Adipose-derived mesenchymal stromal cells (AdMSCs) have potent immunomodulatory properties useful for treating diseases associated with inflammation. We demonstrate how to isolate and culture murine AdMSCs and primary mixed glia, stimulate AdMSCs to upregulate anti-inflammatory genes and growth factors, assess migration of AdMSCs, and co-culture AdMSCs with primary mixed prion-infected glia.
Mesenchymal stromal cells (MSCs) are potent regulators of inflammation through the production of anti-inflammatory cytokines, chemokines, and growth factors. These cells show an ability to regulate neuroinflammation in the context of neurodegenerative diseases such as prion disease and other protein misfolding disorders. Prion diseases can be sporadic, acquired, or genetic; they can result from the misfolding and aggregation of the prion protein in the brain. These diseases are invariably fatal, with no available treatments.
One of the earliest signs of disease is the activation of astrocytes and microglia and associated inflammation, which occurs prior to detectable prion aggregation and neuronal loss; thus, the anti-inflammatory and regulatory properties of MSCs can be harvested to treat astrogliosis in prion disease. Recently, we showed that adipose-derived MSCs (AdMSCs) co-cultured with BV2 cells or primary mixed glia reduce prion-induced inflammation through paracrine signaling. This paper describes a reliable treatment using stimulated AdMSCs to decrease prion-induced inflammation.
A heterozygous population of AdMSCs can easily be isolated from murine adipose tissue and expanded in culture. Stimulating these cells with inflammatory cytokines enhances their ability to both migrate toward prion-infected brain homogenate and produce anti-inflammatory modulators in response. Together, these techniques can be used to investigate the therapeutic potential of MSCs on prion infection and can be adapted for other protein misfolding and neuroinflammatory diseases.
Glial inflammation plays a key role in a variety of neurodegenerative diseases, including Parkinson's, Alzheimer's, and prion disease. Although abnormal protein aggregation is attributed to much of disease pathogenesis and neurodegeneration, glial cells also play a part in exacerbating this 1,2,3. Therefore, targeting glial-induced inflammation is a promising therapeutic approach. In prion disease, the cellular prion protein (PrPC) misfolds to the disease-associated prion protein (PrPSc), which forms oligomers and aggregates and disrupts homeostasis in the brain 4,5,6.
One of the earliest signs of prion disease is an inflammatory response from astrocytes and microglia. Studies suppressing this response, either by removal of microglia or modification of astrocytes, have generally shown no improvement on, or worsened, disease pathogenesis in animal models 7,8,9. Modulating glial inflammation without eliminating it is an intriguing alternative as a therapeutic.
Mesenchymal stromal cells (MSCs) have taken the stage as a treatment for a variety of inflammatory diseases, due to their ability to modulate inflammation in a paracrine manner 10,11. They have shown the ability to migrate to sites of inflammation and respond to signaling molecules in these environments by secreting anti-inflammatory molecules, growth factors, microRNAs, and more 10,12,13. We have previously demonstrated that MSCs derived from adipose tissue (denoted AdMSCs) are able to migrate toward prion-infected brain homogenate and respond to this brain homogenate by upregulating gene expression for anti-inflammatory cytokines and growth factors.
Moreover, AdMSCs can decrease the expression of genes associated with Nuclear Factor-kappa B (NF-κB), the Nod-Like Receptor family pyrin domain containing 3 (NLRP3) inflammasome signaling, and glial activation, in both BV2 microglia and primary mixed glia 14. Here, we provide protocols on how to isolate both AdMSCs and primary mixed glia from mice, stimulate AdMSCs to upregulate modulatory genes, assess AdMSC migration, and co-culture AdMSCs with prion-infected glia. We hope that these procedures can provide a foundation for further investigation of the role of MSCs in regulating glial-induced inflammation in neurodegenerative and other diseases.
Mice were bred and maintained at Colorado State's Lab Animal Resources, accredited by the Association for Assessment and Accreditation of Lab Animal Care International, in accordance with protocol #1138, approved by the Institutional Animal Care and Use Committee at Colorado State University.
1. Isolating and infecting primary cortical mixed glia with prions
Stimulating AdMSCs with TNFα or interferon-gamma (IFNγ) for 24 h induces changes in the expression of anti-inflammatory molecules and growth factors. Treating AdMSCs with TNFα or interferon-gamma (IFNγ) increases TNF-stimulated gene 6 (TSG-6) mRNA, whereas TNFα, but not IFNγ, causes an increase in transforming growth factor beta-1 (TGFβ-1) mRNA. Stimulation with TNFα or IFNγ induces an increase in vascular endothelial growth factor (VEGF)mRNA, but no change in the expression of fibroblast growth factor (FGF) mRNA (Figure 1A). These data suggest that AdMSCs likewise respond to factors in prion-infected brain homogenate, likely a milieu of cytokines and damage-associated molecular patterns (DAMPs) 18,19,20. We show that culturing AdMSCs for 24 h in media containing 0.1% normal brain homogenate (NBH) or Rocky Mountain Laboratories (RML) strain mouse-adapted scrapie brain homogenate increases expression of TSG-6 mRNA. Treatment with RML, but not NBH, increased mRNA for TGFβ-1, VEGF, and FGF (Figure 1B).
To assess the role AdMSCs play in decreasing inflammation in both microglia and astrocytes, primary mixed glia were infected with 0.1% NBH or 22L mouse-adapted scrapie. At 7 days post infection, a co-culture system was established by adding AdMSCs to inserts and co-cultures for 7 days (Figure 2A). Analysis of the mixed glia demonstrated a significant increase in mRNA for the inflammatory cytokines CCL2, CCL5, and IL1β, and the astrocyte marker S100β, in cells infected with 22L compared to NBH. No significant changes were seen in mRNA expression for TNFα. After co-culturing for 7 days with AdMSCs, a decrease in CCL2, CCL5, and IL1β mRNA was seen in both NBH-treated and 22L-infected cells, but no significant changes were seen in TNFα mRNA expression. Co-culturing with AdMSCs also decreases the pan-astrocyte marker S100β and the reactive astrocyte marker C3 in both NBH and 22L-treated glia (Figure 2B).
The BV2 microglia cell line was treated with NBH or RML to assess changes specific to microglia after culturing with AdMSCs. Because stimulating AdMSCs with TNFα causes an increase in the production of the anti-inflammatory gene, as described above, AdMSCs were stimulated for 24 h prior to co-culturing with BV2 cells. At 6 days post prion exposure, BV2 cells were co-cultured with AdMSCs (Figure 2C). RNA was isolated after 24 h. Although we do not show any differences in inflammatory markers between NBH and RML-treated BV2 cells, we show that AdMSCs drastically decrease the transcription of pro-inflammatory markers in BV2 cells. A significant decrease was seen in the inflammatory cytokines IL1β, IL-6, and TNFα and the complement protein C1qa for both NBH-treated and RML-treated BV2 cells. Co-culturing with AdMSCs decreased the M1 microglial gene CD-16 and increased mRNA for Arg-1, a marker for M2 microglia, in both NBH-treated and RML-treated BV2 cells (Figure 2D).
Not only does stimulating MSCs increase the production of anti-inflammatory molecules and growth factors, but it also improves their ability to migrate to areas of inflammation 13,21. Here, we demonstrate an in vitro assay to assess AdMSC migration toward the prion-infected brain. AdMSCs can be stimulated with TNFα for 24 h, then serum-starved for 4 h, then added to inserts above wells containing media only, or media with 1% NBH or RML (Figure 3A). After 24 h, some cells migrate toward media only and NBH-containing media, but there is increased migration of cells toward RML-containing media (Figure 3B,C).
Figure 1: Induction of the expression of anti-inflammatory genes and growth factors after stimulating AdMSCs with cytokines or prion-infected brain homogenate. AdMSCs at passage 3 were stimulated for 24 h with (A) 10 ng/mL TNFα, which induced the expression of mRNA for TSG-6, TGFβ-1, and VEGF, or with 200 ng/mL IFN-α for 24 h induced the expression of TSG-6 and VEGF. (B) Stimulation for 24 h with 0.1% NBH induced expression of TSG-6 mRNA. Stimulation with 0.1% RML prion-infected brain homogenate induced expression of mRNA for TSG-6, TGFβ-1, VEGF, and FGF. One-way ANOVA with post-hoc Tukey's test, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, error bars = SEM. Combined data from four separate experiments, each with three technical replicates. This figure is adapted from Hay et al.22 Abbreviations: AdMSCs = adipose-derived mesenchymal stromal cells; TNFα = tumor necrosis factor alfa; IFN-α = interferon-gamma; TSG-6 = TNF-stimulated gene 6; TGFβ-1 = transforming growth factor beta-1; VEGF = vascular endothelial growth factor; NBH = normal brain homogenate; RML = Rocky Mountain Laboratory prion strain. Please click here to view a larger version of this figure.
Figure 2: Decrease in markers of inflammation in prion-treated primary mixed glia and BV2 microglia following co-culture with AdMSCs. (A) Primary mixed glia were infected with 0.1% NBH or 22L prion-infected brain homogenate for 72 h, thoroughly washed, then cultured in normal media for 96 additional h. AdMSCs were added to inserts above the corresponding wells. Glia-AdMSC cocultures were incubated for an additional 7 days, for a total of 14 days from the initial infection. (B) Glia infected with 22L showed increased expression of mRNA for the inflammatory markers CCL2, CCL5, and IL1β, in addition to the astrocyte marker S100β compared to those treated with NBH. Co-culturing with AdMSCs decreased expression of the inflammatory cytokines CCL2, CCL5, and IL1β, but not TNFα, for both NBH and 22L-infected cells, as well as decreasing expression of S100β and C3, which together are markers of reactive astrocytes. (C) BV2 microglia were treated with 0.1% NBH or RML prion-infected brain homogenate for 72 h, thoroughly washed, then cultured in normal media for 72 additional h. AdMSCs were stimulated with 10 ng/mL TNFα for 24 h, then added to inserts above corresponding wells. BV2-AdMSC co-cultures were incubated for 24 h, for a total of 7 days from initial brain homogenate treatments. (D) BV2s co-cultured with AdMSCs showed a decrease in mRNA for the inflammatory markers IL1β,IL-6, and TNFα, and the complement protein C1qa. Additionally, AdMSCs decreased the M1 microglia gene CD-16 and increased the M2 microglia marker Arg-1. Two-way ANOVA with post-hoc Tukey's test, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, error bars = SEM. Combined data from three separate experiments, each with three technical replicates. This figure is adapted from Hay et al.14. Abbreviations: AdMSCs = adipose-derived mesenchymal stromal cells; IL = interleukin; TNFα = tumor necrosis factor alfa; NBH = normal brain homogenate; 22L = 22L mouse-adapted scrapie prion strain. Please click here to view a larger version of this figure.
Figure 3: AdMSCs migration towards prion-infected brain homogenate in an in vitro model. AdMSCs were stimulated for 24 h with 10 ng/mL TNFα, then serum-starved for 4 h. (A) They were plated in inserts with a pore size of 8 µm above media alone, or media containing 1% NBH, or 1% RML prion-infected brain homogenate, and incubated for 24 h. (B) AdMSCs were stained with crystal violet solution and cells that had migrated to the bottom side of the insert were imaged and counted manually. (C) Migration of AdMSCs toward NBH-containing media is not significantly more than toward untreated media. However, AdMSCs show significant migration toward media containing 1% RML. Data are a combination of three biological replicates each with three technical replicates, with four fields of vision for each technical replicate. One-way ANOVA with post-hoc Tukey's test, **p < 0.001, ****p < 0.0001, error bars = SEM. Scale bars = 10 µm (B). This figure is adapted from Hay et al.14. Abbreviations: AdMSCs = adipose-derived mesenchymal stromal cells; TNFα = tumor necrosis factor alfa; NBH = normal brain homogenate; RML = Rocky Mountain Laboratory mouse-adapted scrapie prion strain. Please click here to view a larger version of this figure.
Table 1: Murine primer sequences for reverse transcriptase quantitative PCR. Please click here to download this Table.
Here we demonstrate a reliable and relatively inexpensive protocol for assessing the effects of adipose-derived mesenchymal stromal cells (AdMSCs) in decreasing prion-induced inflammation in a glial cell model. AdMSCs can easily be isolated and expanded in culture for use in as little as 1 week. This protocol consistently produces a heterologous population of cells that express markers consistent with those of mesenchymal stromal cells by immunofluorescence and flow cytometry, and retain immunological function when introduced to cytokines or prion-infected brain homogenate 14. The data shown here indicate that AdMSCs upregulate anti-inflammatory cytokines and growth factors when exposed to these stimuli. This property can be harnessed to "stimulate" AdMSCs to produce molecules that promote the decrease in inflammation. Stimulating AdMSCs with proinflammatory cytokines may increase their production of specific genes of interest 14,23,24, including anti-inflammatory genes and growth factors such as TSG-6, TGFβ-1, and VEGF. Treatment with brain homogenate from prion-infected animals increases the production of TSG-6, TGFβ-,1 VEGF, and FGF (Figure 1), as detected by qRT-PCR. Different prion strains can have different structures, tissue tropisms, and clinical signs, which poses difficulties when developing therapeutics25. However, they induce a similar inflammatory profile in animal models19. To highlight their universal effectiveness in response to inflammation, we have demonstrated that AdMSCs respond to both 22L and RML, two strains of mouse-adapted scrapie, to upregulate the expression of anti-inflammatory molecules14,22. We also see similar protection although less significant in normal, non-prion infected, brain homogenates (NBH). These brain homogenates contain circulating inflammatory cytokines and tissue, stimulating BV2 cells to phagocytose-a known phenomenon of microglial cells26,27.
Co-culturing with AdMSCs can be done with prion-infected primary mixed glia or BV2 microglia to reduce inflammatory markers (Figure 2). However, there are specific limitations to each type of culture. Once primary mixed glia reach confluency, they can grow for up to 2 months and can maintain prion infection from 7 to 28 days 14. From our experience, BV2 cells can be difficult to culture for more than 7 days before they become 90-100% confluent and can spontaneously activate and die. Lastly, AdMSCs can be stimulated prior to co-culturing which may enhance their anti-inflammatory profile 14,23,24, but they also show anti-inflammatory abilities without stimulation. Here, we demonstrate the abilities of AdMSCs to decrease inflammation in an in vitro model of BV2 microglia and primary mixed glia that have been exposed to prion-infected brain homogenates. Co-culturing with AdMSCs results in a decrease in mRNA for inflammatory cytokines such as CCL2, CCL5, and IL1β, as well as a decrease in the markers of reactive astrocytes, S100β and C3 in primary mixed glia. Co-culturing AdMSCs with BV2 microglia leads to a reduction in mRNA for IL1β and IL-6. TNFα and C1qa, two molecules produced by microglia that contribute to the development of neurotoxic astrocytes, were also decreased when BV2 microglia were co-cultured with AdMSCs. mRNA for the M1 microglia marker CD-16 was decreased, and the M2 microglia marker Arg-1 was increased in co-cultured BV2 cells. Together, these data suggest that AdMSCs polarize microglia to an M2 protective phenotype and astrocytes to a homeostatic non-toxic phenotype. Our recent work using AdMSCs to combat in vivo prion disease in a mouse model further corroborates these in vitro findings22. Further investigation by our research group is forthcoming to characterize the specific factors that are being secreted by AdMSCs responsible for modifying glial inflammation. Previous literature demonstrates that BV2 microglia do not up-regulate inflammatory genes at the same level that primary microglia do in response to inflammogens such as LPS28 and is supported by these data when comparing BV2 microglia (Figure 3) to primary glia (Figure 2). Furthermore, previous research demonstrates that inflammatory response is further exacerbated in mixed glial cultures containing both astrocytes and microglia14, suggesting that prion-induced inflammatory response is increased due to glial-glial communication.
We demonstrate the ability of TNFα-stimulated AdMSCs to migrate toward prion-infected brain homogenate in an in vitro model (Figure 3). AdMSCs can be stimulated prior to migration assays and this may promote their migratory abilities 13,14,21, or MSCs can remain unstimulated and migrate to prion brain homogenate in vitro, as was done in the protocol from which this migration assay was adapted 29. The factors used to stimulate AdMSCs, the length of time allotted for migration, and the stimulus the cells migrate towards are some of many variables that can be assessed using this AdMSC migration assay.
The mechanisms AdMSCs utilize to decrease markers of glial inflammation appear to vary depending on the length of co-culturing and the type of cells the AdMSCs are co-cultured with. Therefore, this protocol lends itself to be adapted to answer a variety of questions in the context of prion disease and can be expanded to investigate the use of AdMSCs in other inflammatory diseases. Here, we demonstrate changes in primary mixed glia derived from C57Bl/6 mice. These changes occur after cells are exposed to 22L prion-infected brain homogenate for 7 days and cultured with AdMSCs for an additional 7 days. This protocol focuses on changes at the mRNA level, but evaluating changes on the protein level with enzyme-linked immunosorbent assay or western blot may be beneficial. Factors to consider manipulating include isolating glia from various strains of mice and changing the length of AdMSC exposure or time of infection, as mixed glia may retain infectious prions for more than 28 days 14. Here, we use BV2 cells, a microglia cell line, but primary microglia and ex vivo microglia may have more biological relevance. Additionally, the source of prions can be modified to use purified prions 30 or brain homogenates from various strains, as these should demonstrate variation in inflammatory properties 19. Moreover, MSC co-cultures have been used to downregulate inflammation in macrophages exposed to toxins such as lipopolysaccharide 31, and this protocol has the potential to be adapted for uses outside of prion and other neuroinflammatory diseases.
The authors have nothing to disclose.
The authors thank Lab Animal Resources for their animal husbandry. Our funding sources for this manuscript include the Boettcher Fund, the Murphy Turner Fund, CSU College of Veterinary Medicine, and the Biomedical Sciences College Research Council. Figure 2A, Figure 2C, and Figure 3A were created with BioRender.com.
0.25% Trypsin | Cytiva | SH30042.01 | |
5 mL serological pipets | Celltreat | 229005B | |
6-well tissue culture plates | Celltreat | 229106 | |
10 cm cell culture dishes | Peak Serum | PS-4002 | |
10 ml serological pipets | Celltreat | 229210 | |
15 mL conical tubes | Celltreat | 667015B | |
50 mL conical tubes | Celltreat | 667050B | |
BV2 microglia cell line | AcceGen Biotech | ABC-TC212S | |
Cell lifter | Biologix Research Company | 70-2180 | |
Crystal violet | Electron Microscopy Sciences | 12785 | |
Dispase | Thermo Scientific | 17105041 | |
DMEM/F12 | Caisson Labs | DFL14-500ML | |
DNase-I | Sigma Aldrich | 11284932001 | |
Essential amino acids | Thermo Scientific | 11130051 | |
Ethanol (100%) | EMD Millipore | EX0276-1 | |
Fetal bovine serum (heat inactivated) | Peak Serum | PS-FB4 | Can be purchased as heat inactivated or inactivated in the laboratory |
Formaldehyde | EMD Millipore | 1.04003.1000 | |
Glass 10 mL serological pipet | Corning | 7077-10N | |
Hank’s Balances Salt Solution | Sigma Aldrich | H8264-500ML | |
Hemocytometer/Neubauer Chamber | Daigger | HU-3100 | |
High Glucose DMEM | Cytiva | SH30022.01 | |
low glucose DMEM containing L-glutamine | Cytiva | SH30021.01 | |
MEM/EBSS | Cytiva | SH30024.FS | |
non-essential amino acids | Sigma-Aldrich | M7145-100M | |
Paraformaldehyde (16%) | MP Biomedicals | 219998320 | |
Penicillin/streptomycin/neomycin | Sigma-Aldrich | P4083-100ML | |
Phosphate buffered saline | Cytiva | SH30256.01 | |
Recombinant Mouse IFN-gamma Protein | R&D Systems | 485-MI | |
Recombinant Mouse TNF-alpha (aa 80-235) Protein, CF | R&D Systems | 410-MT | |
RNeasy mini kit | Qiagen | 74104 | |
Sigmacote | Sigma Aldrich | SL2-100ML | Coat inside of glass pipets by aspirating up and down twice in Sigmacote and allowing to dry thoroughly. Wrap in aluminum foil and autoclave pipets 24 h later. |
Stemxyme | Worthington Biochemical Corporation | LS004106 | Collagenase/Dispase mixture |
Sterile, individually wrapped cotton swab | Puritan Medical | 25-8061WC | |
Thincert Tissue Culture Inserts, 24 well, Pore Size=8 µm | Greiner Bio-One | 662638 | |
Thincert Tissue Culture Inserts, 6 well, Pore Size=0.4 µm | Greiner Bio-One | 657641 |