Macrophages Educated with Exosomes from Primed Mesenchymal Stem Cells Treat Acute Radiation Syndrome by Promoting Hematopoietic Recovery

Abstract

In the setting of radiation-induced trauma, exposure to high levels of radiation can cause an acute radiation syndrome (ARS) causing bone marrow (BM) failure, leading to life threatening infections, anemia and thrombocytopenia. We have previously shown that human macrophages educated with human mesenchymal stem cells (MSCs) by co-culture can significantly enhance survival of mice exposed to lethal irradiation. In this study, we investigated whether exosomes isolated from MSCs could replace direct co-culture with MSCs to generate exosome educated macrophages (EEMs). Functionally unique phenotypes were observed by educating macrophages with exosomes from MSC’s (EEMs) primed with bacterial lipopolysaccharide (LPS) at different concentrations (LPS-low EEMs or LPS-high EEMs). LPS-high EEMs were significantly more effective than uneducated macrophages, MSCs, EEMs, or LPS-low EEMs in extending survival after lethal ARS in vivo. Moreover, LPS-high EEMs significantly reduced clinical signs of radiation injury and restored hematopoietic tissue in the BM and spleen as determined by complete blood counts and histology. LPS-high EEMs showed significant increases in gene expression of STAT3, secretion of cytokines like IL-10 and IL-15 and production of growth factors like FLT-3L. LPS-EEMs also showed increased phagocytic activity, which may aid with tissue remodeling. LPS-high EEMs have the potential to be an effective cellular therapy for the management of ARS.

Introduction

Radiation, delivered therapeutically, accidentally, or maliciously during a terrorist attack, can lead to an acute radiation syndrome (ARS) with life threatening toxicities. High-dose radiation causes damage to highly proliferative cells found in the bone marrow, gastrointestinal tract and skin. ¹ Current standard of care involves supporting victims with antibiotics and transfusions until they can undergo an allogeneic bone marrow transplant (BMT) from a suitable donor. Unfortunately, the entire BMT process can often take weeks to identify and collect cells from a donor, and would be difficult to perform on a large-scale in the event of a widespread exposure. ² Moreover, allogeneic BMTs have their own set of complications, including engraftment failure, opportunistic infections and/or graft-versus-host-disease (GVHD), making them potentially as toxic as ARS itself. Consequently, identification of new therapeutic agents has received priority for treatment of ARS. However, to date the only approved agents are colony-stimulating factors like G-CSF and GM-CSF.^3,4 Unfortunately, these factors require the patient’s own hematopoietic stem cells to be effective ⁵, and do not address restoration of other critical hematopoietic cells including lymphocytes, erythrocytes and platelets.

“Off-the-shelf” cell-based therapies are attractive alternatives for repair of tissue injuries after ARS. Ideally, such cell-based therapies could be cryopreserved and then thawed for infusion after radiation exposure. Theoretically, since the patient is already immunosuppressed by radiation, the cell product should not require extensive tissue matching to the recipient. Among the cells actively explored for ARS are mesenchymal stem cells (MSCs) derived from the bone marrow (BM).^4,6 MSCs are capable of self-renewal and differentiation into osteocytes, chondrocytes and adipocytes. ⁷ MSCs have robust immunomodulatory and anti-inflammatory properties, as well as paracrine effects conductive to tissue regeneration.⁸ While the use of MSCs have shown promise in preclinical rodent models of radiation injury,^9–13 the lack of consistent cell engraftment indicate that MSCs may achieve their therapeutic effects via communication with other effector cells. Despite strong pre-clinical potential of MSCs, they have not yet been approved to treat ARS.

We have previously reported that co-cultivation of human macrophages with human MSCs generates MSC-educated macrophages (MEMs) ¹⁴ capable of prolonging survival in a lethal radiation injury xenogeneic model. ¹⁵ Trans-well studies indicated that secreted factors from MSCs were involved in alternatively activating macrophages to MEMs, although not to the same degree as direct co-culture.¹⁴ Recent studies have shown that a major mode of cell-to cell communication by MSCs involves the secretion of membrane bound extracellular vesicles (EVs). ^16,17 In addition, MSCs can shift to either a pro- or anti-inflammatory phenotype through the activation of surface Toll-like receptors (TLRs), by products released from pathogens such as lipopolysaccharide (LPS) or from damaged tissue via poly I:C. ¹⁸ In this study, we hypothesized that EVs could replace the need for MSC co-culture and educate macrophages to become radio-protective.

Here we determined, based on vesicle size, that EVs produced from unprimed or LPS primed MSCs consisted largely of exosomes. We either used exosomes from unprimed MSCs to produce exosome-educated macrophages (EEMs), or exosomes from MSCs primed with 2 different doses of LPS to produce exosome-educated macrophages (LPS-low or LPS-high EEMs). We found that LPS-high EEMs improved survival from lethal radiation injury in mice compared to vehicle, MSCs, macrophages, EEMs and LPS-low EEMs. The protective effects by LPS-high EEMs were due to restored hematopoietic tissue in the spleen and bone marrow, in part through secretion of cytokines and growth factors as well as increased phagocytic activity, which is needed for tissue remodeling.

Materials and Methods

Macrophage generation

Human peripheral blood mononuclear cells (PBMCs) were collected by density gradient separation using Ficoll-Paque Plus (endotoxin tested) (GE Healthcare Bio-Sciences, Piscataway, NJ, USA) from healthy stem cell donors using an institutional review board (IRB)-approved protocol. Red blood cells were lysed by incubating cells in ACK lysis buffer (Lonza, Walkersville, MD, USA) for 3 minutes. To reduce platelet contamination, cell suspensions were centrifuged at 300–700 rpm for 10 minutes and cell pellets were re-suspended and incubated with anti-human CD14 microbeads (Miltenyi Biotec, Auburn, CA, USA) for 15 minutes at 4°C. After washing unbound antibody, cell separation was done using an autoMACS Pro Separator (Miltenyi Biotec). Purified CD14⁺ monocytes were either plated into six-well cell culture plates at a concentration of 0.5–1 × 10⁶ per well for in vitro studies or 10⁷ per T75 cm² filter cap cell culture flask for in vivo studies (Greiner Bio-One, Monroe, NC, USA) in Iscove’s modified Dulbecco’s media (Gibco, Life Technologies, Grand Island, NY) supplemented with 10% human serum blood type AB (Mediatech, Herndon, VA, USA), 1× nonessential amino acids (Lonza, Walkersville, MD, USA), 4 mM L-glutamine (Invitrogen, Carlsbad, CA, USA), 1 mM sodium pyruvate (Mediatech), and 4 ug/mL recombinant human insulin (Invitrogen). Monocytes were cultured for 7 days at 37°C with 5% CO₂, without cytokines, to allow differentiation to macrophages. Cells were harvested by removing media, washing with phosphate-buffered saline (PBS, Hyclone, Logan, UT, USA) then using Accumax cell dissociation enzyme (Innovative Cell Technologies, Inc, (San Diego, CA, USA) to detach macrophages from the flask followed by the use of a cell scraper.

Isolation and characterization of EV’s from MSCs

MSCs were isolated from human bone marrow from the pelvis of a healthy bone marrow donor () and identity was confirmed by flow cytometry as previously described.¹⁴ Passage 4–8 BM-MSCs were grown in 75-cm² flasks to near complete confluence, washed once with PBS, and placed in StemPro MSC serum-free media (SFM) CTS (A103332–01, Gibco Life Technologies) for 18–24 hours. The conditioned culture media (CM) was collected. To prime MSCs to produce LPS exosomes, SFM was supplemented with either 100 ng/ml (LPS-low) or 1.0 ug/ml (LPS-high) with E. coli LPS O111:B4 (L4391 Sigma, St Louis, MO, USA). EVs were isolated from 4 different bone marrow isolates of MSCs (MSC-EV), LPS primed MSCs (LPS-low or high) or macrophages (macrophage-EV) by a 2-step centrifugation process as described. ¹⁹ Briefly, the CM was centrifuged using an Allegra® X-15R centrifuge (Beckman Coulter, Indianapolis IN, USA) at 2000x g at 4°C for 20 minutes. Clarified supernatant was centrifuged in an Optima™ L-80XP Ultracentrifuge (Beckman Coulter) at 100,000 g _avg at 4°C for 2 hours with a SW 28 rotor to pellet exosomes. The supernatant was carefully removed, and EV-containing pellets were resuspended PBS and pooled. For use and quantitation, we typically suspended the final EV pellet at 100 ul PBS /10 mLs of CM.

EVs were characterized for protein and RNA concentration using a NanoDrop spectrophotometer (Thermo-Fisher, Waltham, MA, USA). Mean and mode particle diameter and particle concentration were assessed at ZenBio Inc. using an IZON qNano Nanoparticle Characterization instrument (Zenbio Inc, Research Triangle Park, NC, USA) and compared using a Nanosight NS300 (Malvern, UK). Residual LPS levels were determined using a chromogenic LAL assay (VRL/Eurofins, Centennial CO) in EV’s of both unstimulated and LPS stimulated MSCs. To visualize the EVs by transmission electron microscopy (TEM), the re-suspended EVs were layered on a 30% sucrose cushion and re-centrifuged at 100,000 g _avg at 4°C for 2 hours, collected the upper portion of the cushion and re-centrifuged. The pellet was re-suspended in a small volume of PBS, whole mounted on Formvar EM grids and stained with uranyl acetate as described. ¹⁹

Education of macrophages by EVs derived from MScs or by MSC co-culture

For education of macrophages with exosomes, day 7 macrophages were supplemented with fresh media and treated for 3 days with either MSC-EVs, MSC-LPS-low EVs, MSC-LPS-high EVs or macrophage-EV’s preparations. Educated macrophages were never directly treated with LPS, however the residual LPS concentration from MSC-LPS-high EV preparations was 1.2ug/ml +/− 0.65 (mean +/− SD), similar to the concentration used during MSC priming. Six microliters of the EV preparation were used to educate 1 × 10⁶ macrophages in 10mL media. Macrophages not treated with exosomes were designated as “control” macrophages. Macrophages treated after EVs treatment were designated as EV/exosome-educated macrophages (EEMs), LPS-low EEMs, LPS-high EEMs and macrophage-EEMs, respectively. The amount of the exosome preparation (particles/ml of culture) used for education was standardized between groups.

Cytokine/chemokine/growth factor multiplex ELISA

Day 10 macrophages (10⁶/well) were grown in 6-well plates. The cells were washed with PBS, replaced with culture media, and incubated for 24 hours. The culture media was recovered, centrifuged at 300 × g for 10 minutes to remove any floating cell debris and assayed for secreted factors using a Milliplex MAP cytokine/chemokine multiplex magnetic bead panel (HCYTOMAG-60K, Millipore, Burlington MA), including epidermal growth factor (EGF), fibroblast growth factor (FGF-2), EOTAXIN, Transforming growth factor beta (TGF-β),granulocyte-colony stimulating factor (G-CSF), FMS-like tyrosine kinase 3 ligand (FLT-3L), granulocyte-macrophage colony-stimulating factor (GM-CSF), chemokine (C-X3-C motif) ligand 1 (FRACTALKINE), interferon alpha 2 (IFNα2), interferon-gamma (IFNγ), growth related oncogene (GRO), C-C motif chemokine 22 (MDC), platelet-derived growth factor (PDGF-BB), Soluble CD40 ligand (sCD40L), IL-1ra, IL-1a, IL-1b, IL-2, IL-4, IL-5, IL-6, IL-7, IL-8, IL-9 IL-10, IL-12p40, IL-12p70, IL-13, IL-15, IL-17, interferon gamma-induced protein 10 (IP-10), monocyte chemoattractant protein 1 (MCP-1), monocyte chemoattractant protein 3 (MCP-3), macrophage inflammatory protein 1a (MIP-1a), macrophage inflammatory protein 1b ( MIP-1b), regulated on activation, normal T cell expressed and secreted (RANTES), tumor necrosis factor alpha (TNF-α), tumor necrosis factor beta (TNF-β) and vascular endothelial growth factor A (VEGF-A) Twenty-five microliters of culture media was assayed in duplicate as directed by the manufacturer and detected on a Luminex xMAP platform.

Phagocytic assay

Phagocytic assays were performed using the pHrodo Green E. coli Bioparticle conjugate system (cat# P35366, Invitrogen) according to manufacturer recommendations. Activation of fluorescence of the pHrodo Green in this system is pH dependent and is detected only when internalized within the phagosome of the cell. Green E. coli Bioparticle conjugate was reconstituted in PBS, diluted in media, and incubated with the cells for 1 hour at 37C. Cells were then washed with cold PBS three times to reduce non-specific attachment collected using a cell scraper. Collected macrophages were then treated with Fc Receptor blocker for 10 minutes and stained for 20 minutes at 4° C with PerCP 5.5-CD14 and with APC-CD90 (5E10, cat # 328113, Biolegend) for gating of MEMs. CD14 positive/pH pHrodo Green positive cells at Ex/Em of 509/533 nm were detected on the MACSQuant analyzer 10 (Miltenyi Biotec) and analyzed using FlowJo software (FlowJo, Inc, Ashland, OR).

Acute radiation syndrome in vivo model

Male and Female NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ (NSG) mice were purchased from The Jackson Laboratory (Bar Harbor, Maine) and used at 8–16 weeks of age. All animals were bred and housed in a pathogen-free facility throughout the study. The Animal Care and Use Committee at the University of Wisconsin approved all experimental protocols.

On day 0, age matched NSG male and female mice received 4 Gy lethal total body irradiation using an X-RAD 320 X-ray irradiator (Precision X-Ray, North Branford, CT, USA). Four hours after radiation injury, mice were treated intravenously in the tail vein with 100 ul of either PBS (PBS), or 1×10⁶ human MSCs, 1×10⁶ control macrophages, 1×10⁶ LPS-stimulated macrophages, 1×10⁶ EEMs, 1×10⁶ LPS-low EEMs or 1×10⁶ LPS-high EEMs. Mice were monitored at least 3 times a week for clinical scores and survival. Clinical scores were determined based on a modified clinical scoring system as previously described.²⁰ The mice were euthanized if the cumulative clinical score was a six or greater. For assessing complete blood counts (CBC), mice were bled by nicking the tail vein and collecting blood in a microtainer K₂ EDTA tube (cat#365974 Becton Dickinson, Franklin Lakes, NJ). Mice were bled before radiation challenge to get baseline values (control) and then post-radiation on day 4. Surviving LPS-EEM treated mice were also collected on day 32 and days 50–53. Whole blood was assayed on a Hemavet 950FS analyzer (Drew Scientific Inc., Miami Lakes, FL,) and mean values were determined for erythrocytes, leukocytes and thrombocytes.

Diagnostic necropsy and histologic preparation

Gross necropsy of organ systems consisting determination of both organ weight and organ weight as function of percent body weight (%BW) as well as the external examination of the integument, cardiovascular, respiratory, digestive, lympho-hematopoietic, uro-genital, endocrine, central nervous system and musculoskeletal. Gross necropsy was performed on the following groups: Un-irradiated NSG mice, NSG mice post radiation challenge on day 9 (PBS treated and LPS-high EEM treated), day 31 (LPS-high EEM treated), and day 53 (moribund LPS-high EEM treated).

Histology focused on the preparations of slides from sections of spleen and bone marrow from long bones (femur, tibia, humerus), sternum and the pelvis (ilium). Tissues were fixed in 10% neutral buffered formalin and processed on a Sakura Tissue-Tek VIP 6 processor and embedded on a Sakura Tissue-Tek TEC embedding station. Slides were cut on a Leitz 2235 microtome at 5–6 microns and stained with H&E using a Sakura Tissue-Tek DRS automatic stainer and manually cover slipped. Tissues were visualized using a Nikon Eclipse 50 I microscope at multiple magnifications using Nikon objectives; 4x/0.2 - Plan Apo, 10x/0.45 Plan Apo, 20x/0.75 Plan Apo, 40x/0.95 Plan Apo. Photographs were taken using a SPOT model 10.2 camera aided with SPOT acquisition software for MAC 5.2. The bone marrow cellularity was scored by a blinded pathologist using a semi-quantitative scale: No loss = 0, Minimal loss = 1, Mild loss = 2, Moderate loss = 3, Marked loss = 4, Severe loss = 5.

Statistical analysis

Statistics were performed using GraphPad Prism version 7.0 (GraphPad Software, San Diego, CA). Data were reported as mean ± SEM. Groups of 3 or more were compared using an ordinary one-way ANOVA or Kruskal-Wallis test with Dunn’s multiple comparisons post-test. Mantel-Cox log-rank test was used for the comparison of the Kaplan-Meier survival curves. A p-value less than 0.05 was considered statistically significant for all tests.

Results

Extracellular vesicles (EVs) from MSCs consist primarily of exosomes

TEM images indicated MSC-EVs had the typical appearance of exosomes; circular shape with convex center and the majority (>90%) of EV preparation consisted of exosomes sized vesicles (<200 nm) (Figure 1A). Quantifying EVs of MSCs or macrophages using either a resistive pulse sensing instrument (Figure 1B) or a visual nanoparticle tracking analysis (Figure 1C) indicated that the EV’s generally ranged in size from 50–185 nm. Based on this analysis, we designated EV’s as exosomes. The qNano instrument analysis of multiple preparations from different MSC isolates yielded similar profiles in terms of mean particle sizes (mean particles size of 128 nm and mode particle size of 91 nm) and particle density (Table S1). Exosomes from LPS primed MSCs were similar to exosomes from un-primed MSC-exosomes in terms of size (mean particle size of 168 nm and mode particle size of 109 nm) with mean particle concentrations; 1.4 × 10¹¹ particles/ml vs 1.1 × 10¹¹particles/ml for the exosomes from unprimed and LPS primed MSCs respectively. LPS levels in EVs from both unstimulated and LPS stimulated MSCs were found to be 0.008ug/ml +/− 0.006 and 1.2ug/ml +/− 0.65, respectively. Macrophages were also found to primarily produce exosome sized particles, with yields about 5-fold more based on cell number (10⁶).

Figure 1: EVs isolated from human MSCs consist primarily of exosomes sized vesicles.

EVs were isolated from BM-MSC’s, and EVs were analyzed by TEM and two different instruments to quantify the mean particle diameter. (A) Representative TEM of the EV preparations indicated that the particles had the typical cup-shaped vesicular appearance of EVs and the size was generally less than 200 nm. (B) The preparations characterized by resistive pulse sensing using the qNano Nanoparticle instrument indicated the majority of the particles were within the 60–130 nm range. (C) The preparations characterized by dynamic light scattering using the Nanosight NS300 indicated that the size of the majority of particles were 95 nm with a range of 50 to 129 nm and generally matched the same profile as using the qNano Nanoparticle instrument. Based on these characterizations the majority of the particles consisted primarily of exosomes-sized EVs.

EEMs express a combination of classical M2 and M1 markers dependent on LPS-stimulation of MSCs

Next, we used exosomes from unprimed MSCs, MSCs primed with a low dose of LPS, and MSCs primed with a high dose of LPS, to generate exosome-educated macrophages (EEMs, LPS-low EEMs and LPS-high EEMs) and compared their phenotypes by examining cell surface marker expression of M2 markers (CD163, CD206, PD-L1 and PD-L2), M1 markers (CD86 and HLA-DR) and functional markers (CD16, CD39, CD73). Compared to controls, EEMs expressed significantly higher mean fluorescent intensities (MFIs) for CD206 (p ≤ 0.0001), PD-L1 (p ≤ 0.0001), and PD-L2 (p ≤ 0.001) (Figure 2). LPS-low EEMs also showed elevated expression of CD206, PD-L1, and PD-L2, but also higher CD16 (Figure 2). In contrast LPS-high EEMs showed no difference in M2 marker expression compared to controls. The M1 markers CD86 and HLA-DR were significantly lower in LPS-low EEMs compared to either controls or EEMs (Figure 2). While M2 marker expression levels between the LPS-low and LPS-high were different, expression of M1 markers in LPS-high EEMs were actually decreased compared to controls or EEMs. There was a significant increase in the percentage of CD73⁺ cells for both LPS-low and LPS-high EEMs compared to either controls or EEMs (Figure S1). Thus LPS-high EEMs do not show increased expression of M1 or M2 markers, but do have an increased percentage of CD73.

Figure 2: Human LPS-high EEMs express low levels of M1 markers CD16, CD86 and HLA-DR.

Day 7 macrophages were either untreated (Control) or treated with exosomes from MSCs to produce EEMs, LPS-low or high EEMs. The Median Fluorescence Intensity (MFI) of CD14+ positive cells for each marker (+/− SEM) is shown. Results pooled from 2 separate experiments, with 4–13 samples/group. Groups compared by Kruskal-Wallis with a Dunn’s post test. *p</=0.05, ** p</=0.005, *** p</=0.0005, **** p</=0.0001) between groups is designated by bars.

LPS-high EEMs protect mice from lethal radiation injury and promote hematologic recovery

To determine if exosomes from MSCs could generate macrophages that mediate radioprotection, EEMs, LPS-low EEMs and LPS-high EEMs were infused as a single treatment to NSG mice after a lethal dosage of radiation. Recipients of LPS-high EEMs showed improved survival (p ≤ 0.0001, Figure 3A), whereas cell treatments with MSCs, macrophages, or EEMs at the same dose were ineffective. Only LPS-high EEMs treatment led to a significant reversal of ARS and prolonged survival, with a median survival of 47.5 days, compared to 8 to 14 days compared with the other groups. While there was an initial delay in recovery of weight and clinical score in the LPS-high EEM treated mice, by day 10, the mean percent weight change (p < 0.05, Figure 3B) and mean clinical score (p < 0.05, Figure 3C) steadily began to improve and remained at near normal levels for several weeks. About 70% of the LPS-high EEM treated mice were alive at this point while 20% of the mice treated with LPS-low EEMs survived long-term, indicating that potency of the exosomes appears to may be related to concentration of LPS used to stimulate the MSCs. Overall, the protective effect from LPS-high EEMs treatment lasted until day 38 (Figure 3C), thereafter the cumulative clinical scores progressively worsened and the remaining mice died between days 50–53.

Figure 3: Treatment with human LPS-high EEMs significantly improves survival, weight loss and clinical scores in mice after lethal radiation injury.

(A-C) On day 0, NSG mice were received 4 Gy of lethal radiation followed by an i.v. treatment 4 hours later with either PBS (vehicle control), 1 × 10⁶ MSCs, 1 × 10⁶ macrophages, 1 × 10⁶ EEMs, 1 × 10⁶ LPS-low EEMs or 1 × 10⁶ LPS-high EEMs. (A) Survival curve of treated mice after radiation. (B) Mean % weight change compared to PBS controls. (C) Mean clinical scores (% weight loss, posture, activity and fur texture) compared to MSC and/or PBS controls. The final mean percent weight change and clinical score were carried over after death to allow for comparison by Kruskal-Wallis with a Dunn’s post test between groups at a given timepoint. Results pooled from 2 separate experiments, with 7–21 mice/group. * p< 0.05, *** p</=0.005, ****p</= 0.0001.

LPS-high EEM treatment was also found to significantly restore CBCs in mice after lethal ARS (Table 1). At day 4, all irradiated mice showed signs of radiation injury and developed general pancytopenia. Interestingly, the greatest significant drop in leukocyte cell subsets (neutrophils, lymphocytes and monocytes) was detected in LPS-high EEM treated mice. Moreover, there was a significant reduction in platelets (Table 1, day 4). Overall, we were unable to identify a candidate cell subtype that could be a predictable biomarker for survival in LPS-high EEMs treated mice. By day 20, the only remaining group available for further CBC analysis were recovering LPS-high EEM treated mice. During this recovery period, as assessed by weight and clinical scores, these mice were re-bled at day 32. Leukocyte subsets were all restored to normal levels and platelets improved, although not quite to normal levels (Table 1, day 32). However platelet volume was significantly greater, suggesting more immature platelets in the periphery. At days 50–53, when the mice became moribund, there was a slight drop in platelets, however the other blood cell subsets did not drop significantly (Table 1, days 50–53). Thus, a single treatment with LPS-high EEMs was able to support normal hematopoiesis, even during later clinical deterioration, indicating the terminal diagnosis during clinical symptom relapse was not due to a failure in hematopoiesis. Histopathological examination of major organs from day 50–53 moribund mice indicated that the probable terminal diagnoses was either hepatic hypoxic necrosis (with a predominantly centrilobular pattern), a skin infection or a combination of both (Fig. S3). No cultures or other testing were performed to identify the causative organism, but skin features were suspicious for Corynebacterium bovis.

Table 1:

Impact of human LPS-high EEM treatment on complete blood counts after lethal irradiation.

Group	Day post radiation	RBC (M/ul)	WBC (K/ul)	Neutrophils (K/ul)	Lymphocytes (K/ul)	Monocytes (K/ul)	Platelets (K/ul)	Platelet vol (fL)
Control	N/A	4.6	1.37	1.06	0.21	0.065	608	4.5
PBS	4	4.3	0.49*	0.14**	0.24	0.03	191**	4.3
EEM	4	3.7	0.19**	0.03**	0.08	0.013	219**	4.2*
LPS-high EEM	4	3.7*	0.21***	0.02***	0.05*	0.01**	187***	4.5
LPS-high EEM	32	4.1	1.68	1.21	0.36	0.05	379*	5.0***
LPS-high EEM	50–53	5.8	1.48	0.92	0.4	0.11	316**	5.0***

^*p</=0.05

^**p</=0.005

^***p</=0.0005

To determine the contribution of residual LPS in the exosome preparations isolated after MSC priming on direct macrophage stimulation, we compared treatment of LPS-stimulated macrophages to no treatment (PBS), unstimulated macrophages and LPS-high EEMs treatment of lethally irradiated mice. We found that LPS-stimulated macrophages resulted in 100% lethality akin to mice receiving no treatment or unstimulated macrophages. Only recipients of LPS-high EEMs showed significantly improved survival (Fig. S4).

LPS-high EEM treatment protected hematopoietic tissue in bone marrow and spleen from radiation injury

To identify which organs and tissues may be protected by LPS-high EEM treatment, we compared histology of BM from long bones (Figure 4A) and spleens (Figure 4B) of normal non-irradiated mice to irradiated mice with or without treatment at different times post-challenge. By gross necropsy, the spleens were most affected by 4Gy radiation exposure, while overt changes to the heart, liver and kidneys were less obvious (data not shown). Spleens from irradiated moribund PBS treated mice, were significantly smaller in mean weight and % spleen BW (9.3 mg and 0.05%) compared to healthy controls (28.1 mg and 0.11%) (p<0.05). In contrast to the irradiated PBS treated mice, mean spleen weight and % spleen BW from LPS-high EEM treated mice during both the recovery phase at day 31 (34.4 mg / 0.15%) or even during clinical symptom relapse at day 50–53 (25.1 mg/ 0.13%) were both similar to healthy controls. Compared to the histologic sections of BMs and spleens from healthy mice (Figures 4A and 4B), moribund PBS treated mice at day 9 post-irradiation showed a marked absence of hematopoietic cellularity in the BM and total lack of extra-medullary hematopoiesis in the spleen with clear hemorrhage (Figures 4A and 4B). In contrast, there were markedly more hematopoietic cells present in the LPS-EEM treated mice at day 9 post-irradiation (Figures 4A and 4B). At this time, cumulative cellularity present in the BM cavity (long bones, sternum and pelvis) graded from 0 to 5 (indicating most to least degree of cellularity) was significantly better in LPS-high EEM mice compared to untreated mice (p<0.01) (Table 2). Improvement continued in these mice at day 30 with strong to moderate hematopoietic activity in the BM of the long bones but also in the pelvis and sternum. The cumulative cellularity improved to near normal levels in LPS-EEM treated mice at scores of 0.5±0.5 (Table 2), with an intense hematopoietic component present in spleen (Figures 4A and 4B). Interestingly even during clinical symptom relapse at day 53, hematopoietic tissue in the BM and spleen was still distinctly present (Figures 4A and 4B) in the LPS-high EEM treated mice, similar to what we observed in CBCs. Cellularity in the long bones, sternum and pelvis in these mice was high at 1.4±0.8.

Figure 4: Human LPS-high EEM treatment protects against tissue damage in the BM and spleen of mice after lethal radiation injury.

On day 0, NSG mice were received either no radiation (normal healthy) or 4 Gy of lethal radiation followed by an i.v. treatment 4 hours later with PBS or with 10⁶ cells of LPS-high EEMs. Histology on tissue preparations of BM from femurs and spleens of healthy mice were compared to day 9 post radiation PBS controls, and LPS-high EEMs, day 31 LPS-high EEM treated mice or day 53 LPS-high EEMs. (A) Representative 20x images of H&E stained femoral BM sections from each group. (B) Representative 20× images of H&E stained spleen sections.

Table 2:

Bone marrow cellularity

Group	Day post radiation challenge	Long bones (femur, tibia, humerous)	Sternum	Pelvis	Cumulative
Healthy controls	N/A	0.0 +/− 0.0	0.0 +/− 0.0	0.0 +/− 0.0	0.0 +/− 0.0
Untreated controls	9	4.6 +/−0.4	4.3 +/−0.7	4.3 +/−0.4	4.3 +/−0.7
LPS-high EEMs	9–10	4.0 +/−1.0	3.3 +/−0.7	4.1 +/−0.2	3.8 +/−0.6**
LPS-high EEMs	31	1.0 +/−0.9	0.25 +/−0.7	0.3 +/−0.1	0.5 +/−0.7****
LPS-high EEMs	50–53	1.9 +/−0.4	0.9 +/−0.7	1.5 +/−1.3	1.4 +/−0.8****

Key: No loss = 0, Minimal loss = 1, Mild loss = 2, Moderate loss = 3, Marked loss = 4, Severe loss = 5

LPS-EEMs have a distinct gene expression and cytokine profile

In order to examine potential mechanisms of protection against ARS and promotion of hematopoietic recovery, we next examined the gene expression profile of the macrophage subsets by qPCR. Both LPS-low and LPS-high EEMs showed a similar profile with significant increases in VEGFA (LPS-low p = 0.018 and LPS-high p = 0.006) and STAT1 (LPS-low p = 0.022 and LPS-high p = 0.033) compared to control macrophages (Figure S2). However, there were differences between LPS-low and LPS-high EEMs; LPS-high EEM only showed increased expression of STAT3 (p = 0.048) as compared to control macrophages (Figure S2).

To verify changes in gene expression at the protein level, we also quantified production of cytokines and growth factors by each macrophage subset. LPS-low EEMs and LPS-high EEMs secreted significantly higher level(s) of factors involved in tissue repair compared to control macrophages, including cytokines, [IL-4 (p < 0.001), IL-8 (p < 0.01)], growth factors [EGF (p < 0.001), FGF-2 (p < 0.001), VEGF-A (p < 0.001)], T and NK cell growth factors [IL-7 (p < 0.01), IL-15 (p < 0.0001)], platelet growth and activation factors, [sCD40L (p < 0.001), PDGF-BB (p < 0.05)] as well as hematopoietic growth factors [FLT-3L (p < 0.05), G-CSF (p < 0.05) and GM-CSF (p < 0.01)] (Figure 5). Moreover, many of these secreted factors were also significantly higher than that found for EEMs. While the types of factors secreted by either the LPS-low or LPS-high EEMs were very similar, the concentrations secreted by LPS-high EEMs were greater than LPS-low EEMs. Interestingly IL-10 (p = 0.05), FLT-3L (p < 0.05) and IL-15 (p < 0.05) were found to be significantly higher in LPS-high EEMs (Figure 5).

Figure 5: Human LPS-high EEMs secrete high levels of anti-inflammatory cytokines and growth factors by multiplex ELISA.

Day 7 macrophages were either untreated (control) or treated with exosomes from MSCs for 3 days to produce EEMs and LPS-low or high EEMs. Cells were then washed and supernatants collected after 24 hours and assayed. Samples were run in triplicate and compared by ANOVA with a Dunn’s post test. *p</=0.05, ** p</=0.01, *** p</=0.001, ****p</= 0.0001) were determined by compared to controls, EEMs or LPS-low-EEMs designated as *, #, or $, respectively.

LPS-high EEMs show increased phagocytic activity

Active phagocytic function in macrophages is important for both wound healing and tissue re-modeling. ²¹ Phagocytic activity, as measured by the percentage of cells containing internalized pHrodo Green-E.coli particles (% cells), was found to be highest in LPS-high EEMs (Figure 6). Phagocytic activity between the controls and EEMs was not significantly different. However, LPS-low EEMs (p≤0.0001), and LPS-high EEMs (p≤0.0001) were significantly more phagocytic than controls. In contrast, M1 macrophages were significantly less phagocytic (42%) compared to controls (59%, p≤0.0005). Thus, the macrophage subsets that are effective in this ARS model seem to have increased phagocytic activity, perhaps indicating a functional biomarker and role for tissue remodeling.

Figure 6: LPS-high EEMs are strongly phagocytic using pHrodoGreen E.coli bio particles.

Day 7 macrophages were either untreated (control) or treated for 3 days using exosomes from MSCs to produce EEMs and LPS-low or high EEMs or stimulated with M1 factors; (PMA/IFN-gamma/LPS) to produce M1 stimulated macrophages. Day 10 macrophages were treated with pHrodoGreen E.coli bioparticles and the ratio of CD14+ cells positive for phrodoGreen E.coli bioparticles (designated as percent (%) cells) was determined by flow cytometry. Samples were pooled from 2 separate experiments and compared by ANOVA with a Dunn’s post test, 3–5 samples/group. *p</=0.05, ** p</=0.005, *** p</=0.0005, **** p</=0.0001) between groups is designated by bars.

Discussion

The goal of this study was to develop a cell-based therapy that would be effective at preventing or treating ARS. Macrophages are an important cell subset involved in wound healing and tissue remodeling. At the site of injured tissue, macrophages can clear the site of pathogens or cell debris, regulate inflammation, and promote tissue repair.^22,23 Previously we have shown that macrophages educated after co-culture with MSCs were more effective than MSCs alone in treating ARS in vivo in part by promoting fibroblast proliferation.¹⁵ However a co-culture methodology can be challenging to translate to the clinic, since one would have to eliminate MSCs from the macrophage culture prior to infusion, or infuse two different type of cells simultaneously. Herein we show that exosomes isolated from MSCs can replace co-cultivation, and when derived from MSCs primed with LPS, can educate macrophages to mediate protection from ARS in vivo, in part by preserving host hematopoiesis.

MSC-EVs are produced in large numbers and consist of both large (1000 nM) micro-vesicles and smaller (100 nM) exosomes known to contain micro-RNA, proteins and DNA that can influence the gene expression and functionality of recipient cells. ^24–27 MSC-EVs can influence tissue responses to injury, infection and disease, and their function appears to reciprocate the direct use of MSC’s. ^24,28 Indeed, there is a great deal of interest in the therapeutic potential of EVs alone, for the treatment of various diseases and tissue repair. ²⁹ Numerous animal studies have indicated that MSCs-EV treatments are effective as a sole therapy in brain, heart, liver and kidney injury models.²⁴ For ARS treatment, MSC-EVs have been shown to rescue murine marrow hematopoietic cells after sub-lethal radiation. ³⁰ Moreover, the contents of MSC-EVs are not static, and can be altered by MSC exposure to different stimuli. A recent study indicated that MSCs primed with bacterial LPS, a TLR-4 ligand, produced MSC-EVs that were better at promoting wound healing in diabetic rats and thought to be due to the polarization of endogenous macrophages in the tissue.³¹ Our analysis of multiple human isolates indicated that MSCs reproducibly produced exosomes of similar size and quantity. Unlike a previous report³¹ we did not detect any significant increase in exosome concentration after priming of MSCs at either low or high dosages of LPS (Table S1). Overall this consistency in exosome secretion provides an attractive translatable option for commercial production by MSCs.

Current dogma indicates that stimulation of MSCs with pro-inflammatory signals like LPS should promote a pro-inflammatory phenotype. Indeed, direct injections of LPS are being studied in clinical trials as an adjuvant to stimulate inflammation and boost innate immunity. ³² Therefore, one might expect that macrophages treated with exosomes from MSCs stimulated with LPS would generate a M1 macrophage that would not only be ineffective, but more destructive in an ARS model. However, when macrophages were educated with exosomes from LPS primed MSCs, we observed a reparative phenotype that extended survival, repaired tissue damage in the spleen and bone marrow, and promoted hematopoiesis. This outcome was dependent on LPS and its concentration, as exosomes from MSC primed with a high concentration of LPS (LPS-high EEMs) improved weight loss, showed lower clinical scores and extended overall survival in an ARS model, with quantifiable improvements by CBC and repairing of hematopoietic tissue in the BM and spleen. Notably there was initially a decrease in “leaky-SCID” lymphocytes observed in vivo which may have decreased the overall inflammatory response to radiation, leading to a significant restoration in the leukocyte cell panel, along with significant increases in platelet volume; all indications of the resumption of hematopoietic activity. Studies have indicated the importance of macrophages in supporting and maintaining hematopoietic stem cells. ^33–35 We hypothesize that since the MSCs are strong homeostatic regulators, exposure to strong inflammatory environmental ques, such as high LPS levels, compel them to direct effector cells like macrophages to respond in a counter-balancing way and become more anti-inflammatory.

To examine this hypothesis, we characterized the macrophage subsets by cell surface marker expression, gene expression, cytokine/growth factor secretion and phagocytic activity. Distinctly in LPS-high EEMs, increased gene expression of STAT3 was noted compared to control macrophages. Increased protein expression of IL-10, an important anti-inflammatory cytokine, VEGF-A, a growth factor involved in angiogenesis, and IL-8, a cytokine involved in stimulating phagocytosis, were also observed.³⁶ By flow-cytometry, the LPS-high EEMs displayed a unique surface profile; expressing very low levels of M1 markers (CD86 and HLA-DRs) and the pro-inflammatory marker, CD16.³⁷ Another relevant surface marker was CD73, expressed in significantly higher percentages of LPS-high EEMs. CD73 is an ecto-nucleotidase which converts AMP to adenosine and thought to be both immunosuppressive and angiogenic^15,38 each important in wound repair. Since MEMs were previously described to be effective in treating ARS in vivo, ¹⁵ the shared phenotype of LPS-high EEMs and MEMs (low expression of M1 markers with high expression of CD73) may be an important “release criteria” that could be used to designate successful education by MSCs. Nevertheless, it should be noted that while characteristics found in LPS-high EEMs were significantly different from controls, often there were no differences between LPS-low and LPS-high EEMs, indicating some biologic features are a result of education with LPS-stimulated MSCs independent of LPS dose.

LPS-high EEMs were found to secrete significantly greater amounts of cytokines and growth factors compared to the other treatment groups. These factors likely helped re-established normal blood counts in the mice and possibly involved in the restoration of clinical scores in the mice with ARS. Factors important for tissue re-modeling such as EGF, FGF-2, PDGF-BB, sCD40L and VEGF-A were elevated in LPS-high EEMs. Moreover there were increases in growth factors such as IL-7 and G-CSF, which function to restore T cells and neutrophils respectively. As seen by gene expression and flow cytometry, the overall secretory profile of LPS-high EEMs were very similar to LPS-low EEMs; two growth activating factors: FLT-3L, a stem cell growth factor which stimulates the growth of blood progenitors, and IL-15, which regulates the activation and proliferation of T-cells and NK cells, were significantly higher in LPS-high EEMs. These two factors may be the major drivers of hematopoietic recovery and could be blocked in preclinical models to see if the radio-protective effects are abrogated.

Recent studies suggest a strong link between phagocytic macrophages and tissue repair; deficits in macrophage phagocytic function can lead to the pathogenesis of non-healing wounds often seen in diabetes or during aging. ²¹ The LPS-high EEMs were found to possess the highest percentage of phagocytic cells compared to all the other groups. An important function of wound macrophages is the ability to remove neutrophils after the decontamination phase of wound healing and evidence indicates that neutrophils negatively influence repair in part because they are capable of destroying normal tissue. ²¹ Besides actively ingesting neutrophils, phagocytic macrophages can also induce apoptosis, which prevents secondary necrosis, and thought to be essential for complete repair. ²¹ Besides the LPS-high EEMs, the LPS-low EEMs and MEMs shown previously also had significantly higher phagocytic ability and interestingly were the only groups to show any efficacy in the ARS model. ¹⁵ These results indicate the importance of phagocytic capacity for the treatment of radiation damage from ARS.

In summary, we show that exosomes from MSCs stimulated by LPS can educate macrophages into a radio-protective phenotype capable of affecting cytokine secretion and increasing phagocytic activity in vitro and promoting hematopoiesis and tissue repair in vivo. These effects were associated with improved survival from ARS in vivo. To move these findings toward an “off the shelf” cell therapy for ARS, we envision the use of frozen allogeneic macrophages that were pre-educated with exosomes from LPS-stimulated MSCs. Because residual LPS is present in exosome preparations isolated after MSC priming, some of the changes in macrophage phenotype by flow cytometry, gene expression and cytokine/growth factor secretion may be confounded by LPS alone, or in conjunction with the contents of the exosomes. This caveat should be considered when characterizing any cells generated from LPS-primed MSC-EVs. Future studies should focus on establishing biomarkers that predict successful generation of radio-protective cells ex vivo and cell subsets in vivo that contribute to improved hematopoiesis, and in better understanding the mechanisms used by LPS-high EEMs to protect against lethal ARS.

Highlights.

• Activation of human MSCs, in a LPS-dose dependent manner, generates exosomes that can polarize human macrophages into radioprotective cells, called LPS-low or LPS-high EEMs, that improve survival from lethal acute radiation syndrome in vivo.

• Human LPS-high EEMs demonstrate increased gene expression of STAT3 and protein expression of IL-10, IL-15 and FLT3L as well as high phagocytic activity.

• Mice with lethal acute radiation syndrome who receive a single infusion of human LPS-high EEMs show improved hematopoiesis by complete blood cell counts and histopathologic examination of the bone marrow and spleen.