|Home | About | Journals | Submit | Contact Us | Français|
When human umbilical cord blood (HUCB) cells are systemically administered following middle cerebral artery occlusion (MCAO) in rats, they produce a reduction in infarct size resulting in recovery of motor function. Rats receiving HUCB cells have a less severe inflammatory response compared to MCAO stroke rats. The purpose of this study was to determine the interaction between HUCB cells and the main resident immune cells of the brain (microglia) under normoxic and hypoxic conditions in vitro. Primary microglial cultures were incubated for 2 h in no oxygen (95% N, 5% CO2) and low glucose (1%) media. Mononuclear HUCB cells were added to half the cultures at the beginning of the hypoxia conditions. Microglial viability was determined using fluorescein diacetate/propidium iodide (FDA/PI) labeling and cytokine expression using ELISA. In some studies, CD11b+ or CD19+ cells isolated from the HUCB mononuclear fraction with magnetic antibody cell sorting (MACS) were used instead of the mononuclear fraction. Co-culturing mononuclear HUCB cells with microglia decreased viability of the microglia during hypoxia. In the microglial monocultures, hypoxia significantly increased release of IL-1β compared to normoxia, while adding HUCB cells in the hypoxia condition decreased IL-1β concentrations to the same level as in the normoxia monocultures. Both CD11b+ and CD19+ HUCB cells decreased microglial viability during normoxia and hypoxia. Our data suggest that HUCB cells may produce a soluble factor that decreases viability of microglia.
The intact brain is an immune privileged island secured by the blood–brain barrier (BBB). Within the central nervous system (CNS), there are no antigen-presenting cells, and there are few cells expressing major histocompatibility complex (MHC) class I and II molecules . However, this immune privilege is disrupted by stroke when microglia become activated, and express MHC antigens and co-stimulatory molecules . The activated microglia acquire a phagocytic phenotype in order to clear necrotic tissue and debris . These activated microglia also secrete cytotoxic cytokines, free radicals, and nitric oxide, resulting in disruption of the BBB . In addition, microglia also secrete chemokines to recruit leukocytes from the circulatory system, where they participate in the neuroinflammatory response.
Among these secreted cytokines, the proinflammatory cytokines, interferon-gamma (IFN-γ), interleukin-1 beta (IL-1β), and the anti-inflammatory cytokine, IL-10, have all been detected in increased levels in the stroked brains of rats following middle cerebral artery occlusion (MCAO) . Rats treated with microinjections of lipopolysaccharide (LPS) and IFN-γ into the hippocampus exhibited delayed neuronal apoptosis, which peaked at 7 days following administration . When IFN-γ was knocked out in Down syndrome mice, neurons showed improved growth and viability in vitro . In addition, IFN-γ also promotes microglial death under stressful conditions . Similar to IFN-γ, IL-1β is expressed shortly after focal cerebral ischemia. It increases infarct volume following ischemia in rodents and inhibiting it prevents neuronal cell death caused by injury [9,10]. IL-1β also mediates microglial proliferation . The anti-inflammatory cytokine IL-10 can suppress the production of a variety of proinflammatory cytokines including TNF-α, IL-1β, and IL-8 [12–14]. Changes in the balance between pro- and anti-inflammatory cytokines may affect microglial proliferation, protein secretion, and phagocytic activity.
There is an increasing literature on the ability of marrow- and blood-derived cells to reduce lesion size and improve outcome [15–17]. We have shown that systemic administration of human umbilical cord blood (HUCB) produced better behavioral recovery than observed in control animals following MCAO in adult rats [16,17] and hypoxia/ischemia in neonatal rats [18,19]. Since HUCB is rich in progenitor cells , it was thought that this functional recovery could be attributed to differentiation of the transplanted cells and replacement of lost cells. However, immunohistological examination found that few infused HUCB cells survived [16,17], and among those that did only a few expressed neural markers . Therefore, it is unlikely that these cells were able to cross the BBB, proliferate, and differentiate enough to replace the cell loss caused by stroke. Interestingly, HUCB-infused animals exhibited a decreased infiltration of CD45/CD11b+ and CD45/B220+ cells into the brain . The changes were accompanied by decreased expression of various proinflammatory cytokines, decreased nuclear transcription factor-κB (NF-κB) DNA-binding activity, and increased expression of anti-inflammatory cytokines . Therefore, the functional recovery of the HUCB-treated rats could be the result of suppression of the inflammatory response.
We propose that HUCB cells induce neuroinflammatory suppression by modulating microglial survival and cytokine expression. In this study, primary microglia isolated from embryonic Day 17 rat pups underwent simulated hypoxic/ischemic conditions in vitro both with and without co-culture with HUCB cells and microglial survival was measured. In order to identify the important component of the HUCB cell preparation, HUCB subpopulations were also used.
Microglia were isolated from fetal rat (Sprague Dawley) brains at embryonic Day 17 according to the procedure described by Mackay . This work was carried out under the purview of the University of South Florida Institutional Animal Care and Use Committee. The brains were minced in Earle’s Balanced Salt Solution (EBSS; Invitrogen, Carlsbad, CA) containing 20 μg/mL DNase (Sigma Aldrich, St. Louis, MO) and 0.3% Bovine Serum Albumin (Invitrogen, Carlsbad, CA) (solution A) on ice. Following this, the homogenate was centrifuged at 500g for 2 min, and the pellet was collected and incubated in EBSS containing 0.025% trypsin (Sigma Aldrich), 60 μg/mL DNase, and 0.3% Bovine Serum Albumin (solution B) at 37°C for 15 min to dissolve the connective tissue between the cells. Trypsinization was stopped with 10% fetal calf serum (FCS; Invitrogen), and the solution was centrifuged again at 500g for 2 min. The pellet was resuspended in solution A and triturated through a fire-polished Pasteur pipette. The solution was allowed for 4 min to settle and the supernatant was collected. The pellet was then resuspended in solution A, and the process was repeated twice. The collected supernatant was combined together and centrifuged at 500g for 5 min. The pellet containing the cells was resuspended in culture medium (Dulbecco’s modified Eagle’s Medium (DMEM; Invitrogen) supplemented with 10% (v/v) FCS) and plated at a density of 2.5 × 105 cells/cm2 in a poly-l-lysine (Sigma Aldrich)-coated flask. Isolated cells were cultured at 37°C with 5% CO2 without medium change for the first 5 days. Media was then changed, and cells were maintained in culture until confluence. The confluent culture was placed on a rotary shaker and agitated for 1 h at 37°C. The media was then collected and spun down. The cells were resuspended in serum-free DMEM and plated at a concentration of 100,000 cells/mL into culture flasks and maintained at 37°C with 5% CO2 for 1 h. Loosely adherent oligodendrocytes were floated from the wall of the culture dishes using gentle agitation by hand. The medium was removed gently and serum-free DMEM was added to the dishes. This procedure was repeated twice. The remaining adherent cells were purified microglia. These cells were cultured in DMEM with 10% FBS and 0.1% gentamicin until use.
Microglia cells were grouped into hypoxia monoculture, hypoxia co-culture, normoxia monoculture, and normoxia co-culture conditions. For hypoxia conditions, the gas-tight chamber  and the low glucose color-free DMEM medium were flushed with 5% CO2 and 95% N2 for 10 min. All plates were then washed and filled with 1 mL of the pretreated medium. For co-culture conditions, an insert was set into the microglial plate, and freshly thawed HUCB cells were pipetted directly into the inserts at a concentration of 500,000 cells/well. The 0.4 μm Millipore (Millipore, Inc., Billerico) membrane allowed medium to pass freely, while the cells were held in the insert. Hypoxia groups were exposed to 5% CO2 and 95% N2 at 37°C for 2 h, while the normoxia groups were treated to normoxic conditions at 37°C for 2 h. Medium was harvested for further use. As a control, the same amount of freshly thawed HUCB cells were pipetted into the inserts with the same amount of clear DMEM and cultured under hypoxic or normoxic conditions for 2 h without microglia. The medium from this control was also collected.
We also examined the effect of CD11b+ (monocyte/macrophage), CD4+ (T cells), CD8+ (T cells), and CD19+ (B cells) subpopulations isolated from the HUCB mononuclear fraction on cell survival and cytokine production in microglial cultures. Cryopreserved HUCB cells (Saneron CCEL Therapeutics, Inc, Tampa, FL) were thawed into 1× BD Image buffer (BD Bioscience, San Jose, CA), and the cell concentration was adjusted to 107 cells per 90 μL. The HUCB cells were incubated with 10 μL anti-CD8-linked microbeads (BD Bioscience), anti-CD4-linked microbeads (BD Bioscience), anti-CD11b-linked microbeads (Miltenyi, Biotech Auburn, CA), or anti-CD19 microbeads (BD Bioscience) at room temperature for 15 min. The solution was centrifuged at 1,000g for 10 min, and washed with 1× BD Image buffer, and supernatant removed. The pellet was resuspended, and the cell concentration adjusted to 2 × 107 cells/mL in cold 1× BD Image buffer. The labeled cell suspension was transferred to the positive fraction collection tubes (12 × 75 mm tubes), and placed into the magnetic holder for 6 min. The positively labeled cells were attached to the wall of the tube, and negative cells sank to the bottom of tube or remained suspended in the medium. The supernatant was removed, and the cells on the wall were collected and resuspended. This procedure was repeated twice with 2-min incubations. The collected cells were placed into the insert at 100,000 cells/well.
Cell viability was determined by Fluorescein Diacetate (Invitrogen, Carlsbad, CA)/Propidium Iodide (Sigma Aldrich, St. Louis, MO) (FDA/PI) staining. In brief, FDA stock solution (5 mg/mL in acetone) and PI stock solution (1 mg in 50 mL PBS) were prepared and stored at 4°C. FDA working solution (5 μL FDA stock solution in 1 mL PBS) was prepared freshly before use, and 100 μL was added to 30 μL PI stock solution as FDA/PI working solution. Then 30 μL FDA/PI was added into each well containing 600 μL PBS, and cells were incubated at room temperature for 5 min. For HUCB cells on the membranes of insert, the insert was transferred into another well containing 600 μL PBS, and 30 μL FDA/PI was added into each well. Photomicrographs were taken under epifluorescence from five fields of each sample: center, left, right, upper, and lower. The number of live (FDA-labeled) or dead (PI-labeled) cells were counted using Image Pro Plus software (Media Cybernetics, Bethesda, MI). The experiment was repeated three times. The viability was determined as mean ± SEM percentage of live cells.
The cytokine ELISA was specific for rat cytokines and was performed according to the manufacturer’s protocol (Amersham Biosciences, Uppsala, Sweden). In brief, the standard dilution buffer was added, and the wells reserved for chromogen blank were left empty. One hundred microliters of the standard solutions were added to the appropriate wells, and 50 μL samples of medium with 50 μL of standard dilution buffer were added into the plates. Then, 50 μL of biotinylated anti-IFN-γ solution (or any other cytokine to be tested) was added into each well. The plates were covered and incubated for 30 min at room temperature. They were then washed four times, and 100 μL streptavidin-HRP working solution was added to each of the appropriate wells. After 45 min of incubation, the plates were washed four times, and 100 μL stabilized chromogen from the kit was added in each well for an incubation of 30 min. The reaction was stopped with 100 μL of the provided stop solution. The concentration of cytokines was determined on a plate reader with light absorbance at 450 nm. For each condition, there were six samples, and each sample was examined in triplicate.
For those data that are normally distributed, the data are presented as means ± SEM analyzed with analysis of variance and post-hoc Tukey tests. For nonparametric data, the data are presented as median ± interquartile range and were analyzed using the Kruskal–Wallis test followed by post-hoc analysis with the Mann–Whitney test.
Cell viability of microglial cells was determined using FDA/PI staining. Both monocultured (microglia alone) and co-cultured (microglia and HUCB cells) cells were observed under normoxic and hypoxic conditions. Viable cells stained bright green (FDA-labeled) while dead cells stained bright red (PI-labeled). This can be clearly observed in Figure 1A–1D. When we quantified the number of living and dead cells in each culture condition, we found that the data were not normally distributed; therefore, nonparametric tests were used to examine group differences. We found few dead cells in the microglial monocultures grown under normoxic conditions (91.1% ± 20.3% survival, Fig. 1E). There was no significant difference in cell viability between the normoxic monocultured microglia cells and the normoxic HUCB and microglia co-cultures (81.82% ± 18.3%). There was a significant increase in viability in the microglial monocultures exposed to hypoxia (96.2% ± 2.2%) compared to the normoxic monocultures (P < 0.05). Cell viability of microglial cells decreased significantly in the microglia HUCB co-cultures exposed to hypoxia (15.6% ± 23.6%) compared to the microglia HUCB co-cultures maintained in normoxia (81.8% ± 18.3%, P < 0.001). Additionally, we observed a significant decrease in microglial cell viability in the hypoxia-exposed microglia HUCB co-cultures (15.6% ± 23.6%) compared to the hypoxia-exposed microglia cultured alone (96.2% ± 2.2 %, P < 0.001).
The expression of the cytokines IL-1β, IFN-γ, IL-6, and IL-10 were measured by ELISA from the media of monocultured microglia, and microglia co-cultured with HUCB cells exposed to normoxic or hypoxic conditions (Fig. 2). The data were not normally distributed and were analyzed with the Kruskal–Wallis test followed by the Mann–Whitney test. IL-1β concentrations in media from microglia monocultures during normoxic conditions were 445.9 ± 120.5 pg/mL (median ± IQR) (Fig. 2A). When HUCB cells were added to the culture, concentrations significantly decreased to 236.0 ± 68.8 pg/mL (P < 0.002). Exposure of microglial cells alone to hypoxia significantly increased IL-1β concentrations to 718 ± 167 pg/mL compared to normoxia microglia monocultures (445.9 ± 120.5 pg/mL, P < 0.02). When co-cultures of microglia and HUCB cells were exposed to hypoxia, there was significantly more IL-1β in the culture media (390.5 ± 479.4 pg/mL) compared to the normoxia co-cultured controls (236.0 ± 68.8 pg/mL, P < 0.002).
There was no statistically significant change in the concentration of IFN-γ (Fig. 2B, P = 0.1055), or IL-10 (Fig. 2D, P = 0.0609), even though there appeared to be less of these cytokines in the media of the co-cultures exposed to hypoxia. Similarly, IL-6 concentrations remained unchanged by the experimental manipulations (Fig. 2C, P = 0.5181).
When HUCB cells alone were exposed to both normoxia and hypoxia, none of the cytokines examined were detectable in the media (data not shown).
When microglia were co-cultured with CD11b+ HUCB cells during normoxia, there was a significant decrease in microglial cell viability from 75.8% ± 5.9% in the monocultures to 38.8% ± 7.1% (P < 0.01) as assessed using FDA/PI staining. A similar result occurred during hypoxia exposure with microglial cell viability decreasing from 70.4% ± 6.6% (monoculture) to 41.6% ± 6.5% (co-culture, P < 0.01) (Fig. 3A). However, there was no significant difference in viability between the normoxic and hypoxic conditions.
Cell viability of microglia co-cultured with CD19+ cells exposed to hypoxia was assessed using FDA/PI staining. The data were not normally distributed, so the analysis was performed with nonparametric statistics. When microglia were co-cultured with CD19+ HUCB cells and exposed to normoxia, there was a significant decrease in viability from 91.7% ± 10.29% (monoculture) to 2.6% ± 83.3% (P < 0.002) (Fig. 3B). Microglia cell viability decreased from 95.2% ± 6.2% in the microglial monoculture exposed to hypoxia compared to the 82.5% ± 56.8% viability in the CD19+ HUCB-microglial co-culture during hypoxia (P < 0.002).
CD8+ and CD4+ T cells isolated from HUCB increased cell viability of microglia under normoxic conditions. There were significantly more microglia in the CD8+ HUCB-microglial co-cultures during normoxia (97.9% ± 4.9%) compared to the microglial monoculture (88.3% ± 10.0%, P < 0.006) (Fig. 4A). During hypoxia, there were no significant differences between the microglial monoculture (90.9% ± 7.3%) and the co-culture (94.0% ± 23%). The data for the CD4+ HUCB T-cell studies were normally distributed and analyzed with ANOVA (Fig. 4B). Under both normoxic and hypoxic conditions, there were significantly more microglia in the co-cultures than in the monocultures (89.8% ± 2.2% compared to 95.7% ± 0.9% during normoxia (P < 0.05) and 87.9% ± 1.4 % compared to 94.9% ± 1.1% during hypoxia (P < 0.01)).
In this study, we found that culturing HUCB mononuclear cells with microglia decreased cell viability of the microglial population. Further, this decrease in viability was accompanied by a decrease in the expression of the proinflammatory cytokine, IL-1β, in the media of hypoxia-exposed microglia. We also show that the CD11b+ and CD19+ HUCB cells decreased microglia viability following hypoxia while CD4+ and CD8+ HUCB cells actually increased microglial viability. These results are essential to understanding how HUCB cells suppress the inflammation observed in our previous studies . We also found that neither microglia nor hypoxia had a significant effect on HUCB viability (data not shown).
There is precedence for microglial death in other models of brain injury; LPS injection was found to result in microglial death within 6 h of injection . The decrease in microglia cell survival with exposure to HUCB cells during hypoxia/ischemia was not expected given the in vivo evidence that HUCB cells are neuroprotective. In the MCAO model of stroke in the rat, we have repeatedly shown that animals injected with HUCB cells 24–48 h following stroke have decreases in infarct volume and display better cognitive and motor function than untreated animals [16,17,26]. This neuroprotective effect in vivo has recently been independently verified in vitro . Forty-eight hours after exposing neuronal cultures to hypoxia-ischemia, the addition of HUCB cells decreased apoptosis and increased neuronal survival. This protective effect is not limited to neurons; HUCB cells also increase survival of cultured oligodendrocytes and protected white matter after MCAO as indicated by immunolabeling for myelin basic protein .
Expression of the proinflammatory cytokine, IL-1β, significantly increased in monocultured microglia after hypoxia exposure. IL-1β is a proinflammatory cytokine belonging to the Th1 immune response, and is cytotoxic to neural cells. IL-1β combines with its receptor to play a pivotal role in regulation of the inflammatory response in the CNS. It can recruit leukocytes into the CNS from peripheral blood vessels , modulate glutamate metabolism or uptake to enhance excitotoxicity , and mediate the release of free radicals, proteases, and inflammatory cytokines from glial cells. The rapid increase in IL-1β expression in response to ischemia/reperfusion  might suggest that IL-1β is the cause of cell death initiated after brain damage rather than a consequence. Further, intracerebroventricular administration of IL-1β increases infarct size caused by ischemia ; in contrast, administering IL-1β receptor antagonist (IL-1Ra) decreases infarct size [9,32]. Further in mice in which IL-1β-converting enzyme has been knocked out, thereby decreasing IL-1β production, neuronal damage after focal cerebral ischemia is reduced [33,34]. Interestingly, this IL-1β toxic effect only affects neurons in the CNS, and is not associated with microglial cell death. For instance, in IL-1β knockout mice, microglia had a similar cell death profile as wild-type microglia in response to LPS and ATP . In addition, IL-1β even acts as a pluritoptent growth factor for glia . When infused into the brain, IL-1β stimulates astrogliosis around the site of injection , and promotes oligodendrocyte proliferation and remyelination after CNS injury . In vitro, IL-1β can promote microglial proliferation . Our study data is consistent with IL-1β as a trophic factor for microglia  since IL-1β concentration was greater in the conditions with increased cell survival.
When we examined the effect of individual cell populations within the mononuclear fraction of HUCB on microglial cell survival, we found that both the CD11b+ monocytes and macrophage and CD19+ B cells were likely responsible for the decrease in microglial cell survival; these two populations decreased microglial survival under both normoxic and hypoxic conditions. The decrease in microglial viability during the normoxia condition is in sharp contrast to the effect of the complete mononuclear fraction of HUCB cells on microglia during normoxia. This would indicate that there are other cell types within the mononuclear fraction counteracting the effects of the CD11b+ and CD19+ cells. Indeed, both CD4+ and CD8+ T cells increased microglial survival during normoxia; only CD4+ T cells increased viability during hypoxia. Even with this “survival signal,” however, there is some process or signal induced during hypoxia/ischemia that enables the CD11b+ and CD19+ cells to override the protective signal generated by the CD4+ cells. Further studies are needed to determine the nature of this interaction and the signaling cascades involved.
Another interesting observation was made in the CD11b+ and CD19+ HUCB-microglial co-cultures; during hypoxia, the HUCB cell types decreased survival by ~30% and 20%, respectively, compared to the 75% observed with the whole mononuclear fraction. This could suggest that the two populations produce their effects through separate signaling pathways that when stimulated together have synergistic effects.
HUCB cells may promote neuronal survival in vivo by decreasing the proinflammatory response observed after a stroke. This may occur through a humoral signal that induces microglial cell death, thereby eliminating the source of proinflammatory cytokines. In the presence of the entire mononuclear fraction, this cell death must be triggered by the hypoxic environment but the monocyte/macrophage and B-cell populations both induce this effect under both normoxic and hypoxic conditions. These cell death effects are consistent with the decrease in microglial survival observed in the infarcted hemisphere after systemic HUCB administration  as well as the decreased proinflammatory and increased anti-inflammatory cytokine expressions in vivo . It remains, however, to determine the exact mechanism by which HUCB cells induce microglial cell death.
This study was funded in part by the American Heart Association (A.E.W., grants 0355183B & 0555266B) and the National Institutes of Health (A.E.W., R01 NS052839). Human umbilical cord blood cells were provided by Saneron CCEL Therapeutics, Inc.
Lixian Jiang, Center of Excellence for Aging and Brain Repair, University of South Florida, Tampa, Florida. Department of Neurosurgery and Brain Repair, University of South Florida, Tampa, Florida. Department of Pathology and Cell Biology, University of South Florida, Tampa, Florida.
Tracy Womble, Center of Excellence for Aging and Brain Repair, University of South Florida, Tampa, Florida. Department of Neurosurgery and Brain Repair, University of South Florida, Tampa, Florida.
Samuel Saporta, Center of Excellence for Aging and Brain Repair, University of South Florida, Tampa, Florida. Department of Neurosurgery and Brain Repair, University of South Florida, Tampa, Florida. Department of Pathology and Cell Biology, University of South Florida, Tampa, Florida.
Ning Chen, Center of Excellence for Aging and Brain Repair, University of South Florida, Tampa, Florida. Department of Neurosurgery and Brain Repair, University of South Florida, Tampa, Florida.
Cyndy Davis Sanberg, Saneron CCEL Therapeutics, Inc., Tampa, Florida.
Paul R. Sanberg, Center of Excellence for Aging and Brain Repair, University of South Florida, Tampa, Florida. Department of Neurosurgery and Brain Repair, University of South Florida, Tampa, Florida. Department of Pathology and Cell Biology, University of South Florida, Tampa, Florida. Saneron CCEL Therapeutics, Inc., Tampa, Florida. Office of Research and Development, University of South Florida, Tampa, Florida.
Alison E. Willing, Center of Excellence for Aging and Brain Repair, University of South Florida, Tampa, Florida. Department of Neurosurgery and Brain Repair, University of South Florida, Tampa, Florida. Department of Pathology and Cell Biology, University of South Florida, Tampa, Florida. Department of Molecular Pharmacology and Physiology, University of South Florida, Tampa, Florida.
S.S. and A.E.W. are consultants to Saneron CCEL Therapeutics, Inc. P.R.S. is cofounder of Saneron CCEL Therapeutics, Inc. A.E.W. and P.R.S. are inventors on cord blood-related patents. C.D.S. is Vice President for Research at Saneron CCEL Therapeutics; she optimized the cord blood preparations prior to releasing the samples to A.E.W.