During bacterial infection, the bone marrow hematopoietic activity shifts toward granulocyte production, which is critical for host defenses. Along with this enhancement of granulopoiesis, the bone marrow also increases its release of hematopoietic precursors. At the present time, little is known about the commitment of hematopoietic precursor cells, including hematopoietic stem cells and progenitors, in this response. To investigate the hematopoietic precursor cell response to bacterial infection, bacteremia was established in Balb/c mice by i.v. injection of Escherichia coli. Bacteremia caused a 10-fold increase in the number of lineage (lin)−c-kit+Sca-1+ cells in the bone marrow. This dramatic expansion of the lin−c-kit+Sca-1+ cell pool resulted from both increased mitosis of these cells and inversion from lin−c-kit+Sca-1− cell phenotype. Lipopolysaccharide, tumor necrosis factor-α, and interleukin-6 were potent factors capable of mediating phenotypic inversion of lin−c-kit+Sca-1− cells. Cells in the expanded lin−c-kit+Sca-1+ cell pool contained more colony-forming unit-granulocyte/macrophage. Mobilization of lin−c-kit+Sca-1+ cells into the circulation was significantly enhanced following bacteremia. These results demonstrate that the lin−c-kit+Sca-1+ cell population in the bone marrow constitutes a key component of the host defense response to bacteremia. Functional modifications of these primitive hematopoietic precursors are critical for enhancing granulocyte production following bacterial infection.
Disclosure of potential conflicts of interest is found at the end of this article.
Author contributions: P.Z.: conception and design, financial support, collection and/or assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript; S.N. and G.J.B.: conception and design, financial support, manuscript writing, final approval of manuscript; R.S.: collection and/or assembly of data, data analysis and interpretation, manuscript writing; J.E.S.: conception and design, financial support, final approval of manuscript; D.A.W.: conception and design, financial support, data analysis and interpretation, manuscript writing, final approval of manuscript.
Hematopoietic stem cells are functionally defined by their unique capacity for self-renewal and differentiation into all types of mature blood cells . Under normal conditions, the process of hematopoietic stem cell self-renewal, as well as their conversion into lineage-committed progenitors, is tightly controlled to maintain daily blood cell production [1, 2]. Accumulated evidence has shown that the equilibrium of bone marrow hematopoiesis is altered during bacterial infection, whereby production of phagocytes, particularly granulocytes and monocytes, becomes predominant with inhibition of other lineage (lymphoid and erythroid) development [3, , –6]. Studies have shown that in response to bacterial infection, bone marrow generation of granulocytes or polymorphonuclear leukocytes (PMNs) from their precursors is accelerated . The transit time of PMNs through the marrow mitotic (or proliferative) and postmitotic (maturation/storage) pools to blood is significantly shortened in both experimental animals and hospitalized patients with bacterial infections [8, 9]. Since PMNs constitute the first line of phagocytic defense in the systemic circulation, this enhancement of granulopoietic activity is critically important. Failure to develop an adequate granulopoietic response to infection results in increased morbidity and mortality [10, 11]. Despite the biological significance of changes in hematopoiesis during bacterial infection, relatively little is known about the commitment of the upstream hematopoietic precursor cells, including hematopoietic stem cells, in this response.
During bacterial infection, the infected tissues produce large amounts of cytokines, including granulocyte colony-stimulating factor (G-CSF) and CXC chemokines (interleukin [IL]-8 in humans and keratinocyte-derived chemokine [KC] in mice) [12, , –15]. These mediators are potent stimulants for mobilization of hematopoietic precursor cells from the bone marrow into the systemic circulation [16, –18]. The mobilized hematopoietic precursor cells have been shown to play a significant role in the process of tissue repair [19, –21]. At the present time, knowledge concerning the release of hematopoietic precursor cells from the bone marrow in relation to changes in the marrow pool of these cells, as well as its effects on granulopoiesis during bacterial infection, remains limited.
In this study, we investigated the hematopoietic precursor cell response to bacterial infection using a murine model of i.v. challenge with Escherichia coli. Previous studies have shown that C57BL/6 mouse bone marrow lineage (lin)−c-kit+Sca-1+ cells are highly purified (or enriched) hematopoietic stem cells, whereas lin−c-kit+Sca-1− cells contain more committed progenitors [22, 23]. Similarly, in Balb/c and FVB/N mouse strains, colony-forming unit (CFU)-spleen cells express high levels of the Sca-1 antigen . C57BL/6 mice express the Ly6b (Ly6A) isoform of Sca-1, whereas Balb/c mice possess the Ly6a (Ly6E) haplotype. The difference between the native Ly6A and Ly6E antigens consists of two amino acid differences . In contrast to human hematopoietic stem cells that are highly enriched in CD34+ cell population in the bone marrow, mouse long-term repopulating hematopoietic stem cells are CD34-low/negative . Our current observations show that the bone marrow lin−c-kit+Sca-1+ cell population is rapidly expanded following E. coli bacteremia in Balb/c mice. This increase in the number of marrow lin−c-kit+Sca-1+ cells results from both mitosis of these cells and phenotypic inversion from the lin−c-kit+Sca-1− cell population. Cells in the expanded marrow lin−c-kit+Sca-1+ cell pool are functionally activated for colony-forming unit-granulocyte/macrophage (CFU-GM) formation. Mobilization of lin−c-kit+Sca-1+ cells into the circulation was significantly enhanced from 12 to 48 hours after initiation of bacteremia.
Materials and Methods
Male Balb/c mice (7–10 weeks old; Charles River Laboratories, Wilmington, MA, http://www.criver.com) with a body weight of 26.8 ± 0.3 g were maintained on a standard laboratory diet and were housed in a controlled environment with a 12-hour light/dark cycle. Intravenous challenge with live E. coli (E11775; American Type Culture Collection, Rockville, MD, http://www.atcc.org; ∼1 × 106 CFUs in 50 μl of pyrogen-free saline per mouse) or recombinant murine G-CSF (50 μg/kg body weight in 50 μl of 5% dextrose; Amgen, Thousand Oaks, CA, http://www.amgen.com) was given to mice via penile vein injection under isoflurane anesthesia. Control mice were injected with an equal volume of vehicle. The animals were sacrificed at selected time points thereafter as indicated in Results (Figs. 1, Figure 2.–3, 5, Figure 6.–7). In a subgroup of animals, 5-bromo-2-deoxyuridine (BrdU; 1 mg in 100 μl of phosphate-buffered saline [PBS] per mouse; BD Pharmingen, San Diego, http://www.bdbiosciences.com/index_us.shtml) was i.v. administered along with the bacterial challenge. Upon sacrifice, a heparinized blood sample was obtained by cardiac puncture. Plasma was separated and stored at −80°C. Peripheral blood mononuclear cells (PBMCs) were isolated using Lympholyte-Mammal density separation medium (Cedar Lane, Hornby, ON, Canada, http://www.cedarlanelabs.com) and protocols provided by the manufacturer. Femurs and tibias were collected, and bone marrow cells were flushed out with a total volume of 2 ml of PBS containing 2% bovine serum albumin (BSA; HyClone, Logan, UT, http://www.hyclone.com) through a 23-gauge needle. Bone marrow cells were filtered through a 70-μm nylon mesh (Sefar America Inc., Kansas City, MO, http://www.sefar.com). Erythrocytes in isolated PBMCs and bone marrow cell samples were lysed with Purescript RBC lysis solution (Gentra Systems, Valencia, CA, www.gentra.com). After being washed twice with PBS containing 2% BSA, the remaining nucleated cells were quantified under a light microscope with a hemacytometer. The experiments described here were performed in adherence to the NIH guidelines on the use of experimental animals. Approval of the Animal Care and Use Committee of the Louisiana State University Health Sciences Center was obtained prior to initiating these experiments.
Flow Cytometric Analysis
Nucleated bone marrow cells or isolated PBMCs suspended in RPMI 1640 (Invitrogen, Grand Island, NY, http://www.invitrogen.com) containing 2% fetal calf serum (2 × 106 cells in 100 μl of medium) were added with a mixed panel of biotinylated anti-mouse lineage markers (10 μg/ml of each antibody against CD3e [clone 145-2C11], CD45R/B220 [clone RA3-6B2], CD11b/CD18 [Mac-1, clone M1/70], Gr-1 [Ly-6G/Ly-6C, clone RB6-8C5], or TER 119 [clone TER-119]) or isotype control antibodies (clones A19-3, R35-95, A95-1) (BD Pharmingen). Following incubation for 15 minutes at 4°C, phycoerythrin-conjugated streptavidin (10 μg/ml) and 10 μg/ml of each fluorochrome-conjugated anti-mouse c-kit (CD117, clone 2B8), anti-mouse Sca-1 (Ly-6A/E, clone D7), and anti-mouse CD34 (clone RAM34) (BD Pharmingen) or the matched isotype control antibodies (clones A95–1 and R35–95) were added to the incubation system. The samples were further incubated in the dark for 15 minutes at 4°C. The cells were then washed with cold PBS. For measuring BrdU incorporation, the cells were further processed using a BD BrdU Flow Kit (BD Pharmingen). 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 or a SLR-II flow cytometer with FACSDiva software (Becton, Dickinson and Company, San Jose, CA, http://www.bd.com). In each sample, 500,000 cells were acquired for analysis.
Sorting of lin−c-Kit+Sca-1+ and lin−c-kit+Sca-1− Cells
Pooled nucleated bone marrow cells or PBMCs were suspended in StemSpan serum-free medium (StemCell Technologies, Vancouver, BC, Canada, http://www.stemcell.com). 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 bone marrow or circulating lin−c-kit+Sca-1+ and lin−c-kit+Sca-1− cells was performed on the FACSAria flow cytometer with FACSDiva software. The purity of sorted cell population was 97%–100%.
In Vitro Culture of Bone Marrow lin−c-Kit+Sca-1− Cells
Sorted marrow lin-c-kit+Sca-1− cells from normal mice were plated into a 96-well tissue culture plate with 5 × 104 cells per well in a total volume of 100 μl of StemSpan serum-free medium. The cells were stimulated with the following treatments: (a) 50% plasma from mice that received i.v. saline for 6 hours; (b) 50% plasma from mice that received i.v. E. coli for 6 hours; (c) 15 ng of murine G-CSF (Amgen); (d) a cocktail of growth factors (1,000 ng of heparin, 1 ng of murine stem cell factor [SCF], 2 ng of murine thrombopoietin [TPO], 2 ng of murine insulin-like growth factor-II [IGF-II], 1 ng of human fibroblast growth factor-1 [FGF-1] [R&D Systems Inc., Minneapolis, http://www.rndsystems.com], and 15 ng of murine G-CSF); (e) 10 μg of lipopolysaccharide (LPS; E. coli 0111:B4; List Biological Laboratories, Inc., Campbell, CA, http://www.listlabs.com); (f) 10 ng of murine tumor necrosis factor-α (TNF-α; Biosource International, Inc., Camarillo, CA, http://biosource.com); (g) 10 ng of murine IL-6 (Biosource International); (h) 25 ng (∼200 U) of murine interferon-γ (IFN-γ; Biosource International); and (i) a combination of 10 ng of murine TNF-α, 10 ng of murine IL-6, and 25 ng of murine IFN-γ. Cells in control wells were not exposed to any stimulants. The cells were incubated at 37°C in an atmosphere of 5% CO2 for 24 hours. At the end of culture, cells were stained with fluorochrome-conjugated anti-mouse c-kit and anti-mouse Sca-1 antibodies. Flow cytometric analysis of live (propidium iodide-negative) cells was conducted on an LSR-II flow cytometer with FACSDiva software (Becton Dickinson).
CFU assays of freshly sorted bone marrow or circulating lin−c-kit+Sca-1+ and lin−c-kit+Sca-1− cells, as well as lin−c-kit+Sca-1+ cells derived from inversion of marrow lin−c-kit+Sca-1− cells following in vitro culture for 24 hours with TNF-α (100 ng/ml) or IFN-γ (250 ng/ml), were performed by culturing the cells on Methocult GF M3434 and Methocult GF M3534 media (StemCell Technologies), respectively. One milliliter of Methocult GF M3434 or Methocult GF M3534 medium containing 100 bone marrow lin−c-kit+Sca-1+ or lin−c-kit+Sca-1− cells or 30 circulating lin−c-kit+Sca-1+ or lin−c-kit+Sca-1− cells was plated to a 35-mm Nunclon dish (Nunc, Roskilde, Denmark, http://www.nuncbrand.com). The cultures were conducted for 7 days at 37°C in an atmosphere of 5% CO2. Colonies containing 50 or more cells were then enumerated.
Luminex Assay and Enzyme-Linked Immunosorbent Assay of Plasma Cytokines
Plasma concentrations of cytokines, including TNF-α, IFN-γ, IL-1α, IL-1β, IL-3, IL-6, G-CSF, granulocyte macrophage colony-stimulating factor (GM-CSF), KC, FGF, and vascular endothelial growth factor (VEGF), were measured using a Mouse Cytokine Twenty-Plex kit (Biosource International). All luminex assays were performed on the Bio-Plex Protein Array System (Bio-Rad, Hercules, CA, http://www.bio-rad.com). Plasma concentrations of SCF, TPO, and IGF-II were measured using the Quantikine Mouse SCF and TPO enzyme-linked immunosorbent assay (ELISA) kits and the DuoSet Mouse IGF-II ELISA kit (R&D Systems).
Data are presented as mean ± SEM. Sample sizes are indicated in the figure legends. Statistical analyses of data were conducted using the unpaired Student t test (for comparison between two groups) or one-way analysis of variance followed by the Student-Newman-Keuls test (for comparisons among multiple groups). Differences were considered statistically significant at p < .05.
Alteration of lin−c-Kit+Sca-1+ and lin−c-Kit+Sca-1− Cell Pools
To examine changes in the bone marrow hematopoietic precursor cell populations following systemic infection with E. coli, bone marrow cells were analyzed by flow cytometry on the basis of their phenotypic surface markers. As shown in Figure 1, the lin−c-kit+Sca-1+ cell population in the bone marrow of control mice was very small. Similar observations have been reported in C57BL/6 mice previously . However, the number of lin−c-kit+Sca-1+ cells in the bone marrow was markedly increased following E. coli infection. At 12 hours post-i.v. E. coli, the lin−c-kit+Sca-1+ cell population was increased sevenfold in the bone marrow. Between 24 and 48 hours after the infection, the bone marrow lin−c-kit+Sca-1+ cell population was further expanded to approximately 10 times the control values. The changes in the number of lin−c-kit+Sca-1+CD34− cells were in parallel with the alterations of lin−c-kit+Sca-1+ cell population in the bone marrow following E. coli challenge. In contrast to this increase in the lin−c-kit+Sca-1+ population, the number of lin−c-kit+Sca-1− cells was reduced significantly following E. coli infection. This cell population decreased to approximately 40% of control value at 12 hours post-i.v. E. coli. Between 24 and 48 hours after E. coli infection, the bone marrow lin−c-kit+Sca-1− cell population gradually recovered but remained significantly lower compared with the control values. Similarly, the number of lin−c-kit+ cells (lin−c-kit+Sca-1+ plus lin−c-kit+Sca-1− cells) in the bone marrow was reduced at 12 hours after the E. coli infection. The decrease in these cells was no longer evident by 24 hours. By 48 hours post-i.v. E. coli challenge, the bone marrow lin−c-kit+ cell population was moderately increased compared with the controls.
Alteration of Cell BrdU Incorporation
To understand the mechanisms underlying this expansion of marrow lin−c-kit+Sca-1+ cell pool following bacteremia, an in vivo BrdU incorporation technique was used to determine the activities of hematopoietic precursor cell proliferation. At both 12 and 24 hours after E. coli infection, the percentage of BrdU+ cells in the marrow lin−c-kit+Sca-1+ cell population was significantly reduced compared with those of control mice (Fig. 2), although the absolute number of lin−c-kit+Sca-1+BrdU+ cells in bone marrow was increased in mice with E. coli infection in comparison with the controls. In contrast, the percentage of BrdU+ cells in the marrow lin−c-kit+Sca-1− cell population was moderately increased in mice at 12 and 24 hours post-E. coli infection. However, the absolute number of lin−c-kit+Sca-1-BrdU+ cells in the bone marrow was reduced in mice with E. coli infection at 24 hours in comparison with controls.
The Plasma Cytokine Response
Cytokines are essential in mediating the alteration of hematopoietic activity in response to infection and inflammation. Several cytokines have also been shown to modulate Sca-1 expression by various cell types [27, –29]. Therefore, we measured a panel of related plasma cytokines and growth factors during the early stage of E. coli infection. At 3 and 6 hours post-i.v. challenge, the plasma concentrations of FGF, GM-CSF, IFN-γ, IGF-II, IL-1α, IL-1β, IL-3, SCF, TPO, and VEGF were very low. No differences in these cytokine levels were observed between control and E. coli groups (data not shown). At 3 hours following E. coli bacteremia, the plasma level of G-CSF was markedly increased compared with the control value (Fig. 3). The G-CSF concentration in the plasma was further increased at 6 hours after E. coli infection. IL-6 and TNF-α levels in the plasma were also increased at 3 hours post-i.v. E. coli challenge. This increase in plasma IL-6 and TNF-α concentrations was attenuated at 6 hours following E. coli bacteremia. At both 3 and 6 hours after i.v. E. coli challenge, KC concentrations in the plasma were significantly increased compared with controls.
Inversion of Phenotype Following In Vitro Culture of lin−c-Kit+Sca-1− Cells
The in vivo BrdU data suggest that the observed increase in bone marrow lin−c-kit+Sca-1+ cells during E. coli bacteremia may not be primarily due to accelerated proliferation of existing lin−c-kit+Sca-1+ cells. Other mechanisms may be involved. Since the expansion of the lin−c-kit+Sca-1+ cell population was associated with a significant reduction in the number of lin−c-kit+Sca-1− cells in the bone marrow following E. coli infection, we hypothesized that lin−c-kit+Sca-1− cells might invert to lin−c-kit+Sca-1+ cells during E. coli bacteremia. Figure 4 shows the phenotypic shift of isolated lin−c-kit+Sca-1− cells following 24 hours of culture with conditioned plasma or different cytokines and growth factors. Incubation of lin−c-kit+Sca-1− cells with 50% plasma of control mice did not cause an increase in c-kit+Sca-1+ cells compared with the background level detected using isotype control antibodies. However, exposure of lin−c-kit+Sca-1− cells to 50% plasma of E. coli-infected mice resulted in a significant increase in the number of c-kit+Sca-1+ cells in the culture system. Culture of lin−c-kit+Sca-1− cells with culture medium alone for 24 hours showed a slight increase in c-kit+Sca-1+ cells in the culture system. Addition of G-CSF or a cocktail of growth factors into the culture system did not cause additional increase in the number of c-kit+Sca-1+ cells during the 24-hour culture period. Interestingly, culture of lin−c-kit+Sca-1− cells with LPS caused a remarkable increase in Sca-1 expression by these cells. Similarly, TNF-α, IL-6, and IFN-γ each caused a marked increase in the number of c-kit+Sca-1+ cells in the culture system. However, IL-6-induced increase in Sca-1 expression in the cultured lin−c-kit+Sca-1− cells was relatively weak in comparison with increases induced by LPS, TNF-α, or IFN-γ. Lin-c-kit+Sca-1− cells cultured with a combination of TNF-α, IL-6, and IFN-γ showed the strongest increase in Sca-1 expression by these cells.
Alteration of CFU Activity in Bone Marrow lin−c-Kit+Sca-1+ and lin−c-Kit+Sca-1− Cells
To test changes in functional activity of marrow lin−c-kit+Sca-1+ and lin−c-kit+Sca-1-cells following E. coli bacteremia, lin−c-kit+Sca-1+ and lin−c-kit+Sca-1− cells sorted from mice at 24 hours post-i.v. saline or E. coli were cultured with Methocult GF M3434 and M3534 media. The Methocult GF M3434 medium supports the growth of multiple lineages, including CFU-granulocyte/erythrocyte/macrophage/megakaryocyte (CFU-GEMM), CFU-granulocyte macrophage (CFU-GM), CFU-granulocyte (CFU-G), CFU-macrophage (CFU-M), and burst-forming unit-erythroid (BFU-E), whereas Methocult GF M3534 supports only the growth of granulocyte and monocyte lineages, including CFU-GM, CFU-G, and CFU-M. As shown in Figure 5A, lin−c-kit+Sca-1+ cells isolated from mice at 24 hours post-i.v. E. coli exhibited a significant increase in CFU activity when cultured on Methocult GF M3434 medium. This increase in CFU formation resulted essentially from an increase in CFU-GM (including CFU-G and CFU-M) activity of these cells. CFU counts of isolated lin−c-kit+Sca-1− cells cultured on either Methocult GF M3434 or Methocult GF M3534 medium did not show any statistically significant differences between control and E. coli bacteremic groups (Fig. 5B).
We subsequently tested CFU activity of lin−c-kit+Sca-1+ cells derived from inversion of lin−c-kit+Sca-1− cells following in vitro culture for 24 hours with TNF-α and IFN-γ. As shown in Figure 5C, lin−c-kit+Sca-1+ cells inverted from lin−c-kit+Sca-1− cells showed a significant increase in CFU activity when cultured on Methocult GF M3434 medium compared with freshly isolated lin−c-kit+Sca-1+ cells from control mice. This increase in CFU formation resulted essentially from an increase in CFU-GM activity of these cells, which is similar to the results of lin−c-kit+Sca-1+ cells freshly isolated from E. coli-infected mice.
Circulating lin−c-kit+Sca-1+ cells sorted from mice 24 hours following E. coli bacteremia also showed a higher CFU-GM activity compared with the control value (Fig. 6). Circulating lin−c-kit+Sca-1+ cells of mice with E. coli bacteremia showed a tendency to have an increase in CFU activity when cultured on the Methocult GF M3434 medium, but this change did not reach statistical significance compared with the control value. Sorted circulating lin−c-kit+Sca-1− cells exhibited minimum CFU activities when cultured on both the Methocult GF M3434 and the Methocult GF M3534 media. No difference in CFU activity in circulating lin−c-kit+Sca-1− cells was observed between control and E. coli-infected groups.
Alteration of Bone Marrow Release of lin−c-Kit+Sca-1+ and lin−c-Kit+Sca-1− Cells
During bacterial infection, the increased production of G-CSF and KC may stimulate the bone marrow to release hematopoietic precursor cells. Since the marrow pools of lin−c-kit+Sca-1+ and lin−c-kit+Sca-1− cells are dramatically altered after infection, this alteration may affect release of these cells from the bone marrow into the systemic circulation. As shown in Figure 7, the mobilization of lin−c-kit+Sca-1+ cells into the circulation was significantly increased at 12 hours after E. coli infection. This mobilization of lin−c-kit+Sca-1+ cells was further enhanced at 24 and 48 hours. In contrast to the enhanced mobilization of lin−c-kit+Sca-1+ cells, the release of lin−c-kit+Sca-1− cells from the bone marrow into the systemic circulation was reduced during E. coli bacteremia.
In response to bacterial infection, hematopoietic activity in the bone marrow is directed toward granulocyte production, which is critical for enhancing host defenses against invading pathogens. At the present time, relatively little is known about how hematopoietic precursors are modified by this response. Early studies reported by Quesenberry et al. show that intraperitoneal administration of Salmonella typhosa endotoxin causes a rapid reduction of marrow granulocytic progenitor cells as assayed by the in vitro colony-forming cell technique (growing granulocyte and macrophage colonies in soft agar) . This reduction of granulocytic progenitor cells occurs within 20 minutes after endotoxin, reaches a nadir at 6 hours, and then returns to baseline values by 48 hours. In contrast, Barthlen et al. observed an increase in CFU-GM in the bone marrow of mice at 12 hours following the onset of abdominal sepsis . To further clarify changes in hematopoietic precursor cells, we analyzed the bone marrow cells of mice with E. coli bacteremia using flow cytometry on the basis of the phenotypic cell surface markers. The results of our current investigation showed that E. coli infection caused a dramatic reduction in the number of marrow lin−c-kit+Sca-1− cells. These results are in agreement with previous observations of endotoxin-treated mice reported by Quesenberry et al. . Interestingly, the lin−c-kit+Sca-1+ or lin−c-kit+Sca-1+CD34− cell population in the bone marrow was markedly increased following E. coli infection in our current model despite the decrease in the number of lin−c-kit+Sca-1− cells. At 12 hours post-i.v. E. coli challenge, lin−c-kit+Sca-1+ cell population was increased sevenfold in the bone marrow. During the 24–48 hours after the infection, the number of bone marrow lin−c-kit+Sca-1+ cells was further increased to approximately 10-fold. In our other, parallel studies, we have observed that C57BL/6 mice intravenously challenged with live or heat-inactivated E. coli and Balb/c mice intratracheally challenged with Streptococcus pneumoniae also show a rapid increase in the lin−c-kit+Sca-1+ or lin−c-kit+Sca-1+CD34− cell population in the bone marrow (unpublished data). These data suggest that the expansion of the marrow lin−c-kit+Sca-1+ cell pool is likely a fundamental component of the host defense response to severe bacterial infection.
To understand the mechanisms underlying the expansion of the marrow lin−c-kit+Sca-1+ cell pool, we conducted studies using in vivo BrdU incorporation. The results showed that the absolute number of BrdU-positive lin−c-kit+Sca-1+ cells in the bone marrow was increased at both 12 and 24 hours following E. coli infection. This early increase in BrdU incorporation into marrow lin−c-kit+Sca-1+ cells reflects an increase in mitosis among these cells, which may contribute to the observed expansion of lin−c-kit+Sca-1+ cell pool in the bone marrow. However, this increase in mitosis of lin−c-kit+Sca-1+ cells appears not to be the sole event that causes the increase in the marrow lin−c-kit+Sca-1+ cell pool. The percentage of BrdU-positive cells in the marrow lin−c-kit+Sca-1+ cell population was actually reduced at both 12 and 24 hours post-E. coli infection compared with the controls. A significant population of the increased lin−c-kit+Sca-1+ cells in the bone marrow was BrdU-negative, which suggests that one or more mechanisms other than proliferation of these cells may be involved.
Recent studies have shown that engraftable stem cells and progenitors exist in a reversible continuum, the so-called progenitor/stem cell inversions [31, –33]. In an in vitro culture system, it has been observed that while traveling through the cell cycle, stem cells acquire a progenitor phenotype and lose their stem cell phenotype during the late S and early G2 phases. This phenomenon is reversible by the next G1 phase. Colvin et al. have postulated that the progenitors and the stem cells may, in fact, be the same cell in different reversible functional states . In our studies, E. coli bacteremia-induced expansion of the marrow lin−c-kit+Sca-1+ cell pool occurred in conjunction with a rapid reduction of the lin−c-kit+Sca-1− cell population in the bone marrow. A significant number of cells in the expanded marrow lin−c-kit+Sca-1+ cell pool appeared not to be derived from the proliferation of pre-existing lin−c-kit+Sca-1+ cells. We speculate that the majority (if not all) of this increased number of BrdU-negative lin−c-kit+Sca-1+ cells in the bone marrow may be inverted from lin−c-kit+Sca-1− cells. To verify this, purified marrow lin−c-kit+Sca-1− cells were cultured with plasma of bacteremic mice. Our data showed that exposure to plasma of E. coli-infected mice resulted in a significant increase in conversion of these cells to c-kit+Sca-1+ cells, which strongly supports our hypothesis that the inversion of lin−c-kit+Sca-1− cells to lin−c-kit+Sca-1+ cells may contribute to the rapid expansion of the marrow lin−c-kit+Sca-1+ cell pool following E. coli bacteremia.
To further define factors capable of mediating phenotypic conversion of lin−c-kit+Sca-1− cells during bacteremia, a battery of related plasma cytokines and growth factors was determined during the early stage of bacteremia. The results showed that plasma levels of G-CSF were markedly increased in comparison with controls. However, i.v. administration of recombinant murine G-CSF at a dose of 50 μg/kg did not induce any increase in the number of marrow lin−c-kit+Sca-1+ cells following 48 hours of G-CSF treatment (data not shown). In vitro culture of purified lin−c-kit+Sca-1− cells with recombinant murine G-CSF or G-CSF plus other growth factors (SCF, TPO, IGF-II, FGF-1, and heparin) also did not induce any increase in the conversion of these cell to c-kit+Sca-1+ cells. In our in vivo experiments, the plasma levels of TNF-α and IL-6 were increased in the early stage of E. coli bacteremia. Previously, it has been reported that LPS, IL-6, and IFN-γ stimulate Sca-1 expression by lymphocytes and that TNF-α enhances Sca-1 expression by endothelial cells [27, –29]. Therefore, we tested the effects of these factors in our in vitro culture system. The results showed that LPS, TNF-α, and IFN-γ each caused a remarkable increase in the number of c-kit+Sca-1+ cells in the culture system of marrow lin−c-kit+Sca-1− cells. IL-6 also induced a significant increase in Sca-1 expression in these cultured cells. In a parallel study, mice were intravenously challenged with recombinant murine TNF-α (45 μg/kg or 105 U per mouse), IL-6 (45 μg/kg or 104 U per mouse), or IFN-γ (540 μg/kg or 105 U per mouse) in the absence of E. coli infection. The results showed that 24 hours after challenge, TNF-α and IL-6 at the doses administered each caused a twofold expansion of marrow lin−c-kit+Sca-1+ cell population. IFN-γ at the dose administered caused an increase of more than sixfold in the lin−c-kit+Sca-1+ population in the bone marrow (unpublished data). Since IFN-γ concentration was not increased in the plasma during the early stage of bacteremia, it appears that TNF-α, IL-6, and LPS are major mediators potentially responsible for inducing the rapid phenotypic inversion of lin−c-kit+Sca-1+ cells in our current model of E. coli infection.
Previous studies indicate that accelerated differentiation of marrow hematopoietic progenitors into granulocytes and migration of these precursors to the spleen may account for the reduction of the progenitor cell population in the bone marrow during the early stage of endotoxin challenge [30, 34]. Our results suggest that the enhancement of phenotypic inversion to lin−c-kit+Sca-1+ cells may also be responsible for the rapid reduction of the marrow lin−c-kit+Sca-1− cell pool following E. coli bacteremia. It appears that the lin−c-kit+Sca-1− cell population in the bone marrow serves as a functional reserve supporting the marrow responses to bacterial infection in both directions. These cells may function as the direct precursors for differentiation and production of granulocytes. In addition, they support the expansion of the marrow lin−c-kit+Sca-1+ cell pool via expression of Sca-1 antigen. The decrease in the number of marrow lin-c-kit+Sca-1− cells following bacteremia may result from both the accelerated differentiation of these cells into the downstream granulocytic cells and phenotypic conversion of these cells to the lin−c-kit+Sca-1+ cells.
During bacterial infection, acceleration of myeloid lineage differentiation with enhancement of granulocyte production becomes a predominant feature of hematopoiesis [3, –5]. To understand the expansion of the lin−c-kit+Sca-1+ cell pool in relation to the granulopoietic activity in the bone marrow following E. coli infection, isolated bone marrow lin−c-kit+Sca-1+ cells were cultured with Methocult GF M3434 and M3534 media. Our results showed that marrow lin−c-kit+Sca-1+ cells of mice with E. coli bacteremia displayed a significant increase in CFU-GM activity compared with those of control animals. These data suggest that cells in the rapidly expanded lin−c-kit+Sca-1+ pool in the bone marrow following E. coli bacteremia are functionally activated for granulocyte lineage commitment. We also tested CFU forming activity in lin−c-kit+Sca-1+ cells inverted from lin−c-kit+Sca-1− cells following 24 hours of culture in vitro with TNF-α or IFN-γ. These inverted cells showed a marked increase in CFU-GM activity compared with lin−c-kit+Sca-1+ cells freshly isolated from control mice, which supports our in vivo observation. Similarly, sorted circulating lin−c-kit+Sca-1+ cells from E. coli-infected mice showed an increase in CFU-GM activity in comparison with the value of saline-treated control mice.
E. coli bacteremia caused a marked increase in KC level in the systemic circulation. Both G-CSF and KC are potent factors mediating the bone marrow release of hematopoietic stem cells and progenitors in mice [16, –18, 35]. In the present study, we observed that the mobilization of lin−c-kit+Sca-1+ cells into the circulation was significantly increased during bacteremia. In contrast, the number of lin−c-kit+Sca-1− cells in the circulation was reduced in the infected animals. These data indicate that the altered lin−c-kit+Sca-1+ and lin−c-kit+Sca-1− cell pools in the bone marrow affect the release of cells from each pool into the bloodstream. The expanded marrow pool of lin−c-kit+Sca-1+ cells effectively endorses mobilization of lin−c-kit+Sca-1+ cells into the systemic circulation.
Our current observation of the hematopoietic precursor cell response to bacteremia is mainly based on changes of cells identified by their phenotypic markers. Further investigation of the alteration of other functions associated with these phenotypic changes, including competitive repopulating capacity of cells in the expanded marrow lin−c-kit+Sca-1+ cell pool and homing of the mobilized lin−c-kit+Sca-1+ cells in the systemic circulation following bacteremia will provide additional information for understanding the significance of the hematopoietic precursor cell response to bacterial infection. This investigation is currently under way in our group.
The lin−c-kit+Sca-1+ cell population in the bone marrow is rapidly expanded in response to bacteremia. Both enhanced mitosis of these cells and inversion from the lin−c-kit+Sca-1− cell phenotype contribute to the observed increase in the number of lin−c-kit+Sca-1+ cells in the bone marrow. Cells in the expanded marrow lin−c-kit+Sca-1+ cell pool exhibit an increase in CFU-GM activity. In addition, mobilization of lin−c-kit+Sca-1+ cells into the systemic circulation is significantly enhanced along with the expansion of the marrow lin−c-kit+Sca-1+ cell pool. These data suggest that the marrow lin−c-kit+Sca-1+ cell population functions as an important component of the host immune defense response to severe bacterial infection.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
We thank Amy B. Weinberg, Rhonda R. Martinez, Jane A. Schexnayder, and Joseph S. Soblosky for expert technical assistance. We also thank Connie P. Porretta for expert assistance with flow cytometric analyses and cell sorting. This work was supported by Public Health Service Grants AA09803, HL075161, HL073770, and HL76100.