Perturbed MafB/GATA1 axis after burn trauma bares the potential mechanism for immune suppression and anemia of critical illness

Short abstract

Anemia of critical illness and immune suppression following burn trauma is orchestrated by common myelopoietic transcriptional mechanisms, specifying the terminal fate of CMPs.

Abstract

Patients who survive initial burn injury are susceptible to nosocomial infections. Anemia of critical illness is a compounding factor in burn patients that necessitates repeated transfusions, which further increase their susceptibility to infections and sepsis. Robust host response is dependent on an adequate number and function of monocytes/macrophages and dendritic cells. In addition to impaired RBC production, burn patients are prone to depletion of dendritic cells and an increase in deactivated monocytes. In steady‐state hematopoiesis, RBCs, macrophages, and dendritic cells are all generated from a common myeloid progenitor within the bone marrow. We hypothesized in a mouse model of burn injury that an increase in myeloid‐specific transcription factor V‐maf musculoaponeurotic fibrosarcoma oncogene homolog B at the common myeloid progenitor stage steers their lineage potential away from the megakaryocyte erythrocyte progenitor production and drives the terminal fate of common myeloid progenitors to form macrophages vs. dendritic cells, with the consequences being anemia, monocytosis, and dendritic cell deficits. Results indicate that, even though burn injury stimulated bone marrow hematopoiesis by increasing multipotential stem cell production (Lin^negSca1^poscKit^pos), the bone marrow commitment is shifted away from the megakaryocyte erythrocyte progenitor and toward granulocyte monocyte progenitors with corresponding alterations in peripheral blood components, such as hemoglobin, hematocrit, RBCs, monocytes, and granulocytes. Furthermore, burn‐induced V‐maf musculoaponeurotic fibrosarcoma oncogene homolog B in common myeloid progenitors acts as a transcriptional activator of M‐CSFR and a repressor of transferrin receptors, promoting macrophages and inhibiting erythroid differentiations while dictating a plasmacytoid dendritic cell phenotype. Results from small interfering RNA and gain‐of‐function (gfp‐globin transcription factor 1 retrovirus) studies indicate that targeted interventions to restore V‐maf musculoaponeurotic fibrosarcoma oncogene homolog B/globin transcription factor 1 balance can mitigate both immune imbalance and anemia of critical illness.

Abbreviations

APC

= allophycocyanin

BM

= bone marrow

CBC

= complete blood count

cDC

= classic dendritic cell

CDP

= common dendritic cell progenitor

CMP

= common myeloid progenitor

CR3A

= complement receptor 3A

DC

= dendritic cell

EB

= erythroblast

Flt3L

= Flt3 ligand

GATA1

= globin transcription factor 1

GMP

= granulocyte monocyte progenitor

Hct

= hematocrit

Hgb

= hemoglobin

ic

= intracellular

ICU

= Intensive Care Unit

IRES

= internal ribosome entry site

LK

= Lin^negSca1^negcKit^hi

LSK

= Lin^negSca1^hicKit^hi

MØ

= macrophage(s)

MafB

= V‐maf musculoaponeurotic fibrosarcoma oncogene homolog B

MEP

= megakaryocyte erythrocyte progenitor

MFI

= mean fluorescence intensity

MHC‐II

= MHC class II

MSCV

= murine stem cell virus

PBD

= postburn day

pDC

= plasmacytoid dendritc cell

SCF

= stem cell factor

siRNA

= small interfering RNA

TBSA

= total body surface area

TfR

= transferrin receptor

Introduction

Critically injured burn patients endure dysregulated immunity predisposing to nosocomial infections and related septic complications [1, 2–3]. Anemia is a persistent consequence in critically ill ICU patients [4, 5] and burn patients requiring multiple transfusions [6, 7–8]. Transfusions further increase overall immune suppression, infection, morbidity, and mortality [9]. Central to the pathobiology of burn trauma is the altered innate and adaptive immune responses mediated by leukocytes [10]. Specifically, antigen presentation of MØs and DCs from burn, trauma, and sepsis patients is less effective, exhibiting an immunosuppressive phenotype [11, 12, 13–14] and reduced DC production [15, 16, 17, 18–19]. Review of the current literature indicates an imbalance in erythrocyte and leukocyte production and function following burn and trauma [20, 21, 22, 23, 24–25]. Destruction of RBCs in critical burn injury, combined with the rapid turnover and short half‐life of leukocytes, increase the demand for BM hematopoiesis.

MØ and DCs have a common BM progenitor [26, 27], which is shared with erythrocytes and megakaryocytes [28]. Production of RBCs starts from the hematopoietic multipotential stem cells (LSK) by sequential lineage commitments through CMPs that give rise to MEPs. CMPs can also progress toward granulocyte or monocyte/MØ development through GMPs or toward DC development through CDPs [29]. We have previously demonstrated that burn injury leads to a reduction in DCs, as well as an increase in GMPs and monocytes [18, 19]; albeit, monocytes exhibit functional deactivation [16, 30]. Therefore, we hypothesized that the persistence of postburn anemia is a result of altered lineage commitment of CMPs away from erythroid lineage, resulting in reduced MEP production and ultimately RBC deficits. Additionally, the terminal fate of CMPs will be shifted away from DC generation, validating the hematopoietic mechanism of the immune dysfunction following burn injury. Furthermore, based on our previous reports [18, 19], we tested the premise that high MafB expression in CMPs after burn injury skews their lineage bias of CMPs toward myeloid (GMPs) and away from erythroid (MEPs) commitment.

MATERIALS AND METHODS

Animals and burn injury

Approval from the Loyola University Medical Center's Institutional Animal Care and Use Committee (Maywood, IL, USA) was obtained for all experimental protocols. Six‐week‐old B₆D₂F₁ male mice (The Jackson Laboratory, Bar Harbor, ME, USA) were housed in our Comparative Medicine Facility with a 12 h light/dark cycle and controlled temperature (20–22°C). Mice were allowed to acclimatize to their environment for 1 wk before experiment start. Food and water were provided ad libitum. All mice were anesthetized using intraperitoneal ketamine (100 mg/kg) and xylazine (2.5 mg/kg) injection and inhaled 5% isoflurane. The dorsal fur was shaved, and animals were randomized into burn and sham groups. Burn animals received a 15% scald burn to their dorsum via immersion in a 100°C water bath for 8 s [18, 31]. All animals were resuscitated with normal saline (2 ml) via intraperitoneal injections. A sham group of mice was administered anesthesia, shaved, and resuscitated. Six animals/group were used for each time point. During the 3 wk postburn periods, no mortality was associated with any experimental group. Burn wounds were not treated with any topical agents. There was no evidence of wound infection or overt morbidity in any of the animals.

Sample collection

Mice were euthanized via CO₂ inhalation at given PBDs (PBD #3, 7, 14, and 21). Blood was collected by cardiac puncture using a heparin‐coated, 1 ml syringe. CBCs were determined using HemaTrue Veterinary Hematology Analyzer (Heska, Loveland, CO, USA). Bilateral femurs were eluted in McCoy's medium (Thermo Fisher Scientific Life Sciences, Waltham, MA, USA). To obtain Lin^neg cells, total BM cells were incubated with a combination of biotin‐conjugated primary antibodies: anti‐IL‐7R, anti‐CD86, anti‐CD11c, anti‐Ter119, anti‐CD19, anti‐B220, anti‐CD11b, anti‐CD90, anti‐CD8a, anti‐Gr1, and anti‐CD3e (BD Biosciences, San Jose, CA, USA); followed by the antibiotin‐conjugated microbeads (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany). The Lin^neg cells were separated by negative selection via autoMACS (Miltenyi Biotec), per the manufacturer's instructions, and labeled with anti‐CD34‐Pacific Blue, anti‐FcγR‐PE‐Cy7 (eBioscience, San Diego, CA, USA), anti‐cKit‐APC, and anti‐Sca1‐PerCP‐Cy‐5.5 (BD Biosciences). The LSK and LK fractions, comprising of multipotential progenitors, were gated using FACSAria (BD Biosciences). The LKs were regated for CMPs (CD34^hiFcγR^low), GMPs (CD34^hiFcγR^hi), and MEPs (CD34^lowFcγR^low) accordingly (Supplemental Fig. 1). Aliquots of aseptically sorted CMPs were used for culture toward myeloid (MØ or DC) and erythroid (EB) differentiation and transcription factor analysis.

Ex vivo culture

Sorted CMPs (5 × 10⁴) were placed in Costar low‐attachment culture plates (Sigma‐Aldrich, St. Louis, MO, USA) or Teflon‐coated culture dishes (Savillex, Eden Prairie, MN, USA) for 5 d in intermediate myeloerythroid cocktail at 37°C with 5% CO₂. On d 6, the culture medium was replaced with specific cocktails for MØs, DCs, and EBs, as explained below, and cells were harvested for phenotype analysis on d 10 for EBs and MØs and d 12 for DCs.

Intermediate myeloerythroid cocktail.

IMDM (Thermo Fisher Scientific Life Sciences) was supplemented with 20% FBS (Gemini Bio‐Products, West Sacramento, CA, USA) and enriched with growth factors SCF (10 ng/ml; Stemcell Technologies, Vancouver, BC, Canada), IL‐3 and IL‐6 (20 ng/ml; Thermo Fisher Scientific Life Sciences), Flt3L (40 ng/ml), and GM‐CSF (50 ng/ml; R&D Systems, Minneapolis, MN, USA).

EB cocktail.

Serum‐free expansion medium (Stemcell Technologies) was enriched with IL‐3 (20 ng/ml), SCF (10 ng/ml), Flt3L (20 ng/ml), GM‐CSF (50 ng/ml), human erythropoietin (2.5 U/ml), and human plasma (5%).

MØ cocktail.

IMDM was supplemented with M‐CSF (20 ng/ml; R&D Systems) and FBS (20%).

DC cocktail.

IMDM was supplemented with GM‐CSF (50 ng/ml; R&D Systems), IL‐4 (20 ng/ml; Thermo Fisher Scientific Life Sciences), and FBS (20%) and LPS (200 ng/ml; BD Biosciences) on d10 for DC maturation.

Penicillin/streptomycin (2%) and Fungizone antimycotic (0.1%; Thermo Fisher Scientific Life Sciences) were added to the culture medium.

Flow cytometry

Cultured cells (5 × 10⁵) were washed with PBS (Thermo Fisher Scientific Life Sciences); treated with a combination of antibodies for 20 min at 4°C for EBs, DCs, and MØs, as stated below;, washed with PBS; and analyzed within 2 h. MØ staining included the following: anti‐CD11b‐PE‐Cy7 (BD Biosciences), anti‐F4/80‐Pacific Blue, and anti‐ER‐MP20‐FITC; DC staining included the following: CD11c‐APC, B220‐APC‐Cy7, and MHC‐II‐PE (eBioscience); and EB staining included the following: CD71‐PE, Ter119‐PerCP‐Cy5.5, and anti‐ CD11b‐PE‐Cy7. Alexa Fluor 430‐Live/Dead was included along with other antibodies to select only live cells and further gated for MØs (F4/80⁺), DCs (F480^neg CD11c⁺ B220^+/− MHC‐II^+/−), and EBs (CD11b^negCD71⁺Ter119⁺) accordingly.

For ic transcription factors, CMPs (5 × 10⁴) were fixed and permeabilized using Cytofix/Cytoperm Plus (BD Biosciences) for 20 min. Nonspecific binding of primary antibodies was prevented by using 1% BSA before treating with rabbit primary mAb anti‐PU.1 (Cell Signaling Technology, Beverly, MA, USA) or anti‐MafB (Abcam, Cambridge, United Kingdom) for 24 h or anti‐GATA1 (Cell Signaling Technology) for 48 h. After rinsing with Perm/Wash Buffer (BD Biosciences), cells were treated for 2 h with anti‐rabbit IgG‐FITC or IgG‐PE secondary antibodies (Abcam). Cells were rinsed and resuspended in 1% paraformaldehyde.

Gene silencing (siRNA)

CMPs (5 × 10⁴) were rinsed with Opti‐MEM (Thermo Fisher Scientific Life Sciences) to remove any residual FBS and then transfected with 40 pmole/10⁶ cells of MafB siRNA (sc‐35840; Santa Cruz Biotechnology, Santa Cruz, CA, USA) by the addition of X‐tremeGENE siRNA Transfection Reagent (Roche Diagnostics, Mannheim, Germany) following kit instructions. After incubating for 4 h, a myeloerythroid cocktail was added for differentiation. Transfection with a control siRNA‐A (sc‐37007) did not promote gene silencing, determined by protein transcription after 72 h.

Morphology

Microscopy.

For morphologic identification of EBs, MØs, and DCs, cultured cells (30 K/300 µl) were cytospun onto glass slides, methanol fixed, stained with May‐Grünwald‐Giemsa, and confirmed by light microscopy at 20× magnification using EVOS FL Cell Imaging System (Thermo Fisher Scientific Life Sciences).

AMNIS.

Following EB staining with fluorochrome‐conjugated anti‐CD71 and anti‐Ter119 antibodies, acquisition was performed using ImageStreamX Imaging Flow Cytometer (EMD Millipore, Billerica, MA, USA) equipped with INSPIRE software. At least 10,000 cells were analyzed per sample at 60× magnification. Data analysis was performed using the image data exploration and analysis software (Amnis). Intensity‐adjusted bright field images were collected.

Confocal microscopy

Following fluorescent‐conjugated antibody staining, aliquots of cells were cytospun onto microscopic slides and preserved using Vectashield H‐1500 mounting medium with DAPI (Vector Laboratories, Burlingame, CA, USA). A Zeiss LSM 510 laser‐scanning microscope (Carl Zeiss MicroImaging GmbH, Jena, Germany) was used to view with C‐Apochromat 40× 1.20 water immersion, and images were acquired using Zeiss LSM 510, version 4.2, SP1 software.

Western blot

BM Lin^negcKit^pos cell lysates were prepared using 1× Laemmli buffer (40 µl/10⁶), followed by a 660 nm protein assay (Bio‐Rad Laboratories, Hercules, CA, USA). Equal amounts (25 µg) of lysates were loaded on Mini‐PROTEAN TGX gels (4–20%), transferred to a polyvinylidene difluoride membrane, and visualized by a standard Western blot protocol. The following rabbit mAb were used to detect the expression of specific antigens: the polyclonal rabbit antibody MafB was purchased from Novus Biologicals (Littleton, CO, USA), and GAPDH (D16H11) and GATA1 (D52H6) were purchased from Cell Signaling Technology.

Gain‐of‐gene function

MSCV‐GATA1‐IRES‐GFP (Gata1 vector)‐incorporated MigR1 plasmid material was gifted by Dr. John D. Crispino (Northwestern University, Evanston, IL, USA). MSCV‐IRES‐GFP (empty vector)‐labeled MigR1 plasmid was obtained as a gift from Dr. Jiwang Zhang (Loyola University). Plasmid DNA was transformed into DH5α10B chemically competent Escherichia coli cells (Thermo Fisher Scientific Life Sciences) by heat shock, according to the manufacturer's instructions. Transformed bacterial cells were plated onto ampicillin‐selective Luria‐Bertani agar plates, which after single colony selection, were expanded in ampicillin‐containing media. Purified plasmid DNA was extracted from the expanded, transformed cells by Maxiprep column purification (Qiagen, Valencia, CA, USA), according to the manufacturer's instructions, in which concentration and purity were determined and confirmed by NanoDrop (Thermo Fisher Scientific Life Sciences) analysis and agarose gel electrophoresis (data not shown). Gata1 and empty vector plasmids were transfected into 3 × 10⁶ Gryphon‐Ampho cells (Allele Biotechnology, San Diego, CA, USA) using FuGENE (Promega, Madison, WI, USA), according to the manufacturer's instructions. Transfected Gryphon‐Ampho cells were incubated at 37°C and 5% CO₂, and retrovirus was collected using Centricon Plus‐70 filtration containers, according to the manufacturer's instructions at 48 and 72 h after transfection, and stored in aliquots for transduction in BM progenitors. Isolated CMPs (1 × 10⁵) were transfected with retrovirus (Gata1 or empty vector) with polybrene (4 μg/μl) by spinfection at 2500 rpm for 4 h and resuspended in EB growth cocktail, as described in Materials and Methods. Transfection efficiency was determined by FACS (Fortessa) analysis (see Fig. 5C and D), and specificity was confirmed by confocal microscopy (see Fig. 6A).

Figure 5.

GATA1 gain of function restores CMP → EB that was decreased by burn injury. GATA1 is a erythroid‐specific, counter‐regulatory transcription factor expressed at low levels in CMPs [28]. Restoration of GATA1 function can improve the EB differentiation that was decreased by burn injury. (A) Approximately 60% of CMPs in either group were GATA1⁺, but the MFI of ic‐GATA1 protein expression in CMPs from burn mice was reduced significantly compared with sham (*P < 0.05, red vs. blue histograms). (B) CMP → EB development was determined by CD71⁺Ter119⁺ expression (gated region). EB development was significantly attenuated in burn vs. sham CMPs; empty vector (EV) transfection (left) in CMPs did not have any effect, showing only 49% of burn and 64% of sham cultures. On the other hand, transfection with GATA1 vector (right) increased the proportion of EBs to 86% in burn and 79% in sham CMPs, respectively. Overall, GATA1 transfection improved the erythroid development of CMPs. (C and D) To determine the contribution of GATA1⁺ cells in EB expansion of burn CMPs, all gfp⁺ and gfp⁻ fractions from the empty vector (D) or the GATA1 vector‐transfected (C) CMP → EB cultures were regated and compared with the proportion of EBs in gfp^pos vs. gfp^neg fractions. The majority of CD71⁺ Ter119⁺ EBs resulted from the gfp^pos fraction compared with the gfp^neg fraction (gfp^pos: 60%; gfp^neg: 22%), accounting for the overall increase in EBs to the gfp‐GATA1 vector in burn CMPs, whereas in empty vector‐transfected cells, CD71⁺ Ter119⁺ EBs remained low with gfp^pos (37%) or gfp^neg (26%) fractions from burn CMPs.

Figure 6.

CMP → EB; stages of EB development were recognized using ImageStreamX Imaging Flow Cytometer (Amnis). gfp‐GATA1‐transfected CMPs were placed in a erythroid development cocktail, supplemented with erythropoietin. (A) Representative confocal images of CMPs taken at 72 h after transfection with gfp‐empty vector (left) or gfp‐GATA1 vector (right) in green and anti‐GATA1 antibody staining (pink) counterstained with DAPI for nucleus (blue). The overlay in each panel indicates that GATA1 protein is only present with the gfp‐GATA1 vector and absent with the gfp‐empty vector. This confirms the specificity of the gfp‐GATA1 vector in CMPs, which is in line with erythroid expansion of burn CMPs. Zeiss LSM 510 laser‐scanning microscope (Carl Zeiss MicroImaging GmbH) was used to view with C‐Apochromat ×40 1.20 water immersion, and images were acquired using Zeiss LSM 510, version 4.2, SP1 software. (B) Histograms exhibit d 10 cultures of CD11B^neg cells to avoid any myeloid cell contamination before selecting the gfp^pos cells. Few representative images from the CD11B^neg gfp‐GATA1⁺ (green) fraction showing differential expression of CD71⁺ (yellow) and Ter119⁺ (red) are presented. EBs (gfp⁺ green) expressing low levels of the CD71 and Ter119 basophilic EB stage are also relatively bigger in size, as seen in brightfield. Whereas EBs start off as larger cells, they get smaller in size as they mature into CD71^+/−Ter119⁺⁺ erythrocytes.

Statistical analyses

Results from all experiments are expressed as means ± sem, containing 4–6 animals/group and repeated 2–3 times. For comparison between groups, multivariate analysis was conducted using ANOVA statistics, followed by Tukey's post hoc test (P < 0.05) using KaleidaGraph Statistical Software, version 4.1.0 (Synergy Software, Reading, PA, USA).

RESULTS

Peripheral blood and BM responses to burn injury

CBCs from sham and burn mice are enumerated in Table 1 . Monocyte and granulocyte counts were elevated significantly from PBD #3. RBC, Hgb, and Hct values were decreased from PBD #7 through PBD #21. Whereas granulocyte counts normalized after the initial increase on PBD #3 and PBD #7 after burn, the monocyte counts remained high, and RBC counts remained low until PBD #21. Peripheral blood profile reflected symptoms of mild to moderate anemia [32].

Table 1.

CBC

Exp. groups	Hgb (g/dl)	Hct (%)	RBCs (106 cells/µl)	Monocytes (%)	Granulocytes (%)
Sham (pooled)	14.0 ± 0.2	43.3 ± 0.5	9.56 ± 0.1	5.7 ± 0.2	19.47 ± 0.7
PBD #3	14.97 ± 0.4	46.8 ± 1.4	10.39 ± 0.4	10.87 ± 1.7 a	53.9 ± 2.4 a
PBD #7	12.7 ± 0.2 a	38.9 ± 1.2 a	8.67 ± 0.3 a	10.03 ± 0.2 a	38.6 ± 0.6 a
PBD #14	13.0 ± 0.1 a	40.6 ± 0.5 a	9.11 ± 0.1 a	9.75 ± 0.8 a	28.38 ± 4.1
PBD #21	12.6 ± 0.4 a	39.3 ± 1.0 a	8.62 ± 0.3 a	9.95 ± 0.6 a	22.33 ± 5.4

Heparinized blood was obtained via cardiac puncture from sham and burn mice harvested on PBD #3, 7, 14, and 21. CBC was determined using HemaTrue hematology analyzer. Results from sham were similar at all different time points; hence, data are pooled. RBC, Hgb, and Hct values were decreased significantly from PBD #3 through PBD #21 compared with sham. Whereas monocyte counts were elevated significantly from PBD #3 through PBD #21, granulocyte counts normalized after the initial increase on PBD #3 and PBD #7 after burn. The divergent and reciprocal responses between high monocyte counts and low RBC counts persisted until PBD #21. n = 6–8 mice/group. Exp. groups, Experimental groups.

^a P < 0.0001 vs. sham.

To study the red cell mass and hematopoietic response, bilateral femurs were eluted after burn. The bright red coloration of total BM pellets fades progressively from PBD #3 to PBD #21 compared with sham indicating red cell depletion following burn injury ( Fig. 1A ). Next, we isolated the myeloid progenitors CMPs, GMPs, and MEPs and the LSK fractions by flow cytometry, as detailed in Supplemental Fig. 1A and B. LSKs were significantly increased following burn, whereas the percentage of CMPs in total BM fraction remained steady. However, a reduction in MEPs with a corresponding increase in GMPs was seen starting on PBD #3 and persisted until PBD #21 (Fig. 1B and C). The disparity in BM GMPs and MEPs is consistent with a concurrent increase in circulating monocytes and granulocytes and a decrease in RBCs (Table 1). Significantly reduced Hct and Hgb levels, along with reduced RBC counts, indicate persistence of anemia during the study period.

Figure 1.

Total BM contents and progenitor profile. Bilateral femurs were harvested and eluted following third‐degree scald burn covering 15% TBSA. (A) Pelleted total BM samples from postburn and sham mice. The shift from a vibrant red in sham and PBD #3 to a slight hue of pink in PBD #21 indicates grossly a decline in both the concentration of RBCs and Hgb. (B) Bar graphs of LSKs expressed as percentage of Lin^neg BM cells from both femurs. Increase in LSKs indicates a continuance of robust hematopoietic response to burn injury. (C) Line graphs represent the percentage of myeloid progenitors on the y–axis and PBDs on the x‐axis. Blue line represents GMPs, which begin to increase from PBD #3. A reciprocal decrease in MEPs (red line) and a steady supply of CMPs (green line) are noted from PBD #3 through PBD #21. Results from sham harvested at all time points were similar; hence, data were pooled for sham. n = 6–8 mice/group. *P < 0.0001 vs. sham; ^P < 0.04 vs. sham; #P < 0.0001 vs. sham, PBD #3, 7, and 14.

Together, peripheral blood and the BM progenitor profile establish anemia and monocytosis with myelopoietic respecification beginning between PBD #3 and PBD #7. BM response is robust with increased production of LSKs. However, the performance of the transcriptional analysis of multipotential progenitors (Lin^negcKit^pos) revealed significant reduction in erythroid‐specific GATA1 expression with a slight increase in myeloid‐specific MafB, as determined by Western blots and flow cytometry (Supplemental Fig. 1C and D). The steady input of CMPs and unaltered PU.1 expressions (Supplemental Fig. 2A and B) after burn injury indicate persistent myeloid commitment. However, the disproportionate GMP vs. MEP production solicits the inquiry of lineage potency and terminal fate differentiation of CMPs.

Burn injury predisposes the terminal fate of CMPs

Next, we inquired the tri‐potential terminal commitment of CMPs following burn injury. Sorted CMPs (5 × 10⁴) from BM of sham and burn mice (PBD #7) were placed in respective growth factor cocktails for terminal differentiation into MØs, DCs, and EBs, henceforward referred as CMP → MØ, CMP → DC, and CMP → EB, respectively. We stratified the tri‐potency of CMPs in liquid culture conditions, as identified by morphology ( Fig. 2A ) and cell surface markers (MØ: F480⁺, DC: CD11c⁺, and EB: CD71⁺Ter119⁺) to confirm terminal differentiations (Fig. 2B). Burn injury reduced the terminal fate of CMP → EB (CD71⁺ Ter119⁺) by 1.8‐fold (burn: 18.7 ± 1.2%, sham: 34.5 ± 1.7%; P < 0.05) and CMP → DC differentiation by 0.7‐fold (burn: 43 ± 5%, sham: 60 ± 3%; P < 0.05); however, CMP → MØ differentiation was increased by 1.7‐fold (burn: 65 ± 7%, sham: 38 ± 4.8%; P < 0.05; Fig. 2C). Results in bar graphs are means ± sem from 4 to 6 animals in each group.

Figure 2.

Burn injury predisposes the terminal fate of CMPs. CMPs have tripotency when placed in physiologic conditions; all 3 lineages are produced from CMPs. Characteristics of sorted CMPs before differentiation and after cultured to CMP → MØ, CMP → DC, and CMP → EB via morphology by May‐Grünwald‐Giemsa staining technique are shown and read by light microscopy at 20× original magnification using the EVOS FL Cell Imaging System (A). Flow cytometric inspection, using cell‐surface markers F480^hiERMP20^low (MØs), CD11c ^hiF480^low (DCs), and CD71⁺ Ter119⁺ (EBs) compared with sham (left) and burn (right), is shown (B), and the means ± sem of CMP derivatives are depicted as bar graphs (C). Despite standard ex vivo conditions, burn injury skewed the terminal fate of CMPs away from EB generation (burn: 18.7 ± 1.2%, sham: 34.5 ± 1.7%). DC production was also impaired following burn injury (burn: 43 ± 5%, sham: 60 ± 3%), whereas MØ production was augmented by burn (burn: 65 ± 7%, sham: 38 ± 4.8%). *P < 0.05 vs. Sham.

Burn injury dictates the phenotype of CMP → DCs

CMPs can generate cDCs and pDCs. However, it is not known whether burn injury influences the phenotype of CMP→ DCs. Representative FACS results for membrane‐bound MHC‐II, CD11c, B220, and F480, as well as ic GATA1 and MafB from sham and burn CMP→ DCs, are shown in Fig. 3A and B . CD11c, ic‐GATA1, and MHC‐II expressions were reduced significantly in burn vs. sham CMP→ DC cultures. After eliminating MØ contamination (F480⁺ cells) and cell lysis (dead cells) as a reason for reduced DC numbers in burn, CD11c⁺ cells were gated on B220⁺ expression to differentiate between DC subsets and characterize pDCs. We noticed a 2.5‐fold increase in pDCs in burn, whereas the overall CD11c⁺ DC production was reduced in burn compared with sham. We next examined MHC‐II expression in the context of ic‐MafB and ic‐GATA1. Whereas there was no change in ic‐MafB expression between sham and burn, we noticed a significant decrease in ic‐GATA1, as well as MHC‐II in burn CMP → DCs. Confocal images of CMP → DCs, triple stained with anti‐MHC‐II, ic‐GATA1, and DAPI, are shown in Fig. 3C and D. MHC‐II and GATA1 expressions were reduced significantly in burn compared with sham. Overall, CD11c⁺ DCs remain significantly reduced, but burn injury preferentially influenced more pDCs expressing a low MHC‐II‐expressing phenotype.

Figure 3.

Burn injury dictates the phenotype of CMP‐derived DCs. Burn injury not only hampers DC production but also dictates the phenotype to pDCs. (A) Histograms representative of FACS results for ic‐GATA1 and ‐MafB and membrane‐associated MHC‐II, CD11c, and F480 in CMP → DCs from sham (blue) and burn (red) mice. CD11c, ic‐GATA1, and MHC‐II expressions was reduced significantly in burn vs. sham CMP → DC cultures. F480 (negative) and MafB (no change) were measured to exclude any possible MØ contamination in DC cultures. (B, left) Live/Dead gating shows that ∼80% of cultured cells were viable in both sham and burn. (B, middle) B220⁺CD11c⁺ double‐positive cells were gated to characterize pDCs. A 2.5‐fold increase in pDCs was found in burn (16%) compared with sham (44%). (B, right) MHC‐II is decreased significantly in burn and is independent of MafB expression in CMP → DCs. (C and D) Confocal images of CMP → DCs, triple staining with fluorescent antibodies for anti‐MHC‐II (red) GATA1 (green), and DAPI for nucleus (blue) are shown. Low expression of MHC‐II and GATA1 in burn (D) vs. sham (C) is seen in the intensities of green and red colors individually and in merged images, respectively. Zeiss LSM 510 laser‐scanning microscope (Carl Zeiss MicroImaging GmbH) was used to view with C‐Apochromat 40×_1.20 water immersion, and images were acquired using Zeiss LSM 510, version 4.2, SP1 software.

Early precursors of all DC subtypes (pDC and cDC) are defined by Flt3⁺ cells [33], whereas cell‐surface M‐CSFRs are associated in pDC progenitors [26, 27] and can determine the phenotype of DC progeny. Therefore, the examination of the prominence of M‐CSFRs in burn CMPs will explain pDC bias. Additionally, our preliminary data indicated a significant increase in MafB with a down‐regulation of GATA1 expression in burn CMPs (Supplemental Fig. 2A, C, and D).

M‐CSFR expression is augmented in CMPs through MafB modulation in burn injury

CMPs from sham and burn mice were analyzed for membrane‐bound M‐CSFR and ic‐MafB expressions. As expected, CMPs from burn mice had more M‐CSFR⁺ cells compared with sham mice. As MafB⁺M‐CSFR⁺ cells were increased in CMPs following burn injury, we reasoned that MafB up‐regulation following burn is influencing myeloid‐specific M‐CSFRs, driving the M‐CSF‐responsive differentiation potential of CMPs to MØ. We tested this concept using MafB‐siRNA. Sorted CMPs were placed in a proliferation cocktail (SCF + IL‐3) and treated with MafB‐siRNA or control‐siRNA. MafB silencing reduced M‐CSFR expression in sham CMPs from 38.8 to 12.5% ( Fig. 4A , left) and burn CMPs from 55.6 to 28% (Fig. 4A, right), emphasizing a direct role for MafB on M‐CSFR expression in CMPs. Approximately 50% inhibition was achieved after 48–72 h of MafB‐siRNA transfection. Figure 4B shows representative confocal images at 60 h after transfection of sham and burn CMPs. The histogram in Fig. 4B, lower left, represents the MFI of MafB obtained by flow cytometry (sham = 1387, burn = 2105, burn + siRNA = 1050, antibody control = 493). The results shown are 1 of 4 different experiments.

Figure 4.

MafB‐dependent erythromyeloid bifurcation of CMPs following burn injury. Freshly sorted CMPs from sham and burn mice were analyzed for membrane‐bound M‐CSFR and ic‐MafB expressions. (A). CMPs from burn mice (upper right) had more MafB⁺M‐CSFR⁺ cells compared with sham mice (upper left). MafB silencing also reduced M‐CSFR expression in sham CMPs from 38.8 to 12.5% (lower left) and burn CMPs from 55.6 to 28% (lower right), respectively. Results emphasize the direct role of MafB on M‐CSFR expression in CMPs. (B) Representative confocal images of MafB (green) in sham and burn CMPs (upper panels) and burn CMPs + MafB siRNA at 60 h after transfection (lower right panel). Histogram in lower left panel represents the MFI of MafB obtained by flow cytometry. B, Burn; B+si, MafB siRNA; S, sham. (C) CMPs from sham and burn mice were placed in an intermediate cocktail for 5 d. Myeloid and erythroid commitment bias were checked using anti‐CD11B (CR3A) and anti‐CD71 (TfR), as shown in the histograms, with CD11B and CD71 on the x‐axes, respectively. CD11B expression was the highest in burn CMPs (blue line) compared with sham (red line) and is reversed to sham levels after MafB silencing in burn CMPs (green line). Conversely, MafB silencing increased TfR (green line), which was otherwise down‐regulated following burn (blue line) compared with sham (red). (D) The cultures were extended in an Epo‐containing erythroid cocktail for 4 more days. Representative FACS plots from each condition, sham and burn, with and without siRNA MafB assessing erythroid commitment with CD71 on the y‐axis and Ter119 on the x‐axis. (E) Mean results are depicted in box plots enumerating the percentage of CD71⁺ cells (upper) and MFI of Tfrs (lower) in CMPs from sham and burn mice with scramble siRNA and MafB siRNA. *P < 0.0001 vs. sham; ^P < 0.01 vs. PBD #7. ctRNA, Control RNA.

Commitment bifurcation of burn CMPs is reset by MafB‐siRNA

Next, to evaluate the role of high MafB expression in burn CMPs toward terminal myeloid vs. erythroid fate, we placed CMPs from sham and burn mice, with and without MafB‐siRNA, in the intermediate myeloerythroid cocktail (SCF, IL‐3, Flt3, IL‐6, GM‐CSF). At the end of 5 d, commitment bias toward myeloid and erythroid lineages was checked using anti‐CD11B (CR3A) and anti‐CD71 (TfR), as shown in histograms, respectively (Fig. 4C). MafB‐siRNA decreased myeloid bias while augmenting erythroid bias of CMPs. CD11B expression was the highest in burn CMPs compared with sham and was reversed to sham levels after silencing MafB in burn CMPs. Conversely, MafB‐siRNA increased CD71‐expressing cells, which were otherwise decreased following burn compared with sham (Fig. 4C, lower). Representative FACS plots from sham and burn CMP‐derived cultures, with and without MafB‐siRNA, are presented in Fig. 4D. Box plots in Fig. 4E enumerate the percentage of CD71⁺ cells (upper) and MFI of Tfrs (lower) in CMPs from sham and burn mice with control‐siRNA and MafB‐siRNA. The results shown here are median (middle line) and the lowest and highest values (whiskers) from n = 4. Experiments were repeated 2 other times with similar results.

The direct role of MafB in myeloid (granulocyte/monocyte) vs. erythroid bias of CMPs following burn is revealed with MafB gene silencing. MafB‐siRNA in CMPs from burn mice partially recovered erythroid‐specific TfR expression. Furthermore, CD11B and M‐CSFR expressions are reduced independently with MafB knockdown (siRNA) in CMPs from burn mice. The central mechanism orchestrating CMP → MØ vs. CMP → EB fate can be explained by the direct role of high MafB‐expressing CMPs following burn.

Gain of function with GATA1 restores CMP → EB that was decreased by burn injury

To study the plasticity of CMPs and the role played by transcription factor dynamics, we introduced a stable transfection with gfp‐GATA1 retrovirus. GATA1 is a erythroid‐specific counter‐regulatory transcription factor expressed at low levels in CMPs [28]. Approximately 60% of CMPs in either group were GATA1⁺, but the MFI of ic‐GATA1 protein expression in CMPs from burn mice was reduced significantly ( Fig. 5A ). To confirm the pliability of CMPs, we tested whether increasing the GATA1 axis can now respecify burn CMPs to differentiate toward erythroid lineage. CMPs from sham and burn mice were placed in a proliferation cocktail (SCF + IL‐3) overnight before transfecting with gfp‐GATA1 retrovirus or gfp‐empty vector. The cultures were maintained in intermediate myeloerythroid cocktail for 5 d, and then replaced with EB cocktail for 5 more d. CMP → EB development was determined on d 10 by measuring CD71⁺Ter119⁺ expressions by FACS. Transfection with gfp‐GATA1 increased the proportion of EBs to 86% in burn and 79% in sham CMPs, respectively. Empty vector transfection in CMPs did not have any effect, showing only 49% of burn and 64% of sham CMP → EB cultures as CD71⁺Ter119⁺ EBs. Overall, GATA1 transfection improved the erythroid potential of CMPs (Fig. 5B).

Next, we pregated for gfp^pos and gfp^neg fractions from empty vector or GATA1 vector‐transfected CMP → EB cultures and compared the proportion of EBs in gfp^pos vs. gfp^neg fractions. Approximately 30% transfection efficiency was reached, irrespective of GATA1 or empty vector (histograms in Fig. 5C and D). Nonetheless, the majority of CD71⁺ Ter119⁺ EBs resulted from the gfp^pos fraction compared with the gfp^neg fraction (gfp^pos: 60%; gfp^neg: 22%), accounting for the overall increase in EBs to gfp‐GATA1 vector in burn CMPs, whereas in empty vector‐transfected cells, CD71⁺Ter119⁺ EBs remained low with gfp⁺ (37%) or gfp^neg (26%) fractions from burn CMPs. Furthermore, the specificity of gfp‐GATA1 vector vs. gfp‐empty vector in CMPs was confirmed by anti‐GATA1 antibody staining. Figure 6A gives the representative confocal images of CMPs taken at 72 h after transfection, indicating that GATA1 protein is only present with gfp‐GATA1 vector and absent with gfp‐empty vector.

To avoid any myeloid cell contamination, we gated for CD11B^neg cells before selecting for gfp⁺ cells (histograms in Fig. 6B). Stages of EB development are defined by differential expressions of CD71 and Ter119, as CD71^hiTer119^med, CD71^medTer119^med, and CD71^medTer119^hi, and they get smaller in size progressively as they mature into erythrocytes. We confirmed the development of CMP → EB cultures, using ImageStreamX, which simultaneously images the cell in brightfield besides flow cytometry. Accordingly, few representative images from the CD11B^neg gfp‐GATA1⁺ fraction showing differential expressions of CD71⁺ and Ter119⁺ are presented in Fig. 6B. EBs (gfp^pos) in the basophilic stage (CD71^hiTer119^med) are relatively bigger in size, as seen in brightfield. As they progress through the maturation process toward orthochromatic (CD71^medTer119^med) and reticulocyte stages (CD71^medTer119^hi), CD71 is reduced, and Ter119 is intensified and eventually enucleated to form Ter119⁺ erythrocytes.

Response to graded burn size

Next, we inquired whether an anemic response is graded with respect to burn size. Based on the survival curve, 14/15 mice were dead in the 30% TBSA group by 24 h after injury. Despite adequate resuscitation measures, we incurred high mortalities (7/15) in the 20% TBSA group by PBD #14, whereas all animals survived in the 15% TBSA and 10% TBSA groups for up to 21 d ( Fig. 7A ).

Figure 7.

Graded burn injury on survival and BM responses. (A) Larger burns (30%) were detrimental in small animals, such as mice. Whereas the majority (80%) of them survived 20% TBSA burns on PBD #7, only 50% survived beyond 7 d. Whereas no mortality was found in 10 and 15% TBSA groups when compared with shams, all 3 are overlapping in the survival curve. (B) Bar graphs showing the impact of graded burn size on erythroid and myeloid cell distributions per million total bone reduction in erythroid cells (CD11B^neg Ter119⁺) with a shift favoring myeloid (CD11B⁺ Ter119^neg) cells compared with sham. The trend was graded, with increasing burn size exhibiting significant differences in myeloid cells in mice with 20% TBSA compared with 10 and 15%, respectively. n = 7–15/group. *P < 0.001 vs. sham; ^a P < 0.03 vs. 20% TBSA; ^b P < 0.01 vs. 20% TBSA.

To avoid selecting out just the survivors, we then chose the PBD #7 time point to harvest blood and femurs when the majority of mice in the 20% group survived. We assessed BM myeloid (CD11B⁺ Ter119^neg) and erythroid (CD11B^neg Ter119⁺) responses to varying burn sizes of 10, 15, and 20% TBSA full‐thickness scald burn. Whereas erythroid bias was reduced significantly compared with sham in all burn groups, the myeloid response was inversely and significantly increased, which was more pronounced in the 20% TBSA than in the 10% TBSA groups (Fig. 7B). Moreover, peripheral blood also exhibited significant perturbations in 10% TBSA burn harvested on PBD #7 compared with sham mice. Whereas RBCs, Hct, Hgb, and granulocyte counts exhibited a graded response between 10 and 20% TBSA burn, perturbation in monocytes was high even with a 10% TBSA burn ( Table 2 ).

Table 2.

Graded burn size and anemic response

Exp. groups (PBD #7)	Hgb (g/dl)	Hct (%)	RBCs (106 cells/µl)	Monocytes (%)	Granulocytes (%)
20% TBSA	12.8 ± 0.23 a , b	40.0 ± 0.59 a , c	8.9 ± 0.19 a , b	10.6 ± 0.67 a	47.3 ± 1.1 a , c
10% TBSA	13.9 ± 0.18 d	43.7 ± 0.72 e	9.6 ± 0.15 e	10.3 ± 0.45 a	31.1 ± 3.2 d
Sham	14.7 ± 0.07	45.7 ± 0.24	10.2 ± 0.07	5.8 ± 0.25	22.1 ± 2.0

Heparinized blood was obtained via cardiac puncture from sham and burn (10% and 20% TBSA) mice harvested on PBD #7. Divergent and reciprocal responses in erythroid and myeloid cells increased with increasing burn size.

^a P < 0.0001.

^b P < 0.001.

^c P < 0.0001 vs. 10% TBSA burn.

^d P < 0.01.

^e P < 0.05 vs. sham.

DISCUSSION

BM response is stimulated after a critical burn injury. LSKs are increased, and there is a steady proportion of CMPs, albeit gaining heterogeneity leaning toward GMPs and away from MEPs, which is reflected in peripheral blood monocytes, granulocytes, and RBCs during the entire observation (21 d). Specifically, elevated MafB in CMPs following burn trauma augments M‐CSFRs (CD115) and CR3A (CD11B) and inhibits Tfrs (CD71) biasing the trilineage potency of CMPs to produce less erythrocytes and more MØs vs. DCs, dictating a pDC phenotype. The current finding establishes myeloerythroid reprioritization [31] after burn injury as a basis for anemia of critical illness [34] and immune dysfunction [16] and uncovers the hematopoietic mechanisms of blunted erythropoiesis reported in burn patients [35].

In the current study, burn size of 10% and 15% TBSA induced significant perturbations in erythroid and myeloid cells in blood and BM and to a greater degree, in 20% TBSA, indicating a graded anemic response with respect to burn size. The drop in Hgb levels in burn mice is somewhat less profound than in burn patients, as mice were not subjected to surgeries, phlebotomies, and dressing changes, all of which exacerbate anemia in patients. Unlike in burn patients with ≥20% TBSA, these small animals met with high mortalities (>50%) by PBD #14, thus supporting 15% TBSA full thickness burn as the relevant model to study erythropoietic and myelopoietic perturbations in BM circumventing survival benefit. Moreover, in mice with 15% TBSA full‐thickness burn, blood Hgb fell below levels considered as anemic [32].

Results are compelling to perceive that similar implications may be applied in ICU patients who also present with erythropoietin refractory anemia, based on a report from a prospective observational cohort study, where anemia continued to persist for 6 mo after discharge in >½ of ICU survivors with no pre‐existing hematologic disorders [36].

Centered on pioneering work by Weissman's group [28], we have used flow cytometric investigations of liquid cultures to delineate further the terminal tripotential fate of CMPs. This has allowed us to investigate the predominant role played by lineage‐specific transcription factors MafB and GATA1 following burn pathology. Previously, we detected a burn‐mediated transcriptional imprint in the BM GMPs, where myeloid transcription factor MafB was up‐regulated altering their terminal fate bias to produce more MØs over DCs in a mouse model of burn injury [18]. Subsequently, we showed that monocytes obtained from burn patients expressed high levels of MafB, impeding monocyte‐derived DC differentiation potential [19]. GMPs and monocytes give rise to MØ and cDCs (also called myeloid DCs) [27, 37] with no erythroid potential, whereas CMPs can produce pDCs and cDCs [29], as well as MØs and erythrocytes [28]. The current finding in CMPs corroborates our previous studies on the role of MafB in respecifying MØ, DC, and RBC development in burn pathology ( Fig. 8 ).

Figure 8.

Schematic summary. Tripotential lineage commitment of CMPs into RBC, MØ, and DC is respecified by the altered transcription factor MafB /GATA1 axis following burn injury.

Similar to GMPs, an increase in MafB and M‐CSFR expressions facilitates a MØ bias in burn CMPs, but it seems contradictory for a pDC bias. Administration of Flt3L stimulates DC production [25], albeit in BM; a more committed CDP for cDC and pDC exists downstream of CMPs [29, 38]. CDPs (Lin^negc‐Kit^medFlt3⁺M‐CSFR⁺) and stem and progenitor cells (Lin^negc‐Kit^hi) can generate both pDC and cDC in response to the cytokines Flt3L and GM‐CSF. When supplemented with M‐CSF in place of GM‐CSF, CDPs could generate only pDCs, whereas Lin^negc‐Kit^hi cells generated both pDC and cDC [26], thus suggesting the role of M‐CSFR in pDC development. In fact, human BM progenitors with pDC potential [39] express M‐CSFR. Therefore, it is likely that increased M‐CSFR expression in burn CMPs dictates the pDC phenotype compared with CMPs from sham‐treated mice, and furthermore, the M‐CSF‐rich milieu, such as in inflammatory settings, for example, burn injury, can influence the fate of the DC phenotype. In a recent report, burn injury had a more accentuated effect on pDCs than on cDCs [40].

We have demonstrated a reduction in MHC‐II expression in burn CMP → DC compared with sham. Loss of the MHC‐II molecule impairs key cell‐to‐cell interaction in T cell priming [41]. Current studies did not investigate the direct effect of MafB on MHC‐II expression in CMP‐derived DCs following burn, as a result of the confines of the siRNA technique in long‐term cultures. We have previously rationalized in a mouse model of burn injury that MafB was up‐regulated in the BM GMPs, and their terminal fate bias was altered, producing DCs with diminished MHC‐II expression [18]. Furthermore, monocytes, the progeny of GMPs from burn patients, also expressed high levels of MafB, producing monocyte‐derived DCs with diminished MHC‐II expression [19]. The silencing of MafB in monocytes and GMP cultures improved CD11c⁺ MHC‐II⁺ DC generation [18, 19]. Together, our findings that burn injury abrogates MHC‐II expression in CMP → DCs could explain the predisposition to infection in burn patients.

With regards to methods for confirmation, a logical approach is to investigate in Mafb^−/− mice. However, in vivo knockout of MafB gene is fatal as a result of respiratory arrest at birth [42], and this prohibits the in vivo knockout model. Therefore, we pursued gene silencing of MafB (siRNA, in vitro) and induced overexpression of counter‐regulatory transcription factor GATA1 to provide sure enough evidences. Given the time‐sensitive commitment, high‐proliferation rate, and the significance of MafB in CMPs (myeloid cells), siRNA was best suited for quick and partial silencing over the long‐lasting RNA interference effect with short hairpin RNA. Results obtained from MafB siRNA in burn CMPs support the hypothesis that the burn‐mediated increase in the primary intrinsic transcription factor MafB biases the terminal fate of CMPs toward monocytosis by augmenting myeloid‐specific CR3A (CD11B) and monocyte‐specific M‐CSFR (CD115). On the other hand, MafB siRNA in burn CMPs improves erythrocyte production by increasing erythroid‐specific Tfrs (CD71) in particular. This conception is consistent with the study where erythroid differentiation of HD3 cells was inhibited as a result of 2‐ to 3‐fold repression of the endogenous Tfr gene when MafB was induced in the chicken EB cell line [43].

PU.1 plays a central role along the myelomonocytic pathway by antagonizing C/EBP‐α. Likewise, PU.1 overexpression blocked DNA binding of the GATA1 fusion protein, as well as GATA1‐mediated erythroid differentiation of G1ER cells [44]. Therefore, it is reasonable to speculate PU.1 as the potential player in erythromyeloid bifurcation, but PU.1 levels remain unchanged in CMPs following burn injury (Supplemental Fig. 2), and moreover, MafB levels are higher in CMPs (Fig. 4 and Supplemental Fig. 2), eliminating the possible repression by PU.1 in early myeloid progenitors in burn pathology. We show that PBD #7 CMPs have a reciprocal relationship between myeloid‐specific MafB and erythroid‐specific GATA1. Our studies support that MafB plays a role in dampening GATA1 and/or interfere in erythroid‐specific genes, such as Tfrs. Gain‐of‐gene function using the gfp‐GATA1 stable transfection demonstrates the high degree of plasticity exhibited by CMPs and the transcription factor dynamics in lineage commitment and terminal fate decisions after a critical burn injury. An abnormal MafB/GATA1 axis respecifies the myeloerythroid commitment pattern, leading to DC deficits and an increase in deactivated monocytes, thus dictating immune dysfunction [10] presumably initiated by extrinsic factors following burn injury. Nonetheless, whether the interplay is direct needs to be determined.

As illustrated in the schematic summary (Fig. 8), altered transcription factor dynamics in CMPs is the central mechanism causing anemia, monocytosis, and DC deficits following burn injury. Specifically, burn‐induced MafB in CMPs acts as a transcriptional activator of M‐CSFR and a repressor of TfR promoting MØ and inhibiting erythroid differentiation while dictating a pDC phenotype. Therefore, a perturbed MafB/GATA1 axis following burn injury unfolds a plausible mechanism causing erythropoietin‐resistant anemia and immune suppression [16], hematopoietic reprioritization responses [31], and a limited supply of CFU‐erythroid in burn patients [35]. Therefore, the identification and regulation of burn‐mediated extrinsic factors that can restore MafB/GATA1 balance can be a therapeutic target for anemia of critical illness and reduce transfusion needs.

AUTHORSHIP

N.B.J., K.M., and R.S. designed, conducted, and analyzed experiments and wrote the paper. A.S. and L.K.H. conducted animal procedures. K.M., R.L.G., and R.S. designed, conducted, analyzed, and interpreted experiments and wrote the paper. J.A.P. consulted on and contributed during the manuscript preparation.

ACKNOWLEDGMENTS

This work was supported by U.S. National Institutes of Health (NIH) Grants R01DK097760‐01 (to K.M.) and NIH Training Grant T32 GM008750 (to R.L.G.). The authors thank Heather LaPorte (Loyola University Chicago) for processing mouse blood in the HemaTrue Veterinary Hematology table‐top analyzer for CBC results. The authors are grateful for Patricia Simms and the staff of the Flow Cytometry Core Facility at Loyola University Chicago for expert assistance with cell sorting and AMNIS images. The authors thank Linda Fox (Loyola University Chicago) for her assistance in confocal imaging. The authors also thank Shirin Hasan (Loyola University Chicago) for technical assistance with parts of this study and Dylan Tromblay (Loyola University Chicago) for preparing keys to figures.

Supporting information

Supplementary Material Files

Supplementary Material Files