C57BL/6J mice were purchased from Jackson Laboratories (Bar Harbor, ME, USA). The global LXRα-KO mice were a gift from Dr. Basilia Zingarelli (Cincinnati Children's Hospital Medical Center, Cincinnati, OH). All mice were maintained and bred under specific pathogen-free conditions in the Division of Laboratory Animal Resources at the University of Cincinnati Medical Center. All animal experiments conformed to the Guidelines for the Care and Use of Laboratory Animals prepared by the National Academy of Sciences, published by the National Institutes of Health, and approved by the University of Cincinnati Animal Care and Use Committee.

Mouse polymicrobial sepsis was surgically induced by CLP as described previously (27). In brief, to determine whether rGDF3 had any preventive effects in vivo against polymicrobial sepsis, recombinant mouse GDF3 Protein (20 μg/kg body weight) (R&D Systems, Cat. # 9009-GD-010) or BSA vesicle was injected into the tail vein of wild-type (WT) mice (8-week old, male) 3 h before CLP. Considering the cyclic hormonal changes during the ovulatory cycle in female mice might cause excess variability, male WT mice were used in the CLP model. Then, mice were anesthetized by intraperitoneal injection of a mixture of ketamine (100 mg/kg) and xylazine (5 mg/kg). A midline laparotomy was then performed. The cecum was exposed and ligated at 1.5 cm from the distal end with a 5–0 sterile silk suture. A single through and through puncture was made at the middle between the ligation and distal end of the cecum with a 21-gauge needle. After puncturing, the cecum was gently compressed to extrude a small amount of cecal content and returned to the abdominal cavity. The abdominal wall incision was closed in layers. After surgery, pre-warmed phosphate buffered saline (0.05 mL/g body weight) was administered subcutaneously. Postoperative pain control was managed with 0.05 mg/kg buprenorphine every 12 h. While, to examine the therapeutic effects of rGDF3 in WT mice upon CLP surgery, rGDF3 (100 μg/kg body weight), or BSA vesicle was injected into the tail vein 1 h after CLP. In order to assess the actual effect of rGDF3 on the survival of septic mice, antibiotics was not used in the CLP model. The animal survival rate was monitored every 6 h for 7 days after CLP operation. Whole blood and peritoneal lavage fluid (PLF) were harvested from these mice to measure bacterial burden, cytokines, and markers of organ injury at 20–24 h post-CLP surgery. Bacterial burden assay was performed as described previously (28).

Lung tissues were collected from mice at 24 h post-CLP. All lungs were perfused via the heart, inflated and fixed with 10% buffered formalin for more than 2 days, then embedded in paraffin, and cut into 5-μm sections. Tissue sections were stained with hematoxylin and eosin (H&E). To evaluate the lung injury, 6–8 independent random lung fields were evaluated per mouse for neutrophils in alveolar spaces, neutrophils in the interstitial spaces, hyaline membranes, proteinaceous debris filling the airspaces, and alveolar septal thickening and weighted according to the relevance ascribed by the official American Thoracic Society workshop report on features and measurements of experimental acute lung injury in animals (29).

To quantify the spleen and lung edema, whole spleen and lung tissues were collected at 24 h post-CLP, rinsed to remove surface blood and patted dry, and the immediate weights of the samples were recorded as the wet weight. The tissues were air dried for 48 h at 60°C, and the weights were recorded as the dry weight. A wet/dry weight ratio for each individual mouse was calculated.

Serum levels of alanine aminotransferase (ALT), a biochemical marker of liver injury, were measured using an ALT assay kit (MyBioSource, Cat. # MBS264717). Serum levels of creatinine (Cr), a biochemical marker of kidney injury, were determined using a creatinine assay kit (MyBioSource, Cat. # MBS2504918), according to manufacturers' protocols.

Blood and PLF were collected at 20 h after CLP surgery and serum samples were prepared by centrifuged at 3,000 g for 15 min at 4°C. All serum was stored in −80°C until use. ELISAs were performed to measure the levels of TNF-α and IL-6 in sera and PLF using commercially available kits [Peprotech, Cat. # 900-K54 (TNF-α), Cat. # 900-K50 (IL-6)], according to the manufacturer's protocols.

For the isolation of peritoneal macrophages (PMs), 5 mL PBS was used to lavage the abdominal cavity of euthanized mice for three times. The PLF was collected and centrifuged (2,000 rpm, 5 min). The cell pellet was then resuspended in 10 mL fresh prepared Dulbecco's modified Eagle medium (DMEM) media with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. The cells were allowed to adhere to the substrate by culturing them for 4 h at 37°C. Non-adherent cells were removed by gently washing three times with warm PBS. Adherent cells were cultured for 18 h, and then used for experiments.

Bone marrow-derived macrophages (BMDMs) were isolated cultured/differentiated into macrophages as described previously (30). In brief, mice were terminally anesthetized, bone marrow from the tibias and femurs of mice was flushed out using cold sterile wash medium (DMEM), and filtered through 70 μm Nylon cell strainer. Subsequently, these mononuclear cells were centrifuged at 500 g for 5 min at room temperature (RT). The resulting cell pellet was re-suspended in complete culture medium (DMEM supplemented with 10% of L929 cell culture supernatant, 10% FBS, 1% penicillin/streptomycin solution, and 1% HEPES), and cultured in a humidified atmosphere of 5% CO₂ at 37°C for 7 days, culture medium was changed on Day 4 of culture.

RAW264.7 macrophages (mouse macrophage cell line) were cultured in incubator at 37°C with 5% CO₂. The cells grew in DMEM, containing 10% FBS and 1% penicillin/streptomycin.

BMDMs and RAW264.7 macrophages seeded into 96-well plates were treated with different rGDF3 doses for 18 h. Cell viability was measured by addition of 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenol)-2-(4-sulfophenyl)-2H-tetrazolium (MTS, Promega, USA) for 1.5 h at 37°C. The absorbance was measured spectrophotometrically at 490 nm with a microplate reader (ELx800, Bio-Tek, USA) according to the manufacturer's protocols.

To analyze the phagocytic capacity of RAW264.7 macrophages and BMDMs isolated from WT mice, cells were seeded into 96-well plates (4 × 10⁴ cells/well; or 1 × 10⁵ cells/well) or on coverslips in 24-well plates (5 × 10⁴ cells/well; or 2 × 10⁵ cells/well) in DMEM medium and allowed to adhere for 24 h. Cells were treated with BSA, rGDF3 (20 ng/mL), GW3965 (1 μmol/L) or GSK2033 (2 μmol/L) + rGDF3 (20 ng/mL) for 18 h, followed by the addition of pHrodo red E. coli BioParticles (Thermal Fisher Scientific, Cat. # P35361) diluted in medium according to the manual. Subsequently, cells were incubated with BioParticles for 1.5 h at 37°C. The fluorescence intensity was measured using a GloMax®-Multi Detection System (Promega). While cells cultured on coverslips were stained with CellMask™ Green Plasma Membrane Stain (Invitrogen, Cat. # C37608) as described in the manual. Then, cells were fixed with 4% paraformaldehyde (15 min RT). After washing three times with PBS, coverslips were mounted onto slides using a ProLong™ diamond antifade mountant reagent with DAPI (Invitrogen, Cat. # P36962). Subsequently, slides were imaged with a confocal LSM 710 (Carl Zeiss Microimaging, Jena, Germany). Images were recorded with ZEN (Black) and analyzed with ImageJ software (Wayne Rasband, National Institutes of Health, Bethesda, MD).

PMs isolated from WT mice were seeded into 6-well plates (8 × 10⁵ cells/well) in DMEM medium and allowed to adhere for 18 h. Cells were treated with BSA or rGDF3 (20 ng/mL) for 18 h, followed by the pHrodo red E. coli BioParticles according to the manual. Subsequently, cells were incubated with BioParticles for 1.5 h at 37°C. Then, cells were collected and fixed with 2% paraformaldehyde (15 min on ice). After washing with FACS buffer, Flow cytometry analysis was performed.

To determine bactericidal activity, a classical CFU assay was conducted with minor modifications as described previously (31). Briefly, E. coli DH5 alpha were grown overnight in LB at 37°C. After that, bacteria were quantified, pelleted, and washed with PBS. BMDMs or RAW264.7 macrophages were placed in 12-well-culture plates and pre-incubated for 6 h to allow to adhere. Subsequently, cells were treated with BSA or rGDF3 (20 ng/mL) in antibiotic-free fresh medium for 18 h. Cells were infected with live E. coli at a multiplicity of infection (MOI) of 20 bacteria/cell ratio for 1 h at 37°C with 5% CO₂. Fresh DMEM containing 100 μg/mL gentamicin was added and incubated for 30 min to kill extracellular bacteria. To determine internal bacteria, macrophages were washed two times with cold PBS and lysed. The cell lysate was diluted serially and plated on the LB plate to determine the colony-forming unit (CFU) count. To assess the bactericidal activity, cells were incubated for another 6 h. The number of bacteria remained after intracellular killing was determined by lysing cells and plating the cell lysate as described in above. The killing percentage was calculated by using the formula [(CFU count at 1 h) – (CFU count at 6 h)/(CFU count at 1 h)] × 100%.

Total RNA was extracted from cultured cells using the RNeasy kit (Qiagen, Cat. # 217004) in accordance with the manufacturer's instructions. cDNA was synthesized from 0.5–1.0 μg total RNA using Superscript II Reverse Transcriptase (Invitrogen, Cat. # 18080044). Then, qRT-PCR was performed in triplicate with the ABI PRISM 7900HT sequence detection system (ABI) using SYBR green (Genecopoeia, Cat. # QP005). Relative mRNA levels were calculated and normalized to 18S rRNA. The primers were used as follows: CD5L forward 5'-TTGGAGAACAACTGTACCCATGGC-3', reverse 5'-AGGCTGAGGGAAAGGTGTCTAAAG-3'; Bcl2a1b forward 5'-GTTTCCAGTTTTGTGGCAGA-3', reverse 5'-CCCAGAACTGTCCTGTCATC-3'; MRPS18C forward 5'-GCATTTATGGAAGGCACATAAC-3', reverse 5'-TGGCAGGAACTTCACAAATAC-3'; 18S rRNA forward 5'-GCAATTATTCCCCATGAACG-3', reverse 5'-GGCCTCACTAAACCATCCAA-3'. Directional polyA RNA-sequencing was performed by the Genomics, Epigenomics, and Sequencing Core (GESC) at the University of Cincinnati.

BMDMs treated with rGDF3 (20 ng/mL) and GW3965 (1 μmol/L) (DMSO used as control) were fixed on the coverslip, blocked with 1% BSA, 0.3% Triton X-100, 0.1% Tween 20 and incubated with primary antibody (LXRα, Invitrogen, Cat. # PA1-330) diluted in 1% BSA, 0.1% Tween 20 overnight at 4°C. After washing, BMDMs were incubated with secondary antibody at RT for 1 h. Thereafter, cells were washed three times with PBS and mounted with ProLong diamond Antifade Mounting medium with DAPI (Invitrogen, Cat. # P36962). Images were captured with Zeiss LSM710 LIVE Duo Confocal Microscope (Live Microscopy Core, University of Cincinnati).

Statistical calculations were performed with GraphPad Prism 7.0 software (GraphPad Software, Inc., USA). The experimental data were expressed as mean ± SEM. Differences between two groups were analyzed using unpaired two-tailed t-test and survival rates were analyzed using log rank test. p < 0.05 was considered to be statistically significant.

Our recent work showed that rGDF3 suppressed inflammatory response and improved cardiac function and survival in endotoxin-induced sterile inflammation mouse model. To explore whether rGDF3 could provide a preventive effect in response to polymicrobial sepsis, WT mice were received rGDF3 (20 μg/kg body weight) or BSA vesicle via the tail vein injection 3 h prior to CLP surgery (Figure 1A), and then mortality was monitored over 7 days. We observed that rGDF3-treated mice exhibited significantly higher survival rate than BSA-treated mice over a period of 7 days post-CLP (Figure 1B). Median survival was 76 h in rGDF3-treated group and 41 h in BSA-treated group (Figure 1B, p < 0.05). Next, we assessed organ injury in BSA- and rGDF3-treated mice at 24 h post-CLP by: (1) conducting histopathological analysis, (2) determining tissue edema, and (3) measuring biomarkers of organ damages in the serum. Firstly, we found that the administration of rGDF3 significantly rescued CLP-induced lung injury, as evidenced by the decrease in neutrophil infiltration, formation of hyaline membranes, thickness of alveolar wall, and the alveolar collapse in lung tissues collected from rGDF3-treated mice (injury score 0.264 ± 0.025), compared to those from BSA-treated mice (injury score 0.452 ± 0.034) (Figures 1C–E). In addition, rGDF3-treated mice exhibited lower ratio of wet weight/dry weight than that from BSA-treated mice (4.325 ± 0.089 vs. 4.558 ± 0.047) (Figure 1F). Similar results were observed for the wet weight/dry weight ratio of the spleen (4.462 ± 0.041 vs. 4.639 ± 0.069) (Figure 1G). Furthermore, we tested serum levels of alanine transaminase (ALT) and creatinine (Cr), biomarkers of liver injury and kidney damage, respectively. As shown in Figure 1H, I, the concentrations of both biomarkers were significantly lower in rGDF3-treated mice, compared to BSA-treated mice at 24 h post-CLP. Collectively, these data indicate that pre-administration of rGDF3 significantly decreases CLP-induced multi-organ injury in mice.

Since bacterial burden and inflammatory responses are important driving factors for the pathogenesis and progression of sepsis (1, 6, 7), we next sought to determine the effects of rGDF3 on these aspects. The bacterial burdens in the blood and PLF were examined at 20 h post-CLP surgery. As shown in Figures 2A,B, the rGDF3-treated group presented significantly lower bacterial load in blood and PLF compared to control group (BSA-treated). Representative images for CFUs in both blood and PLF were shown in Figures 2C,D. In parallel, results from ELISA analysis showed that the levels of inflammatory cytokines IL-6 and TNF-α were significantly suppressed by rGDF3 administration in both sera (Figures 2E,F) and PLFs (Figures 2G,H). Together, these results suggest that the beneficial effects of rGDF3 on survival and organ damage may be associated with its ability to control bacterial burden and inflammatory responses.

Considering that macrophages play an essential role in bacterial clearance during polymicrobial sepsis (12), we next conducted bacterial phagocytosis and intracellular killing assays in three different sources of macrophages: (1) RAW264.7 cell line, (2) BMDMs, and (3) PMs isolated from WT mice. Firstly, we investigated the effect of rGDF3 on the cell viability with MTS assay. As shown in Supplementary Figures 1A,B, stimulation with different rGDF3 doses (10, 20, 50 ng/mL) did not affect the viability of BMDMs and RAW264.7 cells. Next, using the fluorescence-conjugated E. coli BioParticles, we found that treatment of macrophages with rGDF3 at doses of 20 and 50 ng/mL greatly increased the uptake of E. coli BioParticles in both BMDMs (Figure 3A) and RAW264.7 cells (Figure 3B). Of interest, rGDF3 at a lower dose (10 ng/mL) significantly enhanced phagocytosis of E. coli BioParticles in RAW264.7 macrophages but not BMDMs (Figure 3B). To further validate these results, the uptake of fluorescence-conjugated E. coli BioParticles was examined under confocal microscopy. We found that the intensity of red particles in rGDF3-treated BMDMs was significantly higher than that in control BSA-treated cells (Figures 3C,D). Similar findings were also observed in rGDF3-treated RAW264.7 cells (Figures 3E,F). In addition, flow cytometry analysis results further showed that rGDF3-treated PMs displayed higher content of red fluorescence (Figures 3G,H). Thus, these data generated from all three different analysis methods consistently indicate that rGDF3 is able to augment macrophage phagocytosis. Lastly, we assessed the phagocytic and bactericidal activities of BMDMs and RAW264.7 cells using live E. coli, as described in Figure 4A. We found that rGDF3 treatment significantly increased the bacterial uptake by 2.64-fold in BMDMs (Figure 4B) when compared to BSA-treated cells. The numbers of bacteria remained in BMDMs at 6 h after intracellular killing were shown in Figure 4C. Remarkably, rGDF3 stimulation could boost intracellular killing activity by 26.59% in BMDMs, compared to BSA-treated controls (Figure 4D). Similar findings were also observed in rGDF3-treated RAW264.7 cells, in which the uptake of bacteria and intracellular killing activity were significantly increased, compared to BSA-treated cells (Figures 4E–G). Taken together, these data indicate that treatment of macrophages with rGDF3 could improve phagocytosis and bactericidal activities.

To identify potential mechanisms underlying the effect of rGDF3 in promoting macrophage phagocytosis of bacteria, we treated BMDMs with rGDF3 (20 ng/mL) or control BSA for 18 h, followed by isolation of total RNAs for RNA sequencing (RNA-seq). Unexpectedly, we identified only 6 differentially expressed genes, of which 4 were significantly upregulated and 2 were down-regulated (p < 0.05), as listed in Supplementary Table 1 and illustrated in Figure 5A. Of interest, among these 4 up-regulated genes, CD5L has been recently reported to stimulate the phagocytosis of bacteria in macrophages (32). CD5L, also referred to as apoptosis inhibitor 6 (Api6) or apoptosis inhibitor of macrophages (AIMs), is a soluble protein mainly produced by macrophages (32, 33). Importantly, the expression of CD5L is controlled by the activation of LXRα in macrophages (33). Indeed, another LXRα-controlled gene, Bcl2a1b (34), was also significantly upregulated in rGDF3-treated macrophages (Figure 5B). Therefore, we speculated that rGDF3 might enhance macrophage phagocytosis through activating LXRα pathway. To test this hypothesis, we treated BMDMs with GW3965, a potent LXRα agonist, followed by measuring the expression levels of CD5L and Bcl2a1b. qRT-PCR results showed that activation of LXRα by its agonist significantly increased mRNA levels of CD5L and Bcl2a1b, compared with control-treated macrophages (Figure 5C). In contrast, blockade of LXRα in BMDMs by its antagonist, GSK2033, offset rGDF3-mediated upregulation of CD5L and Bcl2a1b (Figure 5D). Given that the activation of LXRα is associated with its translocation to the nucleus (35, 36), as indicated in Figure 5E, we next determined whether rGDF3 could mimic the effect of LXRα agonist GW3965 on LXRα translocation, using immunofluorescence staining. We observed that the number of macrophages with nuclear translocation of LXRα was similar between rGDF3- and GW3965-treated groups (Figures 5F,G). Put together, these results demonstrate that rGDF3 could activate LXRα pathway in macrophages.

Next, we determined whether the activation of LXRα is responsible for rGDF3-mediated increase of bacterial clearance by macrophages. To this end, we pre-treated BMDMs with LXRα antagonist, GSK2033, followed by addition of rGDF3 (20 ng/mL). LXRα agonist (GW3965, 1 μmol/L) and rGDF3 alone were used as controls. Subsequently, phagocytosis of E. coli BioParticles was analyzed. We observed that treatment with both LXRα agonist and rGDF3 alone could significantly enhance BMDMs phagocytosis of E. coli BioParticles (Figure 6A). However, rGDF3 failed to improve phagocytosis of E. coli BioParticles in BMDMs pretreated with GSK2033 (Figure 6A). These results were also validated with confocal microscopy, as shown in Figures 6B,C that the intensity of red fluorescence in GW3965- and rGDF3-treated macrophages was significantly higher than in DMSO-treated cells. Meanwhile, the effects of rGDF3 on phagocytosis of E. coli BioParticles were blunted when BMDMs were pre-treated with LXR antagonist GSK2033, showing similar fluorescence intensity to the DMSO group (Figures 6B,C). Lastly, we isolated BMDMs from LXRα-KO mice (Figure 6D) and assessed CD5L expression by qRT-PCR after the treatment of BSA or rGDF3. As expected, rGDF3-treatment failed to increase the expression levels of CD5L in LXRα-KO BMDMs (Figure 6E). Furthermore, while treatment of LXRα-KO BMDMs with rGDF3 did not improve phagocytosis, loss of LXRα in macrophages significantly reduced the uptake of E. coli BioParticles under basal conditions (Figure 6F). Using living bacteria, we also observed that rGDF3-mediated enhancement of bacterial entry and intracellular killing was offset by knockout of LXRα (Figures 6G–I). Please note here that the entry of living bacteria into BSA-treated LXRα-KO cells (Figure 6G) appeared similar as BSA-treated WT cells presented in Figure 4B. This discrepancy may be ascribed to different batch of living bacteria. Indeed, higher number of living bacteria remained in LXRα-KO cells (Figure 6H), compared to WT-cells (Figure 4C). Accordingly, the percentage of intracellular killing in LXRα-KO cells (43%, Figure 6I) was lower than WT-cells (54%, Figure 4D) under basal conditions. Collectively, these results indicate that the activation of LXRα pathway is essential for rGDF3 to stimulate macrophages phagocytosis and killing of bacteria.

To further determine whether rGDF3 had any therapeutic effects in vivo against polymicrobial sepsis, we injected WT mice with a single dose of rGDF3 (100 μg/kg) or BSA vehicle at 1 h post-CLP surgery, followed by a series of experiments for the analysis of bacterial burden, cytokine levels, organ injury and mortality. We observed that post-administration of rGDF3 into CLP-mice significantly reduced bacterial burdens in the blood (Figure 7A) and PLF (Figure 7D) to a greater degree, compared to BSA-treated control mice. In addition, we found that the pro-inflammatory cytokines (i.e., IL-6 and TNF-α) levels in both serum (Figures 7B,C) and PLF (Figures 7E,F) were significantly lower in rGDF3-treated mice than controls. Furthermore, the ratios of wet/dry weight in lungs (Figure 7G) and spleens (Figure 7H) collected from rGDF3-treated CLP-mice were significantly reduced, compared to those samples from BSA-treated controls. Lastly, post-administration of rGDF3 into CLP-mice significantly improved survival outcome, as evidenced by 31.25% of rGDF3-treated mice (n = 16) survived at day 7 post-CLP, when only 6.25% of BSA-treated mice survived (n = 16). Meanwhile, the median survival was longer in rGDF3-treated mice than that in BSA-treated group (104 vs. 59 h) (Figure 7I, p < 0.05). Collectively, these data indicate that post-administration of rGDF3 could increase bacterial clearance, reduce inflammatory response, attenuate organ injury and improve animal survival in CLP-induced septic mice.