive hematopoietic precursors. As shown in Fig. 4, LPS-induced increases in both the percentage of c-kit⁺ Sca-1⁺ cells and MCF of Sca-1 expression by cultured lin⁻ c-kit⁺ Sca-1⁻ cells were significantly inhibited by the addition of the specific JNK inhibitor SP600125 to the cell culture system. These data strongly support a model in which JNK signaling is a critical component in mediating LPS-induced Sca-1 expression by primitive hematopoietic precursor cells. In preliminary experiments for testing experimental methods, we did not observe any effect of SP600125 alone in inducing Sca-1 expression in cultured Sca-1⁻ cells at the current dose. Concerning the time restriction for the fluorescence-activated cell sorting (FACS) assay and the limited number of marrow lin⁻ c-kit⁺ Sca-1⁻ cells that could be obtained by FACS sorting from animals, we did not include the agent control group, in which marrow lin⁻ c-kit⁺ Sca-1⁻ cells were treated with SP600125 alone, in this study.

Fig 4.

(A) Representative flow cytometric plots of the effects of JNK inhibitor SP600125 on LPS (10 ng/ml)-induced Sca-1 expression by cultured lin⁻ c-kit⁺ Sca-1⁻ cells. (B) Effects of JNK inhibitor SP600125 on LPS (10 ng/ml)-induced changes in percentage of c-kit⁺ Sca-1⁺ cells in cultured lin⁻ c-kit⁺ Sca-1⁻ cells. (C) Effects of JNK inhibitor SP600125 on LPS (10 ng/ml)-induced changes in mean channel fluorescence (MCF) of Sca-1 expression by cultured lin⁻ c-kit⁺ Sca-1⁻ cells. Control, cells cultured with medium only; LPS, cells stimulated with LPS; LPS+Inh, cells stimulated with LPS in the presence of JNK inhibitor SP600125. n = 3 to 5 replicates per group. *, P < 0.05 compared to control group; #, P < 0.05 compared to other groups.

TLR4 in LPS-induced lin⁻ c-kit⁺ Sca-1⁻ cell phenotypic change and proliferation.

Myeloid progenitors and granulocyte lineage-committed progenitors are active cells of proliferation. Their proliferative activity is the key determinant for maximal granulocyte production (25) in the bone marrow. We next determined LPS-induced BrdU incorporation in cultured marrow lin⁻ c-kit⁺ Sca-1⁻ cells without or with TLR4 deletion. As shown in Fig. 5A, LPS exposure caused a significant increase in phenotypic conversion of lin⁻ c-kit⁺ Sca-1⁻ cells to c-kit⁺ Sca-1⁺ cells from the genotypic background control mouse origin. The LPS-induced increase in c-kit⁺ Sca-1⁺ cells was blocked in cell culture with cells from TLR4-deleted mice. These in vitro results are in agreement with our in vivo observations of the Sca-1 response to bacteremia (Fig. 1). LPS also stimulated proliferation of c-kit⁺ Sca-1⁺ cells from genotypic background control mice, as detected by the significant increase in BrdU incorporation into these cells (Fig. 5B). Deletion of TLR4 abolished the activation of c-kit⁺ Sca-1⁺ cell proliferation in response to LPS. In our current culture model, approximately half of the cells expressed Gr1 at the end of 24 h of culture, suggesting the granulocyte lineage commitment of cultured hematopoietic precursors. LPS exposure resulted in a significant increase in BrdU incorporation into total Gr1⁺ cells (Fig. 5C) and the Sca-1⁺ Gr1⁺ cell subpopulation (Fig. 5D) of the genotypic background control origin. This LPS-induced increase in BrdU incorporation into Gr1⁺ cell populations was not seen in cells with TLR4 deletion.

Fig 5.

Bone marrow lin⁻ c-kit⁺ Sca-1⁻ cell phenotypic change and proliferation following 24 h of culture with LPS (10 ng/ml). (A) Change in percentage of c-kit⁺ Sca-1⁺ cells. (B) Change in percentage of BrdU⁺ cells in c-kit⁺ Sca-1⁺ cells. (C) Change in percentage of BrdU⁺ cells in Gr1⁺ cells. (D) Change in percentage of BrdU⁺ cells in Gr1⁺ Sca-1⁺ cells. n = 5 replicates per group. *, P < 0.05 compared to cells from background control mice; †, P < 0.05 compared to cells cultured with medium alone. (E) Representative flow cytometric plots. I, cells from genotypic background control mice cultured with medium; II, cells from genotypic background control mice cultured with LPS; III, cells from TLR4 gene-deleted mice cultured with medium; IV, cells from TLR4 gene-deleted mice cultured with LPS.

Changes in marrow granulocyte production during bacteremia.

We next assessed the proliferative activity of cells in the granulocyte lineage in mice challenged with bacteremia. At 24 h after i.v. challenge with 10⁸ CFU E. coli, the number of granulocytes (Gr1⁺ cells) in the bone marrow was significantly reduced (Fig. 6A). This reduction of the marrow granulocyte pool might result from enhanced mobilization of granulocytes into the circulation. Concomitantly, BrdU incorporation into marrow Gr1⁺ cells was increased in response to E. coli bacteremia. Supported by the enhanced proliferation of cells in the granulocyte lineage, bone marrow release of granulocytes into the circulation was markedly increased (Fig. 6B). In particular, the release of newly produced granulocytes (BrdU⁺ Gr1⁺ cells) into the systemic circulation was significantly increased during bacteremia.

Fig 6.

(A) Changes in granulocyte (Gr1⁺) subpopulations in bone marrow cells (BMCs). (B) Changes in granulocyte (Gr1⁺) subpopulations in peripheral white blood cells (WBCs) following 24 h of bacteremia. n = 4 mice per group. *, P < 0.05 compared to control group of the same cell type; #, P < 0.05 compared to other groups of the same cell type.

Sca-1 and maintenance of granulocyte lineage cell pool following bacteremia.

Along the multistage development of granulocytes in the bone marrow, the expression of Gr1 increases with the maturation of granulocytes. Early-stage granulopoietic progenitors express low levels of Gr1, while fully developed mature granulocyte express the highest levels of Gr1 (26). As shown in Fig. 7, wild-type mice were able to maintain the marrow pool of granulopoietic progenitors (Gr1^lo cells) during host response to bacteremia. The number of marrow granulopoietic progenitors in Sca-1 null mice tended to decrease following bacteremia (though the difference did not reach statistical significance). The total granulocyte (Gr1⁺ cell) pool in the bone marrow was significantly decreased in Sca-1 null mice at both 24 h and 48 h of bacteremia.

Fig 7.

Changes in bone marrow total granulocyte (Gr1⁺) and granulopoietic progenitor (Gr1^lo) cell populations in Sca-1 null mice following i.v. challenge with E. coli (10⁸ CFU/mouse). Groups: I, wild-type background control mice that received i.v. saline; II, Scal-1 null mice that received i.v. saline; III, background control mice that received i.v. E. coli; IV, Scal-1 null mice that received i.v. E. coli. n = 4 or 5 mice per group. *, P < 0.05 compared to mice of the corresponding strain that received i.v. saline.

DISCUSSION

The lin⁻ c-kit⁺ Sca-1⁺ cell population is enriched with HSPCs in the bone marrow of normal mice (20). Generally, patterns of surface receptor expression, signaling cascade activation, and regulatory pathway organization in undifferentiated precursors are different from those identified in mature cell types. Toll-like receptors are well known for recognition of microbial components in initiation of innate immune responses. In particular, TLR4 recognizes LPS from Gram-negative bacteria (27) and TLR2 recognizes peptidoglycan and lipoteichoic acid from Gram-positive bacteria (28). Murine HSCs and certain early-stage progenitors express TLRs and their coreceptors (14). These primitive hematopoietic precursors can be activated to expand and commit to myeloid lineage development in response to the specific TLR stimulation. Our previous studies have shown that lin⁻ c-kit⁺ Sca-1⁺ cells play an important role in initiating the granulopoietic response to severe bacterial infection (10–12, 29). The marrow pool of lin⁻ c-kit⁺ Sca-1⁺ cells is rapidly expanded following bacterial infection, which is primarily driven by upregulation of Sca-1 expression in marrow lin⁻ c-kit⁺ cells (10, 11). To define the role of TLR in this response, we examined alterations of marrow lin⁻ c-kit⁺ Sca-1⁺ cell types in mice with and without TLR4 deletion following E. coli bacteremia. Systemic E. coli infection caused a marked increase in the number of lin⁻ c-kit⁺ Sca-1⁺ cells in the bone marrow of genotypic background control mice. The expansion of the marrow lin⁻ c-kit⁺ Sca-1⁺ cell pool was primarily supported by the significant increase in nonproliferating lin⁻ c-kit⁺ Sca-1⁺ (BrdU-negative) cells in the bone marrow. TLR4 gene deletion abolished the marrow lin⁻ c-kit⁺ Sca-1⁺ cell response to E. coli bacteria, particularly the increase in nonproliferating lin⁻ c-kit⁺ Sca-1⁺ cells in the bone marrow. These findings demonstrate that TLR4 signaling is essential for evoking the lin⁻ c-kit⁺ Sca-1⁺ cell response.

Certain inflammatory mediators, including tumor necrosis factor-α (TNF-α), interleukin-6 (IL-6), and interferon-γ (IFN-γ), have been reported to induce Sca-1 expression by hematopoietic cells (10, 17, 30, 31). During E. coli bacteremia, cells activated through LPS-TLR4 signaling in inflammatory tissues produce various mediators. Therefore, we tested whether LPS could directly induce the Sca-1 response in isolated marrow lin⁻ c-kit⁺ Sca-1⁻ cells. Exposure of marrow lin⁻ c-kit⁺ Sca-1⁻ cells to LPS in an in vitro culture system caused an LPS dose-dependent upregulation of Sca-1 expression by these hematopoietic precursors. This LPS-induced upregulation of Sca-1 expression presented as both the increase in phenotypic conversion of lin⁻ c-kit⁺ Sca-1⁻ cells to c-kit⁺ Sca-1⁺ cells and the enhancement of Sca-1 surface protein expression in the cultured cell population. These data provide strong evidence supporting the role of LPS as the most proximal mediator in inducing the Sca-1 response of primitive hematopoietic precursors in the bone marrow during Gram-negative bacteremia.

We further determined changes in Sca-1 gene expression by nucleated bone marrow cells following E. coli bacteremia and by cultured marrow lin⁻ c-kit⁺ Sca-1⁻ cells following exposure to LPS. Our results demonstrate that the Sca-1 response in marrow cells is regulated at the transcriptional level. Sca-1 mRNA expression in nucleated bone marrow cells was markedly increased following systemic infection with E. coli. Similarly, the expression of Sca-1 mRNA by cultured marrow lin⁻ c-kit⁺ Sca-1⁻ cells was significantly upregulated following exposure to LPS. To delineate the intracellular signal cascade mediating the transcriptional activation of Sca-1 gene expression, we analyzed the promoter region of the Sca-1 gene. Twenty AP-1 binding sites with high optimization values were mapped in the sequence from −1 to −13743 bp upstream from the transcriptional start site in the Sca-1 gene. The AP-1 family is composed of members of the Jun and Fos subfamilies, which belong to the basic leucine zipper domain (bZIP) group of DNA binding proteins (32). c-Jun is the most potent transcriptional activator in the AP-1 family (24). The transcriptional activity of c-Jun occurs through phosphorylation of its N-terminal activation domain by JNKs (33). LPS and LPS-containing particles (including intact bacteria) activate JNKs through interaction with TLR4 (34–36). To verify the role of TLR4-JNK signaling in mediating the Sca-1 response of primitive hematopoietic precursors, we employed the specific JNK inhibitor SP600125 in our cell culture model. Indeed, LPS-induced upregulation of Sca-1 expression by cultured marrow lin⁻ c-kit⁺ Sca-1⁻ cells was significant inhibited (though not completely abolished) by specific inhibition of JNK activity in our current study.

Sca-1 is an 18-kDa, glycophosphatidylinositol (GPI)-anchored cell surface protein of the lymphocyte activation protein-6 (Ly-6) gene family (37). At this time, neither natural ligands to Sca-1 nor intracellular signaling pathways directly associated with Sca-1 have been identified. Previous studies from our group have shown that upregulation of Sca-1 expression is associated with a significant increase in granulocyte lineage commitment of hematopoietic precursor cells during severe bacterial infection (10, 12). Our current results demonstrate that the expression of key gene products for myeloid and granulocyte development in the bone marrow is activated in association with the Sca-1 response during bacteremia. Gene expression of the master myeloid transcription factor PU.1 and granulocyte colony-stimulating factor receptor (G-CSFR) by nucleated bone marrow cells was significantly upregulated in conjunction with enhanced Sca-1 gene expression in mice with systemic E. coli infection. To further verify that Sca-1 facilitates TLR4 signaling in mediating the granulopoietic response, bone marrow lin⁻ c-kit⁺ Sca-1⁻ cells were stimulated with LPS in the presence of specific anti-Sca-1 monoclonal antibodies. LPS-induced upregulation of Sca-1 mRNA expression by cultured lin⁻ c-kit⁺ Sca-1⁻ cells was markedly enhanced in the presence of anti-Sca-1 antibodies, suggesting a feed-forward signaling mechanism. Furthermore, the presence of antibodies against Sca-1 enhanced PU.1 gene expression by cultured marrow lin⁻ c-kit⁺ Sca-1⁻ cells in response to LPS stimulation. These results suggest that Sca-1 may function as a receptor or coreceptor. Thus, engagement of anti-Sca-1 antibodies with Sca-1 may activate Sca-1 signaling. In light of this postulate, upregulation of Sca-1 expression likely plays a role in facilitating TLR4 signaling for evoking the granulopoietic response to infectious stimuli. In support of our observation, previous studies have shown that Sca-1 signaling is involved in regulation of the expression of different transcription factor genes by hematopoietic cells and cardiac stem cells (38, 39). In our in vitro experiment, a combination of two different anti-Sca-1 antibodies (clones D7 and E13) was used. The D7 and E13 antibodies each bind to Sca-1 at a distinct position. The purpose for exposing the cells to a combination of both anti-Sca-1 antibodies is to maximize the antibody-mediated stimulation.

In addition to programming primitive hematopoietic precursor cells to promote their commitment toward myeloid lineage development, enhancing the proliferation of committed myeloid and granulopoietic progenitors to maximize granulocyte production is another element of the granulopoietic response to severe bacterial infection. To test the role of TLR4/Sca-1 signaling in this process, we determined the proliferative activity of marrow lin⁻ c-kit⁺ Sca-1⁻ cells cultured in vitro in response to LPS stimulation. LPS induced a marked increase in Sca-1 expression by cultured cells, which was associated with an increase in BrdU incorporation into the converted c-kit⁺ Sca-1⁺ cells. TLR4 deletion abolished these increases. Previously, we have reported that the Sca-1 response penetrates into the committed granulopoietic progenitor compartment in the bone marrow during bacteremia (12, 13). Our current results showed that LPS also increased BrdU incorporation into the Sca-1⁺ Gr1⁺ cell subpopulation and total Gr1⁺ cells in the culture system. The increase in proliferative activity of granulocyte lineage-committed cells was blocked by deletion of TLR4. These data support the important role of TLR4/Sca-1 signaling in the activation of cell proliferation in committed myeloid and granulopoietic progenitors.

Bacteremia caused a significant increase in the number of granulocytes in white blood cells in our experimental model. The increase in the circulating granulocyte fraction resulted from enhanced delivery of these phagocytes from the bone marrow to the systemic circulation. The proliferation of precursors along the granulocyte lineage was enhanced to support this host defense process. BrdU incorporation into cells bearing the Gr1 granulocyte lineage marker was significantly increased in the bone marrow following bacteremia in our current study. Concomitantly, the level of BrdU-positive granulocytes in the circulation was also elevated. Previous studies have shown that HSCs from Sca-1 null mice exhibit reduced potential for myeloid lineage development (40). Furthermore, the activation of cell proliferation and granulocyte development in myeloid and granulopoietic progenitors in response to bacteremia is attenuated with Sca-1 ablation (12, 13). In this study, the number of granulopoietic progenitors expressing a low level of surface Gr1 tended to decrease in the bone marrow of Sca-1 null mice following bacteremia, which was accompanied by a severe reduction of the marrow granulocyte pool following bacteremia. These data further confirmed the role of Sca-1 in sustaining the granulopoietic response to severe bacterial infection.

In summary, our current investigation demonstrates that JNK–AP-1 signaling initiated through TLR4 activation mediates the Sca-1 response in marrow primitive hematopoietic precursors during E. coli bacteremia. Upregulation of Sca-1 expression is critical for hematopoietic precursor cell programming and for enhancing these cells' commitment toward granulocyte lineage development. TLR4/JNK/Sca-1/PU.1 signaling is essential for mediating the granulopoietic response to severe bacterial infection.