NFI-A disrupts myeloid cell differentiation and maturation in septic mice

NFI-A promotes myeloid-derived suppressor cell expansion and implications for sepsis-induced immunosuppression.

Abstract

Mounting evidence supports that sepsis-associated immunosuppression increases mortality. As potential contributors to poor sepsis outcomes, myeloid-derived suppressor cells, which are Gr1⁺ CD11b⁺ innate-immune cell progenitors unable to differentiate and possess suppressive activities, expand dramatically in septic mice by a process requiring increased microRNA-21 and microRNA-181b expression. The inhibition of these microRNAs in vivo in septic mice restores Gr1⁺ CD11b⁺ cell differentiation and maturation and improves survival. Here, we show that during sepsis-induced generation of myeloid-derived suppressor cells, transcription factor nuclear factor 1 A type represses cyclin-dependent kinase inhibitor p21 to arrest differentiation of Gr1⁺ CD11b⁺ cells. Our findings include the following: 1) Gr1⁺ CD11b⁺ myeloid cells from late septic mice genetically lacking nuclear factor 1 A type cannot suppress CD4⁺ T cell proliferation and activation; 2) the reconstitution of nuclear factor 1 A type in microRNA-21 and microRNA-181b-depleted Gr1⁺ CD11b⁺ myeloid-derived suppressor cells inhibits cyclin-dependent kinase inhibitor p21 and restores the immune-suppressor phenotype; 3) ex vivo nuclear factor 1 A type knockdown in Gr1⁺ CD11b⁺ myeloid-derived suppressor cells from late septic mice restores cyclin-dependent kinase inhibitor p21 expression and promotes monocyte and dendritic cell differentiation; and 4) ectopic nuclear factor 1 A type expression in normal Gr1⁺ CD11b⁺ cells generates an immunosuppressive phenotype. We suggest that therapeutically targeting nuclear factor 1 A type during late sepsis might improve survival.

Introduction

Sepsis is initiated by a systemic hyperinflammatory reaction, which shifts within a few days to a protracted anti-inflammatory and immunosuppressive state [1–4]. Whereas early sepsis deaths accompany the early hyperinflammatory phase, recent findings in humans indicate that most sepsis deaths occur during the immunosuppressive state [1, 5]. This shift to an immunosuppressive state is clinically undetectable and may explain why attempts to treat sepsis by an anti-inflammatory approach failed in human clinical trials [6–8]. Sepsis-induced immunosuppression correlates with tolerance of innate immunity cells to stimulation by bacterial toxins, depletion of adaptive immune effector cells by apoptosis, including T cells and dendritic cells, increased production of immunosuppressive cytokines (e.g., IL-10), increased CD25⁺ regulatory T cells, immune exhaustion of CD8⁺ T cells, and Th2 cell polarization [2, 5]. Another feature of sepsis is generation of MDSCs, which have mixed surface marker phenotypes of neutrophils and monocytes. This broad-based immunosuppressive state prolongs the initial sepsis-inducing infection and promotes risk to opportunistic infections [1, 9]. Its persistence also predicts a poor outcome, whereas its resolution correlates with survival [10–12].

With the use of a murine model of polymicrobial sepsis with antibiotic treatment, which progressively develops early and late sepsis responses [13], we observed a dramatic expansion of MDSCs in late septic mice [14]. Other studies have also reported increases in MDSCs during the inflammatory response to sepsis [15, 16]. These immature myeloid cells, mainly progenitors and precursors of monocytes, granulocytes, and dendritic cells, increase in many acute and chronic inflammatory conditions [17–20]. In mice, Gr1⁺ CD11b⁺ MDSCs inhibit T cells and innate immune cells [17, 21]. MDSCs express high levels of arginase 1 and immunosuppressive cytokines IL-10 and TGF-β [19, 21]. They also use NOS to generate NO [17, 22]. As MDSCs are generated under nearly all inflammatory conditions, some suggest that MDSCs may develop as part of the normal myelopoiesis and that mediators produced in response to an infection or a growing tumor may reprogram these myeloid progenitors to become immunosuppressive. However, data also support that Gr1⁺ CD11b⁺ MDSCs reflect arrested differentiation [19, 21] and are a byproduct of emergency myelopoiesis [17, 23]. We found that increased expression of miR-21 and miR-181b contributes to Gr1⁺ CD11b⁺ MDSC expansion/accumulation in septic mice and that disrupting this miRNA-dependent pathway improves survival [14, 24].

In this study, we investigated the mechanism of miRNA-dependent expansion of Gr1⁺ CD11b⁺ MDSCs. We identify myeloid cell differentiation-related transcription factor NFI-A as a critical switch in the path leading to MDSC expansion. miR-21 and miR-181b combine to up-regulate NFI-A expression, which directs a negative path to block myeloid differentiation. We show further that NFI-A alone can generate the Gr1⁺ CD11b⁺ MDSC phenotype in normal/naive Gr1⁺ CD11b⁺ cells.

MATERIALS AND METHODS

Mice

Male BALB/c mice, 6–8 wk old, were purchased from Harlan Sprague Dawley (Indianapolis, IN, USA). The mice were housed in a pathogen-free facility and were acclimated to the new environment for 1 wk before surgery. All experiments were conducted in accordance with U.S. National Institutes of Health guidelines and were approved by the East Tennessee State University Animal Care and Use Committee.

CLP

Polymicrobial sepsis, which emulates general peritonitis occurring in humans [25, 26], was induced by CLP, as described previously [13], a model that creates a prolonged infection with ∼90% mortality over 4 wk. In brief, mice were anesthetized via inhalation with 2.5% isoflurane (Abbott Laboratories, Abbott Park, IL, USA). A midline abdominal incision was made and the cecum exteriorized, ligated distal to the ileocecal valve, and then punctured twice with a 21-gauge needle. A small amount of feces was extruded into the abdominal cavity. The abdominal wall and skin were sutured in layers with 3-0 silk. Sham-operated mice were treated identically, except that the cecum was not ligated or punctured. Mice received (intraperitoneally) 1 ml lactated Ringers plus 5% dextrose for fluid resuscitation. To generate late sepsis, mice were subcutaneously administered antibiotic (Imipenem; 25 mg/kg body weight) or an equivalent volume of 0.9% saline. To establish intra-abdominal infection and approximate the clinical situation of early human sepsis, where there often is a delay between the onset of sepsis and the delivery of therapy [27], injections of Imipenem were given at 8 and 16 h after CLP, which results in high mortality (60–70%) during the late/chronic phase [13]. The presence of early sepsis was confirmed by transient systemic bacteremia and elevated cytokine levels in the first 5 d after CLP. Late/chronic sepsis (after d 5) was confirmed by enhanced peritoneal bacterial overgrowth and reduced circulating proinflammatory cytokines.

miRNA antagomiRs

The synthesis of antagomiRs against miR-21 and miR-181b has been described previously [24]. antagomiRs were injected via the tail vein at doses of 80 mg/kg body weight in 100 μl vol. Injection was performed 48 h after CLP (to allow for initiation of sepsis), and target miRNA levels were measured 24 h after injection.

MDSC isolation and purification

Bone marrow and spleen Gr1⁺ CD11b⁺ MDSCs were harvested from sham and septic mice immediately after mice were euthanized by use of magnetically assisted cell sorting, according to the manufacturer's protocol (Miltenyi Biotec, Auburn, CA, USA). The bone marrow cells were flushed out of the femurs with RPMI-1640 medium (without serum) under aseptic conditions [13]. The spleens were minced in RPMI-1640 medium. A single cell suspension of the bone marrow or spleen was made by pipetting up and down and filtering through a 70 μm nylon strainer, followed by incubation with erythrocyte lysis buffer. After washing, total Gr1⁺ CD11b⁺ MDSCs were purified by subjecting the single cell suspension to positive selection of the Gr1⁺ CD11b⁺ cells by incubating with biotinylated mouse anti-Gr1 antibody (clone RB6-8C5; eBioscience, San Diego, CA, USA) for 15 min at 4°C. After washing, the cells were incubated with anti-biotin magnetic beads for 20 min at 4°C and subsequently applied to the MS column. Purified Gr1⁺ CD11b⁺ cells were then washed and resuspended in sterile saline. Cell purity was determined by flow cytometry and was >90%.

MDSC differentiation

Total purified Gr1⁺ CD11b⁺ cells were cultured for 6 d with complete RPMI-1640 medium in the presence of 10 ng/ml M-CSF (PeproTech, Rocky Hill, NJ, USA), plus 10 ng/ml rIL-4 (eBioscience). The cell phenotypes were analyzed by flow cytometry. In some experiments, a portion of differentiated cells was washed and stimulated for 12 h with 1 μg/ml LPS, and culture supernatants were used for cytokine measurements by ELISA.

CD4⁺ T cell proliferation

A coculture of CD4⁺ T cells and Gr1⁺ CD11b⁺ cells was used to determine the effect of Gr1⁺ CD11b⁺ cells on T cell proliferation. In brief, spleen CD4⁺ T cells were isolated from normal (naive) mice by positive selection by use of magnetic beads (Miltenyi Biotec). Cells were fluorescently labeled with CFSE dye by use of the Vybrant CFDA SE Cell Tracer Kit (Invitrogen Molecular Probes; Life Technologies, Eugene, OR, USA). Cells were incubated for 10 min at room temperature with 10 μM CFSE dye and then cocultured (at a 1:1 ratio) with Gr1⁺ CD11b⁺ cells, which were isolated from the bone marrow of sham or late septic mice. T cell proliferation was induced by the stimulation with an anti-CD3 plus anti-CD28 antibody (1 μg/ml/each). After 3 d, cells were harvested, and CD4⁺ T cell proliferation was determined by the step-wise dilution of CFSE dye in dividing, CD3-gated CD4⁺ cells by use of flow cytometry.

Flow cytometry

MDSCs were analyzed by flow cytometry. Total or differentiated Gr1⁺ CD11b⁺ MDSCs were labeled by incubation for 30 min on ice in staining buffer (PBS plus 2% FBS) with the appropriate fluorochrome-conjugated antibodies. After washing, the samples were analyzed by a FACSCaliber flow cytometer (BD Biosciences, Sparks, MD, USA). Approximately 15,000 events were acquired and analyzed by use of CellQuest Pro software (BD Biosciences). The following antibodies were used: anti-Gr1 conjugated to FITC, anti-CD11b conjugated to PE, anti-F4/80 conjugated to APC, anti-CD11c conjugated to PE, and anti-MHC II conjugated to FITC. An appropriate isotype-matched control was used for each antibody.

miRNA Northern blots

Levels of miR-21 and miR-181b in whole bone marrow cells were measured by use of a miRNA Northern Blot Assay kit, per the manufacturer's instructions (Signosis, Santa Clara, CA, USA). In brief, 2 μg amounts of miRNA-enriched RNA were separated by 15% Tris-borate-EDTA urea gels, transferred to a nylon membrane, and then immobilized by UV crosslinking. Membranes were hybridized overnight at 42°C with a biotin-labeled miRNA probe, complementary to miR-21 and miR-181b. Membranes were blocked for 30 min at room temperature and then incubated with streptavidin-HRP for 45 min at room temperature. After washing, miRNAs were detected with a chemiluminescence substrate, the bands were visualized by use of the ChemiDoc XRS System (Bio-Rad Laboratories, Hercules, CA, USA), and the images were captured with the Image Lab Software, V3.0 (Bio-Rad Laboratories).

Immunoblots

Equal amounts of protein extracts were mixed with 5× Laemmli sample buffer, separated by a SDS-10% polyacrylamide gel (Bio-Rad Laboratories), and subsequently transferred to nitrocellulose membranes (Thermo Fisher Scientific, Waltham, MA, USA). After blocking with 5% milk in TBST for 1 h at room temperature, membranes were probed overnight at 4°C with the appropriate primary antibody. After washing, blots were incubated with the appropriate HRP-conjugated secondary antibody (Life Technologies, Grand Island, NY, USA) for 2 h at room temperature. Proteins were detected with the ECL detection system (Thermo Fisher Scientific). The developed bands were visualized by use of the ChemiDoc XRS System (Bio-Rad Laboratories), and the images were captured with the Image Lab Software, V3.0. Membranes were stripped and reprobed with β-actin antibody (Sigma-Aldrich, St. Louis, MO, USA) as a loading control.

siRNA-mediated knockdown

Purified Gr1⁺ CD11b⁺ MDSCs were transfected with pools of NFI-A-specific, NF-κB p65-specific, or scrambled (control) siRNAs at a 0.5 μM final concentration (Santa Cruz Biotechnology, Santa Cruz, CA, USA) by use of HiPerFect Transfection Reagent, per the manufacturer's instructions (Qiagen, Valencia, CA, USA), and the Gene Pulser MXcell Electroporation System (Bio-Rad Laboratories). After 24 h, cells were differentiated for 6 d with M-CSF plus rIL-4 (as described above) or stimulated for 12 h with LPS to assess cytokine production.

NFI-A expression plasmid

Full-length mouse Nfia cDNA was cloned in a pEZ-M07 plasmid expression vector downstream of the CMV promoter, and NFI-A protein expression was verified by Western blot. An empty pEZ-M07 vector served as a negative control. Plasmid DNA was suspended in HiPerFect Transfection Reagent (final concentration: 0.5 μg/ml) and transfected into Gr1⁺ CD11b⁺ cells (at 0.5 × 10⁶ cells/ml) as described above.

ELISA

Cytokine concentrations were determined by use of specific ELISA kits (eBioscience), according to the manufacturer's instructions. Each sample was run in duplicate.

Statistical analysis

The Kaplan-Meier survival curve was plotted by use of a GraphPad Prism, version 5.0 (GraphPad Software, La Jolla, CA, USA), and survival significance was determined by a log-rank test. Other data were analyzed by use of Microsoft Excel, V3.0, and differences between 2 groups were analyzed by an unpaired Student's t test. One-way ANOVA was used to analyze data with more than 2 groups. All values are expressed as mean ± sd. P ≤ 0.05 was considered statistically significant.

RESULTS

Expression of NFI-A is induced in Gr1⁺ CD11b⁺ MDSCs during sepsis and inhibited by anti-miR-21 and miR-181b antagomiRs

Our model of polymicrobial sepsis, induced by CLP, develops into early and late sepsis phases [13]. With this model, the mortality rate during the early phase (days 1–5 post-CLP) is 20–30%. Mice that became moribund (those suffering deep hypothermia of <34°C, weight loss of ˃30%, and lethargy) are euthanized after d 6 post-CLP and defined as “late septic” [13].

We previously showed that miR-21 and miR-181b expression in septic mouse bone marrow promotes expansion of immunosuppressive Gr1⁺ CD11b⁺ MDSCs [14] and that simultaneous in vivo inhibition of both miRNAs diminishes Gr1⁺ CD11b⁺ cells numbers in the bone marrow and improves survival [24]. Here, we tested whether miR-21- and miR-181b-induced myeloid differentiation-associated factor NFI-A can arrest Gr1⁺ CD11b⁺ myeloid progenitor differentiation and maturation to support Gr1⁺ CD11b⁺ MDSC expansion. Gr1⁺ CD11b⁺ cells increase slightly during early sepsis, but unlike cells from late sepsis, they can differentiate ex vivo and are not immunosuppressive [14]. As MDSCs expand/accumulate dramatically in late sepsis, and our aim was to investigate late sepsis immunosuppression, we performed most of our analyses in late septic mice.

First, we evaluated Gr1⁺ CD11b⁺ cell differentiation and animal survival after in vivo inhibition of miR-21 and miR-181b by antagomiRs. Sepsis was induced by CLP, and 48 h later (to allow sepsis initiation), mice were injected via the tail vein with a mixture of anti-miR-21 and miR-181b antagomiRs or control mutant antagomiRs at doses of 80 mg/kg body weight. Northern blots showed that levels of both miRNAs diminish in the bone marrow Gr1⁺ CD11b⁺ cells, 24 h after the antagomiR injection (Fig. 1A). Bone marrow Gr1⁺ CD11b⁺ cells were isolated by positive selection, differentiated by the stimulation with M-CSF plus IL-4 for 6 d, and then phenotyped for macrophages and dendritic cell markers. Flow cytometry analysis showed that percentages of differentiated cells were significantly higher in septic mice injected with antagomiRs compared with mutant antagomiRs (Fig. 1B and Supplemental Fig. 1A). The miRNA inhibition by antagomiRs also improved the overall survival rate by 76% over a 4 wk period (Fig. 1C). These results support that miR-21 and miR-181b promote MDSC expansion in late sepsis by attenuating differentiation and maturation of the Gr1⁺ CD11b⁺ myeloid progenitors.

Figure 1. The inhibition of miR-21 and miR-181b in septic mice restores Gr1⁺ CD11b⁺ cell differentiation and improves survival.

Sepsis was induced by CLP. After 48 h, mice received miR-21 and miR-181b or mutant (control) antagomiRs (80 mg/kg by tail vein). (A) At 24 h after antagomiR delivery, 5 randomly selected mice/group were euthanized, and miR-21 and miR-181 levels were measured by Northern blots. (B) Isolated bone marrow Gr1⁺ CD11b⁺ cells from moribund late-septic mice and from antagomiR-treated (surviving) mice (euthanized at the same time) were differentiated for 6 d with M-CSF plus rIL-4 (10 ng/ml each). Macrophage (F4/80⁺ CD11b⁺) and dendritic cell (CD11c⁺ MHC II⁺) percent was determined by flow cytometry. (Left) An example of F4/80⁺ CD11b⁺ or CD11c⁺ MHC II⁺ gated cells; (B) percentages of differentiated cells. Data are from 1 of 3 experiments (7 mice/group) and are expressed as means ± sd. *P < 0.05. (C) Sepsis mortality was monitored for 28 d after CLP, and deaths (moribund state) were separated into early (d 1–5) and late (d 6–28) sepsis and expressed by Kaplan-Meir survival analysis. All moribund (death from euthanasia) mice had profound weight loss (>30%), hypothermia (<34°C), and loss of righting reflex.

Transcription factor NFI-A [28] reduces p21 gene expression [29]. p21 is a CDK inhibitor involved in stem cell differentiation [30]. We measured NFI-A and p21 protein levels in Gr1⁺ CD11b⁺ cells isolated from the bone marrow of septic mice. NFI-A protein was not detected in cells from sham mice, but p21 was constitutively expressed (Fig. 2A). In contrast, Gr1⁺ CD11b⁺ cells from septic mice had high levels of NFI-A protein with diminished p21 expression. Importantly, miRNA inhibition by antagomiRs reduced NFI-A expression and restored p21 protein levels (Fig. 2B). miRNAs inhibit target protein expression by binding complementary 3′ UTR mRNA sequences [31]. Sequence inspection of the 3′ UTR region of NFI-A did not reveal any potential binding sites for miR-21 or miR-181b. Together, the results in Figs. 1 and 2 suggest that miR-21 and miR-181b block Gr1⁺ CD11b⁺ cell differentiation and maturation by indirectly promoting NFI-A expression.

Figure 2. miR-21 and miR-181b knockdown reduces NFI-A protein expression and increases p21 levels.

(A) NFI-A and p21 proteins were assessed ex vivo by Western blot by use of isolated and pooled bone marrow Gr1⁺ CD11b⁺ cells (4 mice/group) from sham, early (d 1–5), and late (after d 5) septic mice. Results represent 1 of 2 experiments. (B) The NFI-A protein probed by Western blots was compared in late (d 6–28 post-CLP) septic mice that were moribund and euthanized (mutant antagomiR group) to time-corresponding, antagomiR-treated mice. Results are representative of 3 experiments.

Late sepsis Gr1⁺ CD11b⁺ cells lacking NFI-A lose immunosuppressive function

Gr1⁺ CD11b⁺ MDSCs repress T cell proliferation and activation [21, 32]. Accordingly, we tested whether late sepsis Gr1⁺ CD11b⁺ cells from mice injected with antagomiRs and with diminished NFI-A expression regain differentiation capacity and lose the ability to repress T cells. Gr1⁺ CD11b⁺ cells were isolated from the bone marrow of late septic mice that received antagomiRs or mutant (control) antagomiRs and were cocultured with splenic CD4⁺ T cells isolated from naive (normal) mice. The cultures were stimulated with a cocktail of CD3 and CD28 antibodies, and CD4⁺ T cell proliferation and activation (measured by IFN-γ production) were assessed by flow cytometry and ELISA, respectively. Gr1⁺ CD11b⁺ cells from septic mice injected with mutant antagomiRs inhibited CD34⁺ T cell proliferation and IFN-γ production compared with cells from sham mice (Fig. 3). In contrast, Gr1⁺ CD11b⁺ cells from septic mice with miRNA inhibition (antagomiR group) did not affect CD4⁺ cell proliferation or IFN-γ production. As Gr1⁺ CD11b⁺ cells from the antagomiR-injected group do not express NFI-A (Fig. 2), we evaluated the effect of re-expressing NFI-A on immunosuppressive function. To do this, cells from late septic mice that received antagomiRs were transfected with an NFI-A expression plasmid for 48 h and then cocultured with CD4⁺ T cells. The NFI-A-transfected cells regained their immunosuppressive phenotype and inhibited CD4⁺ T cell proliferation and IFN-γ production (Fig. 3). Thus, NFI-A supports MDSC repressor activity in the absence of miR-21 and miR-181b.

Figure 3. Gr1⁺ CD11b⁺ cells from septic mice lose their immunosuppressive phenotype after miR-21 and miR-181b knockdown and regain it with NFI-A repletion.

Sepsis was induced by CLP, and mice were injected with antagomiRs or mutant (control) antagomiRs, as described in Fig. 1. After 48 h, bone marrow Gr1⁺ CD11b⁺ cells were isolated from late septic (after d 5) mutant antagomiR and antagomiR-treated mice and then transfected with an NFI-A expression plasmid or empty vector for 48 h. To assess immunosuppression by Gr1⁺ CD11b⁺ cells, spleen CD4⁺ T cells were isolated from normal (naive) mice and labeled with the fluorescent dye CFSE for 10 min at room temperature. Gr1⁺ CD11b⁺ cells were then cocultured (at a 1:1 ratio) with target CD4⁺ T cells, and the mixture was stimulated with anti-CD3 plus anti-CD28 antibodies (1 μg/ml each). After 3 d, CD4⁺ T cell proliferation was determined by the step-wise dilution of CFSE dye in dividing CD4⁺ cells by flow cytometry. T cell proliferation is presented relative to the coculture of Gr1⁺ CD11b⁺ cells from sham mice, and values are normalized to 100%. The culture supernatants were analyzed for IFN-γ levels by ELISA. Data are expressed as means ± sd from 7 mice/group (*P < 0.05, compare mutant antagomiRs with antagomiRs; **P < 0.05, compare antagomiRs + vector with antagomiRs + NFI-A) and are representative of 2 experiments.

Next, we investigated the effect of NFI-A knockdown on Gr1⁺ CD11b⁺ cell differentiation, suppressive function, and p21 expression. Gr1⁺ CD11b⁺ cells were isolated from late septic mice, transfected with NFI-A siRNA, and differentiated with M-CSF plus IL-4. NFI-A knockdown restored p21 protein levels and significantly enhanced cell differentiation and maturation of macrophages and dendritic cells (Fig. 4A and B and Supplemental Fig. 1B). In addition, the suppressive function of Gr1⁺ CD11b⁺ cells was diminished after NFI-A knockdown, as they did not significantly inhibit CD4⁺ T cell proliferation or IFN-γ production in the coculture (Fig. 4C). Late sepsis Gr1⁺ CD11b⁺ cells with NFI-A knockdown (Fig. 4A) re-expressed the p21 protein, similar to those from mice that were injected with antagomiRs (Fig. 2), which were able to differentiate ex vivo (Fig. 1B). To determine whether lack of p21 as a result of up-regulation of NFI-A blocks Gr1⁺ CD11b⁺ cell differentiation, we knocked down p21 in cells isolated from septic mice treated with antagomiRs. Knockdown of p21 significantly attenuated Gr1⁺ CD11b⁺ cell differentiation (Fig. 4D). These data show that p21 is required for myeloid differentiation and support that NFI-A links with reduced expression of p21 to promote Gr1⁺ CD11b⁺ MDSC accumulation in late sepsis.

Figure 4. NFI-A depletion in Gr1⁺ CD11b⁺ cells from septic mice restores their differentiation capacity.

Gr1⁺ CD11b⁺ cells were isolated from the bone marrow of late (after d 5) septic mice (as described in Fig. 1) and transfected with NFI-A-specific or control siRNA. After 36 h, cells were then differentiated for 6 d with M-CSF plus rIL-4 (10 ng/ml each). (A) Western blot analysis of NFI-A and p21 proteins is representative of 3 experiments. (B) Flow cytometry was used to analyze macrophages (F4/80⁺ CD11b⁺) and dendritic cells (CD11c⁺ MHC II⁺). (Left) An example of flow cytometry of differentiated cells gated on F4/80⁺ CD11b⁺ or CD11c⁺ MHC II⁺ staining; (right) percentages of differentiated cells. Data are expressed as means ± sd from 6 mice/group (*P < 0.05, compare NFI-A siRNA with control siRNA) and are representative of 3 experiments. (C) Late sepsis Gr1⁺ CD11b⁺ cells were transfected with control or NFI-A siRNA, cocultured with CD4⁺ T cells, and treated as described in Fig. 3. T cell proliferation and IFN-γ production were determined by flow cytometry and ELISA, respectively. Data are expressed as means ± sd from 5 mice/group (*P < 0.05, compare with sham; **P < 0.05, compare with control siRNA) and are representative of 3 experiments. (D) Knockdown of p21 in Gr1⁺ CD11b⁺ cells lacking NFI-A blocks their differentiation. Gr1⁺ CD11b⁺ cells were isolated from the bone marrow of late septic mice that were injected with antagomiRs, which re-express p21 in the absence of NFI-A (see Fig. 2B). Cells were transfected with p21-specific or control siRNA and then differentiated and analyzed as described above. Data are expressed as means ± sd from 5 mice/group (*P < 0.05, compare with control siRNA) and are representative of 2 experiments.

NFI-A switches naive Gr1⁺ CD11b⁺ cells into MDSCs

Next, we determined whether ectopic expression of NFI-A in normal Gr1⁺ CD11b⁺ cells from naive mice changes their phenotype. Gr1⁺ CD11b⁺ cells were isolated from the bone marrow of naive mice, transfected with an NFI-A expression plasmid, and differentiated into macrophages and dendritic cells. Western blot analysis showed high levels of NFI-A protein expression, 36 h after NFI-A plasmid transfection, at which time, p21 protein levels were diminished (Fig. 5A). Cells transfected with a control vector differentiated into macrophages and dendritic cells when cultured with M-CSF plus IL-4 (Fig. 5B and Supplemental Fig. 1C). These cells were of immunocompetent phenotype, as they produced increased amounts of proinflammatory cytokines (TNF-α and IL-6) but little immunosuppressive cytokines (IL-10 and TGF-β) when stimulated with bacterial LPS (Fig. 5C). In contrast, ectopic expression of NFI-A diminished the p21 expression and attenuated differentiation capacity. LPS-stimulated cells acquired immunosuppressive properties, with low TNF-α and IL-6 and elevated IL-10 and TGF-β levels (Fig. 5C). These results further support that NFI-A reprograms Gr1⁺ CD11b⁺ myeloid progenitors to become MDSCs.

Figure 5. NFI-A overexpression in Gr1⁺ CD11b⁺ cells of naive mice blocks differentiation and induces an immunosuppressive phenotype.

Gr1⁺ CD11b⁺ cells were isolated from the bone marrow of normal (naive) mice and transfected with an NFI-A expression plasmid or an empty vector. After 36 h, cells were differentiated for 6 d with M-CSF plus rIL-4 (10 ng/ml each). (A) Cells pooled from 3 mice/group were analyzed by Western blot for NFI-A and p21 proteins. Results are representative of 3 experiments. (B, left) Flow cytometry analysis of differentiated cells gated on F4/80⁺ CD11b⁺ or CD11c⁺ MHC II⁺ staining; (right) percentages of differentiated cells. Data are expressed as means ± sd from 6 mice/group (*P < 0.05) and represent 1 of 3 experiments. (C) Differentiated cells were stimulated for 12 h with 1 μg/ml LPS (Escherichia coli serotype 0111:B4), and cell supernatants were assessed by ELISA for proinflammatory (TNF-α and IL-6) and anti-inflammatory (IL-10 and TGF-β) cytokines. Data are expressed as means ± sd from 6 mice/group (*P < 0.05) and are representative of 3 experiments.

NFI-A deregulates proteins that support myeloid cell differentiation and maturation

To investigate further the role of NFI-A in Gr1⁺ CD11b⁺ cell expansion, we investigated proteins known to regulate myeloid cell differentiation, including cyclin D1, PTEN, and NF-κB p65. Cyclin D1 is a G1-phase regulatory protein that complexes with CDK4 to initiate G1/synthesis cell-cycle progression and attenuate differentiation and maturation [33, 34]. p21 inhibits the cyclin D1–CDK4 protein complex, arrests cell-cycle progression, and promotes differentiation [35, 36]. Accordingly, we investigated whether the cyclin D1–CDK4 protein complex forms with or without p21 expression and with or without NFI-A expression (Fig. 4A) by immunoprecipitating cell lysates with an antibody against cyclin D1, followed by immunoblotting with cyclin D1 or CDK4 antibody. Cyclin D1 and CDK4 formed a complex in cells lacking the p21 protein and expressing NFI-A, and this complex was disrupted in cells expressing p21 protein but not NFI-A (Fig. 6A).

Figure 6. The transfection of NFI-A in Gr1⁺ CD11b⁺ cells obtained from septic mice after miR-21 and miR-181b knockdown expands understanding of the immunosuppressive phenotype.

(A) Gr1⁺ CD11b⁺ cells were isolated from the bone marrow of late septic mice and transfected with NFI-A-specific or control siRNA, as described in Fig. 4. After 36 h, whole-cell lysates were prepared and immunoprecipitated (IP) with cyclin D1-specific or IgG isotype control antibody by use of protein G-agarose beads and then immunoblotted with a cyclin D1 or CDK4 antibody. Results are representative of 2 experiments. (B) Gr1⁺ CD11b⁺ cells were isolated from late septic mice that were injected with antagomiRs, as described in Fig. 1. The cells were then transfected with an NFI-A expression plasmid (to restore NFI-A levels) or an empty vector as in Fig. 5. Cells from sham mice served as a control. After 36 h, whole-cell lysates were prepared from a portion of these cells and probed for the indicated proteins by Western blotting. The loading control (β-actin) levels were consistent between blots, but only a sample is shown to facilitate comparison. Results are representative of 2 experiments. (C) The remainder of the cells was differentiated for 6 d with M-CSF plus rIL-4 (10 ng/ml each), and the percentages of macrophages (F4/80⁺ CD11b⁺) and dendritic cells (CD11c⁺ MHC II⁺) were determined by flow cytometry. Data are expressed as means ± sd from 6 mice/group (*P < 0.05, compare mutant antagomiRs with antagomiR groups; **P < 0.05, compare antagomiRs + vector with antagomiRs + NFI-A) and are representative of 2 experiments. (D) Gr1⁺ CD11b⁺ cells were isolated from late septic mice and transfected with NF-κB p65-specific or control siRNA. After 36 h, nuclear proteins were isolated from a portion of the cells and used to confirm the NF-κB activation and knockdown by Western blotting. Results are representative of 3 experiments. (E) The remainder of the cells was differentiated for 6 d with M-CSF plus rIL-4 (10 ng/ml each), and the percentages of differentiated macrophages (F4/80⁺ CD11b⁺) and dendritic cells (CD11c⁺ MHC II⁺) were determined by flow cytometry. Data are expressed as means ± sd from 6 mice/group (*P < 0.05, compare NF-κB siRNA with control siRNA) and are representative of 3 experiments. (F) Late sepsis Gr1⁺ CD11b⁺ cells were transfected with control or NF-κB siRNA, cocultured with CD4⁺ T cells, and treated as described in Fig. 3. T cell proliferation and IFN-γ production were determined by flow cytometry and ELISA, respectively. Data are expressed as means ± sd from 6 mice/group (*P < 0.05, compare with sham; **P < 0.05, compare with control siRNA) and are representative of 3 experiments.

PTEN down-regulation expands MDSCs in tumor-bearing mice by activating the Akt/NF-κB pathway and increasing myeloid cell expansion [37], while attenuating differentiation in human epithelial cell lines [38]; thus, persistent activation of NF-κB p65 increases Gr1⁺ CD11b⁺ progenitor numbers but not differentiation, a tumor-promoting path [39]. We did not detect the NF-κB p65 protein in Gr1⁺ CD11b⁺ cells from sham mice, which express high levels of PTEN protein (Fig. 6B). In contrast, NF-κB p65 increased in cells with reduced PTEN and decreased in cells with increased PTEN. Thus, changes in these target proteins associated with mouse tumors paralleled the sepsis-induced Gr1⁺ CD11b⁺ MDSC generation, which increases NFI-A, down-regulates p21 and PTEN, up-regulates NF-κB p65, and modifies macrophage and dendritic cell differentiation (Fig. 6B and C). Finally, we found that the knockdown of NF-κB p65 in Gr1⁺ CD11b⁺ cells isolated from late septic mice that were injected with mutant antagomiRs (i.e., express NFI-A; Fig. 6B) increased Gr1⁺ CD11b⁺ cell differentiation into macrophages and dendritic cells (Fig. 6D and E), further supporting that increases in NF-κB p65 in Gr1⁺ CD11b⁺ cells during sepsis block their differentiation and maturation. In addition, Gr1⁺ CD11b⁺ cells with NF-κB p65 knockdown did not exhibit immunosuppressive activity when cocultured with CD4⁺ T cells, as they did not significantly affect T cell proliferation or IFN-γ production (Fig. 6F).

DISCUSSION

This study clarifies further how MDSCs are generated during sepsis. Previously, we reported that increased miR-21 and miR-181b expression contributes to MDSC generation and a poor sepsis outcome [14, 24]. Here, we show that the CCAAT-related transcription factor NFI-A protein induction, supported by miR-21 and miR-181b, reversibly expands MDSCs by limiting differentiation and maturation of the Gr1⁺ CD11b⁺ macrophage and dendritic cell progenitors into competent innate immune cells. Under different conditions, others found that NFI-A represses myeloid cell differentiation and maturation [40, 41]. Importantly, the reversal of this phenotype allows monocytes and granulocytes to differentiate normally [40–42]. The importance of this axis is supported further by the discovery that RNA interference-mediated silencing of NFI-A in myeloid cell lines and human myeloid progenitors enhances their differentiation to competent monocytes [41]. In addition, NFI-A silencing in human promyelocytic leukemia cells enhances retinoic acid-induced and miR-223-supported granulocytic differentiation [40]. NFI-A binds and inhibits the miR-223 promoter, whereas miR-223 is part of a feedback loop that inhibits NFI-A and promotes Gr1⁺ CD11b⁺ cell differentiation and maturation. Other and our data suggest that this adaptation loop is dysregulated during sepsis. Adding further to the importance of NFI-A in sepsis-induced MDSC generation and the operation of a connected regulation loop is our finding that the transfection of NFI-A into naive Gr1⁺ CD11b⁺ cells blocks differentiation and increases immunosuppression (Fig. 5). Furthermore, miR-223 levels decrease as NFI-A levels increase (data not shown). Others found decreased miR-223 levels in mouse dysfunctional heart tissue during early sepsis with increased mortality [43]. Further studies of how the NFI-A loop functions are needed.

Our finding that NFI-A and p21 proteins are reciprocally expressed confirms reports that NFI-A represses the p21 promoter in proliferating human skin fibroblasts and attenuates their differentiation [29]. p21 represses CDKs, which control the G1 checkpoint and maintain hematopoietic stem cells in a quiescent state [30, 35]. p21 also acts through the cyclin CDK to arrest cell-cycle progression in human fibroblasts and breast cancer cell lines [34, 36]. The deletion of the p21 murine gene increases hematopoietic progenitor numbers by attenuating their differentiation [30], and deletion of the p27 gene, another CDK inhibitor, also enhances hematopoietic progenitor cell numbers in mice [44]. We also find that p27 protein levels, like p21, decrease in late sepsis during increased NFI-A expression. Together, substantial data support a reciprocal relationship of NFI-A and p21/p27 in regulating the cyclin-CDK G1 checkpoint and controlling MDSC generation during late sepsis.

In this study, the cyclin D1–CDK4 protein complex formation increased when PTEN protein levels decreased, and siRNA-mediated NFI-A depletion disrupted this reciprocal axis. The reintroduction of NFI-A expression in cells from the antagomiR-injected mice promoted the cyclin D1–CDK4 complex formation (data not shown) and decreased PTEN protein levels (Fig. 6B). The cyclin D1–CDK4 protein complex, controlled by p21, facilitates cell-cycle progression and attenuates differentiation of lung cancer cell lines [33]. PTEN, which is inhibited by CDK complexes, increases in breast cancer cells; this effect correlates with increased p21 expression [45]. PTEN decreases in tumor-bearing mice promote MDSC accumulation [37]. PTEN is an upstream regulator of the PI3K–Akt kinase pathway and has been shown to inhibit Akt phosphorylation downstream of PI3K [46, 47]. Up-regulation of PTEN in skin keratinocytes reduces Akt phosphorylation/activation and attenuates their differentiation [48]. An important issue relevant to our finding of disrupted PTEN activity in sepsis is that human epithelial cells activate NF-κB p65 via the PI3K–Akt pathway. In this cell system, PTEN expression was decreased and led to phosphorylation/activation of Akt and subsequent activation of NF-κB, whereas inhibition of Akt expression reduced NF-κB activation and promoted cell differentiation [38]. We find increases in Akt phosphorylation in late sepsis Gr1⁺ CD11b⁺ cells (data not shown). Moreover, persistent NF-κB p65 activation in miR-146a-deficient mice promotes chronic myeloproliferation by increasing Gr1⁺ CD11b⁺ progenitors in the bone marrow [39]. Here, we show that the NF-κB p65 is activated in late sepsis Gr1⁺ CD11b⁺ MDSCs at a time when PTEN levels are low. When NF-κB p65 is depleted by siRNA in Gr1⁺ CD11b⁺ cells from late septic mice, pronounced increases in differentiation occurred, and Gr1⁺ CD11b⁺ cells were no longer immunosuppressive. Together, these results suggest that NFI-A promotes Gr1⁺ CD11b⁺ MDSC expansion in sepsis by disassociating myeloid cell-cycle progression from differentiation. Although our results clearly link the up-regulation of NFI-A downstream of miR-21 and miR-181b to NF-κB activation and attenuated differentiation of Gr1⁺ CD11b⁺ cells, further studies are needed to clarify how the NFI-A/p21 axis enhances NF-κB activation during sepsis, specifically how p21 inhibition by NFI-A links to reduced PTEN expression.

The concept of MDSC generation and activation is complex, and several mechanisms generate MDSCs among different pathologies [17, 19, 21]. Cytokines, such as IL-1β, IL-6, IL-10, and TGF-β, and growth factors, such as vascular endothelial growth factor and stem cell factor, promote repressor Gr1⁺ CD11b⁺ MDSCs (reviewed in ref. [20]). MDSCs transferred from tumor-bearing mice into healthy hosts rapidly differentiate into macrophages and dendritic cells [49]. We reported that transferring naive CD34⁺ progenitor cells into mice with late sepsis rescues mice from innate immune incompetence and substantially improves survival [13]. The STAT transcription factor family regulates both innate and adaptive immunity [50]. STAT3 activation is essential for MDSC expansion in tumor-bearing mouse models [49, 51] and also regulates the promoters of miR-21 and miR-181 in human epithelial cell lines to attenuate their differentiation [38]. We find increased phosphorylation/activation of STAT3 in sepsis Gr1⁺ CD11b⁺ MDSCs (data not shown).

We propose a model (Fig. 7) for the Gr1⁺ CD11b⁺ MDSC expansion in late sepsis, which induces acute inflammatory reactions that activate miR-21 and miR-181b expression, together, supporting NFI-A-dependent p21 repression. In the absence of p21, the cyclin D1–CDK4 protein complex remains active to arrest cells at the G1 phase. Concomitant and sustained PTEN inhibition leads to persistent activation of NF-κB p65, which promotes myeloid progenitor expansion at the expense of differentiation. The biologic outcome is accumulation of MDSCs and sustained sepsis immunosuppression.

Figure 7. Proposed model for sepsis-induced Gr1⁺ CD11b⁺ MDSC expansion.

Sepsis-induced miR-21 and miR-181b support increased expression of transcription factor NFI-A, which disrupts p21 function. This is accompanied by persistent NF-κB p65 activation via the cyclin D1–CDK4 complex and PTEN, which enhance myeloid proliferation and attenuate differentiation. Accumulating MDSCs support late sepsis immunosuppression.

Mounting evidence supports that late sepsis immunosuppression increases mortality, both from the primary infection and by opportunistic secondary infections [1, 2]. This concept is robustly supported at the mechanistic level by the report of marked immunosuppression in spleens and lungs obtained from humans dying of sepsis [5] and by the seminal finding that latent viral infection activation occurs in >75% of humans with sepsis [52]. We suggest that therapeutically targeting mechanisms responsible for MDSC generation may counter late sepsis immunosuppression and improve its prognosis.

AUTHORSHIP

C.M. designed and did most of the experiments. E.A. did some experiments. D.Y. contributed to the discussion of the results. Z.Q.Y. contributed to data interpretation. C.E.M. edited the manuscript. M.E.G. interpreted the data and wrote the manuscript.

Glossary

antagomiR

chemically modified, cholesterol-conjugated miRNA inhibitor

APC

allophycocyanin

CDK

cyclin-dependent kinase

CLP

cecal ligation and puncture

G1 phase

Gap 1 phase

MDSC

myeloid-derived suppressor cell

MHC II

MHC class II

miRNA/miR

microRNA

NFI-A

nuclear factor 1 A type

PTEN

phosphatase and tensin homolog

siRNA

small interfering RNA

UTR

untranslated region

DISCLOSURES

The authors declare no competing financial interests.