sign in sign up
<<
Home , Central nervous system, Myelin, Nervous tissue, Oligodendrocyte, Synapse

Direct and Indirect Effects of Immune and Central −Resident Cells on Progenitor Differentiation This information is current as of October 1, 2021. Craig S. Moore, Qiao-Ling Cui, Nebras M. Warsi, Bryce A. Durafourt, Nika Zorko, David R. Owen, Jack P. Antel and Amit Bar-Or J Immunol 2015; 194:761-772; Prepublished online 10

December 2014; Downloaded from doi: 10.4049/jimmunol.1401156 http://www.jimmunol.org/content/194/2/761

Supplementary http://www.jimmunol.org/content/suppl/2014/12/10/jimmunol.140115 http://www.jimmunol.org/ Material 6.DCSupplemental References This article cites 60 articles, 7 of which you can access for free at: http://www.jimmunol.org/content/194/2/761.full#ref-list-1

Why The JI? Submit online.

by guest on October 1, 2021 • Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2015 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Direct and Indirect Effects of Immune and –Resident Cells on Human Oligodendrocyte Progenitor Cell Differentiation

Craig S. Moore, Qiao-Ling Cui, Nebras M. Warsi, Bryce A. Durafourt, Nika Zorko, David R. Owen,1 Jack P. Antel, and Amit Bar-Or

In , successful within the injured CNS is largely dependent on the survival and differentiation of oligodendrocyte progenitor cells. During inflammatory injury, and oligodendrocyte progenitor cells within sites are exposed to secreted products derived from both infiltrating immune cell subsets and CNS-resident cells. Such products may be considered either proinflammatory or anti-inflammatory and have the potential to contribute to both injury and repair processes. Within the CNS, also contribute significantly to oligodendrocyte during development and following inflammatory injury. The overall objective of the current study was to determine how functionally distinct proinflammatory and anti- Downloaded from inflammatory human immune cell subsets, implicated in multiple sclerosis, can directly and/or indirectly (via astrocytes) impact human oligodendrocyte progenitor cell survival and differentiation. Proinflammatory T cell (Th1/Th17) and M1-polarized myeloid cell supernatants had a direct cytotoxic effect on human A2B5+ neural progenitors, resulting in decreased O4+ and GalC+ oligodendrocyte lineage cells. -conditioned media collected from astrocytes pre-exposed to the same proinflammatory supernatants also resulted in decreased oligodendrocyte progenitor cell differentiation without an apparent increase in cell death and was mediated through astrocyte-derived CXCL10, yet this decrease in differentiation was not observed in the more differ- http://www.jimmunol.org/ entiated oligodendrocytes. Th2 and M2 or supernatants had neither a direct nor an indirect impact on oligodendrocyte progenitor cell differentiation. We conclude that proinflammatory immune cell responses can directly and indirectly (through astrocytes) impact the fate of immature oligodendrocyte-lineage cells, with oligodendrocyte progenitor cells more vulnerable to injury compared with mature oligodendrocytes. The Journal of Immunology, 2015, 194: 761–772.

hronic inflammation within the CNS plays a key role in including TNF-a, compared with more mature oligodendrocyte- the pathophysiology of multiple sclerosis (MS). Although lineage cells (4–7). C inflammation can result in demyelination and axonal Cellular interactions between proinflammatory immune cells and by guest on October 1, 2021 damage, it can also stimulate and repair (1). CNS-resident cells were suggested to promote CNS injury. In Given these observations, it is hypothesized that complex neuro- contrast, interactions between “anti-inflammatory” immune cells and immunological interactions between immune cells and neural cells CNS-resident cells promote tissue repair and/or neuroprotection (8). are important determinants of the net tissue effects of CNS in- Although this concept has been supported in models of CNS flammation (1, 2). In MS, successful remyelination in the injured inflammation and is consistent with the presumed mechanisms of CNS is dependent on the survival and differentiation of oligoden- action of several immunomodulatory disease-modifying therapies drocyte progenitor cells (OPCs) rather than process regeneration by (DMTs) used in the treatment of MS, there have been no studies mature, myelinating oligodendrocytes. Our recent in situ anal- directly assessing the contributions of human disease–relevant im- ysis of active MS suggested that OPCs are more vul- mune cell subsets (including T cells, , and microglia) to nerable to injury than mature oligodendrocytes within the same the biology and fate of human OPCs. Given the limited capacity of lesion site (3). In vitro and animal studies also implicated en- OPCs to differentiate and, subsequently, remyelinate with disease hanced vulnerability of OPCs to proinflammatory mediators, duration (9), it is critical to understand the contributions of resident neural cells (including microglia and astrocytes) and peripheral- derived inflammatory immune cells in this process. Unit, Department of and , Montreal Neu- rological Institute, McGill University, Montreal, Quebec H3A 2B4, Canada In MS, infiltrating immune cells belonging to both the adaptive 1Current address: Division of Sciences, Department of Medicine, Imperial and innate immune systems can physically interact with CNS- College London, London, U.K. resident cells, but they also secrete soluble immunoregulatory Received for publication May 8, 2014. Accepted for publication November 10, 2014. and neuroregulatory factors that can influence the biology of cells Address correspondence and reprint requests to Prof. Amit Bar-Or, Montreal Neuro- residing in the CNS (1). Among the cells of the adaptive immune logical Institute, Neuroimmunology Unit, Room 111, 3801 University Street, Mon- system, T cells are to play a particularly prominent role in treal, QC H3A 2B4, Canada. E-mail address: [email protected] MS pathophysiology (8). In attempts to model the contributions of The online version of this article contains supplemental material. these cells in vitro, T cells can be polarized into three distinct Abbreviations used in this article: ACM, astrocyte-conditioned media; DMT, disease- modifying therapy; IHC, immunohistochemistry; MDM, monocyte-derived macro- subsets: Th1, Th17, and Th2. In the context of MS, Th1- and Th17- phage; MS, multiple sclerosis; OPC, oligodendrocyte progenitor cell; SCI, spinal polarized T cells are considered proinflammatory, while promoting cord injury. damage and demyelination in the CNS. Conversely, Th2-polarized Copyright Ó 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00 immune cells are generally considered anti-inflammatory and are www.jimmunol.org/cgi/doi/10.4049/jimmunol.1401156 762 IMMUNE CELL REGULATION OF OLIGODENDROCYTE DIFFERENTIATION believed to protect the CNS from damaging inflammation and #50,000 pg/ml IFN-g, which was significantly less than in Th1-polarized potentially promote repair (8). supernatants. Supernatants of PBMCs that were activated, but not polarized, were used as controls. In the CNS, innate immune cells belonging to the myeloid Polarized myeloid cell supernatants were generated using previously pub- lineage, including blood-derived macrophages and CNS-resident lished protocols (19). Comprehensive characterization of the phenotype and microglia, also can play a significant role in the pathophysiology functional properties of our polarized human macrophages and microglia was + of MS. Similar to T cells, myeloid cells can be polarized into performed previously (11, 12, 19). Briefly, CD14 monocytes were positively selected from PBMCs using anti-CD14 MicroBeads (Miltenyi Biotec, proinflammatory M1 (classical activation) and anti-inflammatory 5 Auburn, CA). Cells were then plated at a density of 5 3 10 cells/ml in RPMI M2 (alternative activation) phenotypes (10–12). We showed pre- 1640 containing 10% FCS, penicillin/streptomycin, and glutamine. For M1 viously that exposure of human microglia to Th1 supernatants can polarization, monocytes were treated with rGM-CSF (5 ng/ml) and cul- induce an M1-like phenotype (13, 14). Although M1 macrophages tured for 5 d. Cells were then activated with IFN-g (20 ng/ml) for 1 h, and microglia secrete proinflammatory and are pre- followed by LPS (100 ng/ml) for 48 h. For M2 polarization, cells were treated with M-CSF (25 ng/ml) for 5 d, followed by stimulation with IL-4 dominant in active MS lesions, phagocytic myelin-laden macro- (20 ng/ml) and IL-13 (20 ng/ml) for 48 h. Human adult microglia were phages expressing M2 markers, such as CD206 and CD163, are cultured and polarized using protocols similar to those described previ- also present in lesions and the (15). Interest- ously (19). TNF-a, IL-12, and IL-10 ELISAs were used to validate po- larized myeloid cell supernatants. For myeloid cell polarizations, ingly, M2 myeloid cells also were shown to be present during + remyelination and secrete factors that can directly enhance oli- unpolarized supernatants were generated from naive CD14 monocyte- derived macrophages (MDMs) and used as a control to ensure that all godendrocyte differentiation in rodents (16). effects were due to specifically polarized myeloid cell subsets rather than In addition to the direct effects of immune cells on OPC biol- nonpolarized, baseline myeloid cell responses. ogy and differentiation, immune cells can interact and activate Downloaded from astrocytes in the CNS (17). Astrocytes are highly prevalent in Human astrocyte culture and generation of inflammatory MS lesions, and reactive is one of the astrocyte-conditioned media major pathological features of MS (17). During both CNS de- Human astrocytes were derived from human fetal CNS tissue (cerebral velopment and injury, astrocytes secrete a variety of factors hemispheres, gestational age 14–20 wk) from the human tissue repository that can influence OPC survival, migration, differentiation, and at the Albert Einstein College of Medicine (Bronx, NY). All procedures related to the use of these cells followed established institutional (McGill subsequent myelination (18). Therefore, astrocytes can be con- University, Montreal, QC, Canada) and Canadian Institutes of Health http://www.jimmunol.org/ sidered important regulators of myelination and remyelination in Research guidelines for the use of human cells. Protocols used to isolate the CNS. astrocytes from CNS tissue were described previously (4). Briefly, fetal In the current study, we assessed how peripheral-derived and CNS tissue was treated with DNase/trypsin and passed through a nylon mesh to obtain a single-cell suspension. Cells were then plated in CNS-resident immune cells can directly and indirectly influence 6 aflaskat73 10 cells/ml in DMEM containing 10% FCS, penicillin/ OPC survival and differentiation. We demonstrated that soluble streptomycin, and glutamine. All cells were kept under sterile conditions products derived from proinflammatory T cell and myeloid cell in incubators at 37˚C and passaged every 7–14 d. To ensure cell purity, populations can directly and indirectly (via astrocytes) impact experiments were conducted on astrocytes at the third or fourth passage. . + OPC survival and differentiation in a negative manner. We also The astrocyte cultures were 98% GFAP , as determined by flow cytometry (20). identified two distinct mechanisms by which these cells can To assay the effects of immune cell supernatants on astrocytic expres- by guest on October 1, 2021 impact OPCs: cytotoxicity mediated by proinflammatory medi- sion of factors relevant for myelin repair, astrocytes were incubated with ators and noncytotoxic inhibition of differentiation mediated by supernatants derived from individually polarized immune cell subsets. For this procedure, astrocytes were trypsinized with a 2.5% trypsin-EDTA distinct chemokines. We further demonstrate that immature hu- 5 man OPCs are more sensitive to these factors than are mature solution, washed, centrifuged, and plated at 10 cells/ml in defined me- dia. Supernatants were added to the astrocytes at 80% confluency for 48 h, oligodendrocytes. after which astrocytes were collected for RNA analysis. Materials and Methods Human adult microglia culture and polarization Total PBMC and myeloid cell polarizations and supernatant Human adult microglia were isolated from temporal lobe tissue from generation patients undergoing surgery for nontumor-related intractable . All tissue obtained was outside of the suspected focal site, and cells were PBMCs were isolated by Ficoll separation from the venous blood of healthy isolated as previously described (19). Histological sections made from human volunteers (aged 19–50 y). PBMCs were polarized to Th1, Th17, or human CNS material were cut and stained for GFAP and CD68 to assess Th2 subtypes using whole-PBMC polarization protocols. All PBMCs astrocyte and microglial activity. Previous studies found no evidence of were cultured at a density of 1 3 106 cells/ml in RPMI 1640 medium with significant astrogliosis or increased microglial activity in control, autopsy- 10% FCS, penicillin/streptomycin, and glutamine. Th1 polarizations were derived tissue. Following isolation, human adult microglia are quiescent, performed using anti-CD3 (1 mg/ml), IL-12 (10 ng/ml), and anti-IL-4 do not secrete proinflammatory factors, and do not express surface markers (5 mg/ml). Th17 polarizations were generated using anti-CD3 (0.3mg/ml), that would be indicative of an activated state (11). Briefly, tissue was anti-CD28 (1 mg/ml), anti–IFN-g (5 mg/ml), anti–IL-4 (5 mg/ml), and IL-23 treated with DNase/trypsin and passed through a nylon mesh to obtain (10 ng/ml). All cytokines and Abs were from R&D Systems (Minneapolis, a single-cell suspension. To remove the myelin, specimens were placed in MN). For Th1 and Th17 polarizations, cells were cultured for 5 d at 37˚C a linear 30% Percoll density gradient. The remaining cell layer was cul- in 5% CO . Following incubation, cells were centrifuged, and the super- 2 tured in 5% MEM with 5% FCS, penicillin/streptomycin, and glutamine. natants were collected and stored at 280˚C. Th2 polarizations were gen- After 24 h, the floating fraction was removed, and the remaining adherent erated using anti-CD3 (1 mg/ml), IL-4 (200 ng/ml), anti–IL-12 (5 mg/ml), cells (i.e., microglia) were grown for an additional 5 d prior to use. Human and anti–IFN-g (5 mg/ml). Cells were cultured for 5 d, followed by the adult microglia were polarized using the same protocols as described addition of fresh media containing IL-2 (20 U/ml) to expand the Th2- previously (19) and detailed above. polarized cultures. After 48 h, supernatants were collected and stored at 280˚C. Confirmation of adequately polarized supernatants was performed Human OPC isolation, culture, and treatment using ELISAs (BD Biosciences, Mississauga, ON, Canada and eBio- science, San Diego, CA). Specifically, IFN-g (Th1), IL-5 (Th2), and IL-17 Human OPCs were isolated from human fetal CNS tissue (cerebral hemi- (Th17) levels were measured in the T cell–polarized superna- spheres, gestational age 14–20 wk) from the human tissue repository at the tants (Supplemental Fig. 1). All Th1-polarized supernatants contained Albert Einstein College of Medicine. All procedures related to the use of these .150,000 pg/ml IFN-g, Th17 supernatants had $1,500 pg/ml IL-17, and cells followed established institutional and Canadian Institutes of Health Th2 supernatants contained $500 pg/ml IL-5. Furthermore, all Th2 Research guidelines. Protocols used to isolate the OPCs from the CNS tissue supernatants contained levels of IFN-g that were below the lower limit of were described previously (3). Briefly, fetal CNS tissue was treated with detection for our assay (100 pg/ml). All Th17 supernatants contained DNase/trypsin and passed through a nylon mesh to obtain a single-cell The Journal of Immunology 763 suspension. A2B5+ neural progenitor cells were positively selected using anti- Immunocytochemistry and quantification A2B5 MicroBeads (Miltenyi Biotec). PDGFRa+ cells were selected with a monoclonal mouse anti-human CD140a Ab (AbD Serotec, Kidlington, U.K.), Using live cultures, cells were stained with O4 (IgM hybridoma; 1:50), followed by rat anti-mouse IgG1 magnetic beads (Miltenyi Biotec). Cells GalC (IgG3 hybridoma; 1:50), or O1 (IgM hybridoma; 1:25) Abs, followed were cultured on poly-L-lysine and extracellular matrix–coated 48-well by fixation with 4% PFA. An MBPAb (Covance, Princeton, NJ; 1:1000) was plates, at 105 cells/well, for 7–10 d in defined media consisting of DMEM- used to stain mature oligodendrocytes. Cells also were stained using F12 supplemented with N1 (Sigma-Aldrich, Oakville, ON, Canada), 0.01% a primary Ab for Ki67 (Millipore, Billerica, MA; 1:50) to assess cell proliferation. Appropriate secondary Abs conjugated with either Cy3 or BSA, penicillin/streptomycin, B27 supplement (Invitrogen, Burlington, ON, + + + Canada), PDGF-AA (10 ng/ml), basic fibroblast growth factor (10 ng/ml), FITC were used. Quantification of DAPI ,O4, and GalC cells was and T3 (2 nM). In experiments designed to assay the direct effects performed using MetaXpress Imaging software. Cell percentages were of polarized immune cell and microglia supernatants on OPC differentiation, obtained by dividing the numbers of positively stained cells by the total A2B5-selected cells were treated with a 1:1 ratio of defined media/polarized cell counts. OPC morphologies were scored by a blinded observer. supernatants for 48 h. Thereafter, cells were immunostained immediately and fixed or cultured for an additional 4 d in defined media. To determine the direct TUNEL assay effects of individual cytokine treatments on OPC differentiation, A2B5- To determine OPC survival, a TUNEL assay was used to quantitate O4+ selected cells were treated with various polarizing agents, including IFN-g apoptotic cells. Cells were fixed in acetone/ (1:1) for 10–15 (7 ng/ml), LPS (35 ng/ml), TNF-a (2 ng/ml), IL-12 (3 ng/ml), and IL-10 (100 min, washed, and incubated for 1 h with TdT (Promega, Madison, WI) pg/ml). Neutralization experiments were performed using a function-blocking and biotinylated dUTP (Roche, Basel, Switzerland) in TdT buffer, anti-human TNF-a Ab (R&D Systems). Briefly, proinflammatory Th1 or M1 according to the manufacturer’s instructions. Cells were blocked in 10% culture supernatants were incubated overnight with either control IgG or anti– . goat serum and 2% horse serum for 5 min, followed by incubation with TNF-a Ab (100 ng/ml). Confirmation of TNF-a blockade ( 75%) was streptavidin-FITC (Jackson ImmunoResearch, West Grove, PA; 1:1000) measured using a human TNF-a ELISA (BD Bioscience). In additional ex- for 30 min. periments exploring signaling pathways involved in the increased of

+ Downloaded from OPCs treated with immune cell supernatants, human A2B5 progenitor RNA isolation and quantitative PCR cells were treated with either immune cell–conditioned media or rTNF-a and grown in the presence or absence of Z-vad–FMK, a pan caspase inhibitor, TRIzol reagent (QIAGEN, Germantown, MD) was used to lyse astrocytes, for 24 h. Following fixation, cells were visualized using Abs for O4, and cell and subsequent RNA isolation was performed under strictly RNase-free death was measured using propidium iodide and TUNEL staining. conditions. RNA was quantified using a UV spectrophotometer to mea- In experiments designed to assay the indirect effects (via astrocytes) of sure concentration and ensure purity (A260/A280 ratio . 1.8). Each RNA polarized immune cell and microglia supernatants on OPC differentiation, sample was diluted to the same concentration and treated with recombi-

astrocytes were treated with immune cell supernatants for 24 h, subjected to nant DNase-I (Invitrogen) to remove any genomic DNA contamination. http://www.jimmunol.org/ a series of thorough washes, and cultured in N1 medium for an additional Finally, murine reverse transcriptase (Invitrogen) was used to generate 48 h. Astrocyte-conditioned media (ACM) was collected and stored at cDNA, according to the manufacturer’s instructions, and stored at -20˚C –80˚C. A2B5-selected cells were treated with a 1:1 ratio of N1 media/ACM until use. for 48 h. In attempts to identify human astrocyte–derived factors responsible TaqMan quantitative PCR was used to measure mRNA expression for the deleterious effects of ACM collected following Th1 exposure, levels for PDGF-A, FGF-2, TIMP-1, IGF-1, IL-6, CNTF, CXCL10, and ELISAs were performed for IL-6, TIMP-1, and CXCL10 (IP-10). To assay LIF. Briefly, cDNA samples were mixed with a TaqMan MasterMix the direct effects of CXCL10 on human OPC differentiation, a CXCR3 (Invitrogen), RNase-free water, and primers for the gene of interest. Each inhibitor, AMG487 (1 mM; Tocris Bioscience, Bristol, U.K.), was applied sample was tested in duplicate and analyzed according to the ΔΔCT to OPCs for 1 h prior to the addition of ACM. method (21). by guest on October 1, 2021

FIGURE 1. Impact of polarized T cell supernatants on proportion of human O4+ cells. T cell supernatants were applied to human fetal-derived A2B5+ cells, and the percentage of cells expressing O4, an oligodendrocyte lineage progenitor cell marker, was measured on day 2. (A and B) The percentage of O4+ cells was decreased significantly following exposure to Th1 supernatants compared with either Th2 supernatants or media alone. (C) At day 6, an additional 4 d following removal of T cell supernatants, the percentage of O4+ cells also was decreased, yet the results were not statistically significant (p = 0.08). TUNEL labeling was performed in all conditions, including application of recombinant human TNF-a (100 ng/ml) as a positive control (D). The number of TUNEL+ cells was increased significantly in A2B5-selected cells treated with Th1 supernatants compared with N1 media alone. Scale bar, 200 mm. Bars represent the mean 6 SEM (n = 3/condition performed in duplicate). *p , 0.05, **p , 0.01. 764 IMMUNE CELL REGULATION OF OLIGODENDROCYTE DIFFERENTIATION

Statistical analyses a positive control. No effects were observed with Th2 super- All results were analyzed using Prism 6 (GraphPad, San Diego, CA). All natants. To explore the mechanism of Th1-mediated OPC apo- comparisons between OPC treatments were analyzed using a one-way ANOVA, ptotic injury, assays were performed using Z-vad, a caspase followedbytheTukeyposthoctest.Comparisonofastrocyticgrowthfactor inhibitor. The deleterious impact of Th1-conditioned media on expression levels was analyzed using one-way ANOVAs, and significant dif- O4+ cells among human A2B5+ progenitor cells was abrogated in ferences were determined using the Dunnett multiple-comparisons test. All comparisons with a p value , 0.05 were considered statistically significant. the presence of Z-vad (Supplemental Fig. 2A). Proinflammatory M1 supernatants decrease the percentage of Results O4+ and GalC+ OPCs compared with M2 supernatants or Proinflammatory Th1 supernatants decrease the percentage of media alone O4+ OPCs compared with Th2 supernatants or media alone To determine how differentially polarized human MDMs and To determine how differentially polarized human T cells can di- microglia can impact differentiation of O4+ progenitor cells rectly impact survival and differentiation of OPCs through sol- through soluble products, human A2B5-selected cells were cul- + uble products, human A2B5 cells were cultured with supernatants tured with supernatants derived from polarized macrophages and derived from polarized immune cell populations. Following a 48-h microglia. Similar to the results obtained with Th1-polarized treatment with Th1 supernatants, the percentage of O4+ cells was supernatants following a 48-h treatment with M1 macrophage significantly decreased compared with either Th2 supernatants or supernatants, the percentage of O4+ cells was significantly de- media control (Fig. 1A, 1B). After four additional days of cul- creased compared with either M2 supernatants or media alone turing, a similar trend was observed, although it was not statisti- (Fig. 2A). Similar results also were observed after four additional Downloaded from cally significant (Fig. 1C). After the 48-h exposure of polarized days of culturing (Fig. 2B). Following a 48-h treatment with hu- cell supernatants, a TUNEL assay also was performed on A2B5- man M1 microglia supernatants, the percentage of O4+ cells also selected cells to measure apoptosis. Compared with N1 media wasdecreasedsignificantlyatboth48h(Fig.2C)and6d alone, TUNEL+ cells were significantly increased following the (Fig. 2D). Prior to application upon OPCs, the profiles of all po- exposure to Th1 supernatants (Fig. 1D). Similar levels were larized myeloid cells were validated by standard assays, as pre- measured following treatment with rTNF-a, which was used as viously published (11), including flow cytometry (data not shown) http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 2. Impact of polarized human macrophage and microglia supernatants on the proportion of human O4+ cells. Exposure of human fetal-derived A2B5+ cells to macrophage M1 supernatants significantly decreased the percentage of cells expressing the oligodendrocyte lineage marker O4 on both day 2(A) and day 6 (B). Similarly (C and D), M1-polarized fetal microglia also significantly decreased the percentages of O4+ cells. Error bars represent the mean 6 SEM; n = 3/condition performed in duplicate. *p , 0.05, **p , 0.01, ***p , 0.001. The Journal of Immunology 765 and cytokine levels (Supplemental Fig. 1A–D). To confirm that OPCs (Supplemental Fig. 2B, 2C). In M1 supernatants, the role of any carryover of residual myeloid polarization factors used to TNF-a subsequently was validated using a function blocking anti– generate the supernatants was not responsible for the observed TNF-a Ab, which resulted in the restoration of the percentages of effects on O4+ cells, A2B5-selected cells were treated in parallel both O4+ (Fig. 4C) and GalC+ cells (Fig. 4D). No significant effects with either M-CSF or GM-CSF as controls. Direct application of were observed when using an IgG isotype control (data not shown). these growth factors did not affect total cell number; however, the In contrast, addition of a TNF-a–blocking Ab to Th1 supernatants numbers of GalC+ cells were increased significantly following the did not reverse the effects of the Th1 supernatants on the percent- addition of growth factors after 6 d in culture (Fig. 3C). None- age of O4+ or GalC+ cells or propidium iodide–labeled cells theless, direct application of M1-polarized supernatants signifi- (Supplemental Fig. 2D–F). This indicates that, unlike M1 myeloid cantly decreased the numbers of both O4+ and GalC+ cells at 48 h cells, which mediate OPC apoptosis in a TNF-a–dependent fashion, and after an additional 4 d in culture (i.e., 6 d total). At day 6, M2 the proinflammatory T cells secrete additional factors that are supernatants had significantly increased total O4+ and GalC+ cell responsible for the increase in apoptosis of human OPCs. numbers compared with media alone. Human astrocytes pre-exposed to polarized T and myeloid cell Implication of TNF-a as a contributing factor in decreasing supernatants secrete several soluble cytokines and growth OPC numbers factors known to influence OPC biology Based on the TUNEL assay results measuring apoptosis of A2B5- In addition to brain-resident and peripheral-derived immune cell selected cells (Fig. 1D), we sought to identify the soluble factor(s) populations, astrocytes are an important and significant contributor present in the Th1 and M1 supernatants that were responsible for of factors (immune and nonimmune) that are known to influence Downloaded from cell death, thus decreasing O4+ and GalC+ OPCs. Several cyto- OPC differentiation both during development and in the inflamed kines and molecules associated with the polarization of different brain. We assessed how the effects of astrocytes on OPCs were immune cell populations were added directly to the A2B5+ cells. impacted when the astrocytes were exposed to distinct disease- After 48 h, no differences in the total cell numbers (DAPI+) were implicated immune cell subsets. After incubating human astrocytes observed (Fig. 4A); however, after normalizing for total cell with Th1, Th17, Th2, or stimulated, yet nonpolarized, T cell super- number, the percentages of both O4+ and GalC+ cells were sig- natants for 48 h, we measured astrocyte gene expression of PDGF-A, nificantly decreased following direct addition of TNF-a (Fig. 4B). TIMP-1, CXCL10, IL-6, FGF-2, IGF-1, CNTF, and LIF. Pro- and http://www.jimmunol.org/ Indeed, Z-vad reversed the soluble TNF-mediated apoptosis of O4+ anti-inflammatory T cell supernatants differentially influenced by guest on October 1, 2021

FIGURE 3. Direct effects of polarized human fetal microglia supernatants on absolute numbers of total O4+ and GalC+ cells. Using MetaXpress Imaging software and ImageJ software, total cell numbers were quantified by counting DAPI+ cells. (A) At both days 2 and 6, no statistical differences in total cell numbers were observed among media treatment alone, growth factor treatment (M-MCF and GM-CSF), and polarized supernatants. Using the same methods, the numbers of O4+ and GalC+ cells were counted using IHC. (B) A significant decrease in the total numbers of O4+ cells was measured at both days 2 and 6 following direct application of M1 supernatants; at day 6, the direct application of M2 supernatants significantly increased the numbers ofO4+ cells compared with media alone. (C) Using a later oligodendrocyte lineage marker, GalC, a significant decrease in the total numbers was also observed at both days 2 and 6 following direct application of M1 supernatants. Direct application of M2 supernatants also significantly increased the numbers of GalC+ cells compared with media alone. Error bars represent the mean 6 SEM; n = 3/condition performed in duplicate. *p , 0.05, **p , 0.01, ***p , 0.001. 766 IMMUNE CELL REGULATION OF OLIGODENDROCYTE DIFFERENTIATION Downloaded from http://www.jimmunol.org/

FIGURE 4. Identification of soluble mediators produced by polarized immune cell subsets involved in OPC differentiation. (A) Following application of a number of M1- and Th1-associated soluble mediators, no significant differences in total cell number (DAPI+ cells) were observed. (B) TNF-a sig- nificantly decreased the percentage of cells expressing both oligodendrocyte lineage markers O4 and GalC compared with control conditions. (C) The ability of M1-polarized microglia supernatants to limit OPC differentiation from human A2B5+ progenitor cells into O4+ cells was decreased significantly by guest on October 1, 2021 when TNF-a was selectively neutralized in the supernatants. (D) TNF-a neutralization in the M1-polarized microglia supernatants also tended to decrease the supernatant effect on GalC+ cells, although it was not statistically significant (p = 0.08). Error bars represent the mean 3 SEM; n = 5/condition. *p , 0.05. astrocytic expression of these factors; Th1 supernatants signifi- Human astrocytes pre-exposed to Th1 or M1 supernatants cantly increased TIMP-1 (p = 0.0028), IL-6 (p = 0.0053), impair OPC differentiation, in part through CXCL10 , CXCL10 (p = 0.0039), and LIF (p 0.0001) expression com- To determine whether astrocytes pre-exposed to Th1 or M1 pared with control untreated astrocytes (Fig. 5A). Increased levels supernatants could functionally impair OPC differentiation, astro- of secreted IL-6, TIMP-1, and CXCL10 were confirmed by cytes were exposed as described above. Following thorough wash- ELISA (Supplemental Fig. 1). Similarly, proinflammatory Th17 ing, ACM were collected and applied to A2B5-selected cells. supernatants increased astrocyte TIMP-1 (p = 0.0047), IL-6 Compared with media controls, ACM collected from astrocytes (p = 0.0124), CXCL10 (p = 0.0028), and LIF (p = 0.0026) ex- pre-exposed to Th1 supernatants significantly decreased O4+ cell pression relative to control. Th1 and Th17 supernatants did not numbers. A reversal of this effect was measured when cells were significantly affect the expression of either IGF-1 or FGF-2 preincubated with a CXCL10 inhibitor (Fig. 6A, 6B). (Table I). To determine whether pro- and anti-inflammatory my- Cells exposed to ACM collected from astrocytes pre-exposed eloid cell supernatants also can influence growth factor expression to M1 microglia supernatants also tended to decrease O4+ cell by astrocytes, these cells were incubated with M1, M2, or unpo- numbers; however, the result was not statistically significant + larized supernatants derived from CD14 macrophages or micro- (Fig. 6C). To assess whether the decrease in O4+ cells was due to . Proinflammatory M1 macrophage supernatants significantly cell death, proliferation, or differentiation, total cell counts, increased astrocyte expression of PDGF-A (p = 0.0177), TIMP-1 TUNEL assays, and Ki67 immunocytochemistry were performed. (p = 0.0159), FGF-2 (p =0.0386),IL-6(p = 0.0231), and CXCL10 No differences in total cell numbers were observed (Supplemental (p = 0.0003) relative to untreated controls (Fig. 5B). IGF-1 ex- Fig. 3). In addition, no differences in TUNEL+ cells within either pression was decreased significantly in all astrocytes exposed to M1 the total cell population or O4+ cells were observed (Supplemental supernatants (p , 0.0001). Similar results also were observed using Fig. 3A–F). A similar analysis of Ki67+ immunocytochemistry M1 adult microglia supernatants (Table I). Conversely, treatment revealed no differences in cell proliferation among the different with anti-inflammatory M2 supernatants did not have a significant treatment groups (Supplemental Fig. 3G–J). effect on the expression of these growth factors. A summary of In addition to measuring differentiation of human OPCs by mRNA expression of all factors is presented in Table I. immunohistochemical markers, further categorization was con- The Journal of Immunology 767

FIGURE 5. Expression levels of growth factors in human astrocytes following exposure of supernatants derived from polarized immune cell population. (A) Following a 48-h application of polarized T cell supernatants to astrocytes, mRNA expression levels of several molecules (IL-6, LIF, TIMP-1, CXCL10, and IGF-1) were increased significantly. The largest increases were observed following treatment with either Th1 or Th17 supernatants; however, IGF-1 Downloaded from expression was solely influenced by Th2 supernatants. (B) Supernatants generated from polarized human MDMs also significantly influenced the expression of several growth factors (PDGF-A, IL-6, LIF, TIMP-1, CXCL10, and FGF-2). Similar to the proinflammatory Th1 conditions, M1 supernatants also significantly increased the expression of several of these genes. Treatment with M2 supernatants had minimal effect on the expression of these growth factors, with the exception of IL-6. Fold changes are relative to untreated astrocytes (set at 1.0; red dotted line). All data are the mean 6 SEM; n = 5–9/ condition. *p , 0.05, **p , 0.01, ***p , 0.001. http://www.jimmunol.org/ sidered by assessing cell morphology and process extension. O4+ and MBP (Fig. 8A). No significant differences in total numbers of cells were categorized into four stages based on the complexity O1+ cells (Fig. 8B) or MBP+ cells (Fig. 8C) or percentage of of their process formation and branching (Fig. 7A): stage 1, cells MBP+/O1+ cells (Fig. 8D) were observed. with round to bipolar morphology; stage 2, cells with few and short processes; stage 3, cells with ramified, longer, and multiple Discussion processes; and stage 4, cells with membrane sheets and/or lacy In MS, therapeutic strategies are needed that promote tissue repair, + processes. When A2B5 cells were treated with ACM pre-exposed including remyelination by OPCs. Although several first-line and to Th1 supernatants, a significant decrease in stage 3 O4+ cells second-line DMTs in MS are effective at targeting the inflammatory by guest on October 1, 2021 was observed; no differences were measured in the other stages component of the disease, including decreasing the frequency and (Fig. 7B). Cultures treated with ACM pre-exposed to M1 super- severity of relapses, these immunomodulatory therapies are severely + natants had significantly decreased numbers of O4 stage 4 cells, limited in their capacity to be either neuroprotective or stimulate + and numbers of stage 3 O4 cells also were decreased; however, repair. Our study demonstrated that immune cells (particularly these were not statistically significant (Fig. 7C). proinflammatory T cells and myeloid cells) secrete cytotoxic

+ + factors that include, but are not limited to, TNF-a, which increases Mature O1 /MBP human oligodendrocytes are resistant to apoptosis of OPCs, thus negatively influencing differentiation. the negative effects of ACM collected from astrocytes Similarly, astrocytes pre-exposed to supernatants derived from pre-exposed to M1 supernatants proinflammatory immune cell subsets (Th1/Th17/M1) also de- To determine whether the negative effects of ACM collected from creased OPC differentiation; however, this mechanism was not due astrocytes pre-exposed to proinflammatory supernatants also influ- to either increasing cell death or decreasing proliferation. Previous enced differentiation of mature oligodendrocytes, ACM was ap- work demonstrated that secreted products from B cells isolated plied to human PDGFRa-selected cells isolated from the human from MS patients are cytotoxic to oligodendrocytes in vitro; how- fetal brain. After 48 h, cells were fixed and stained via immuno- ever, these results were not mediated by several cytokines analyzed, histochemistry (IHC) for the mature oligodendrocyte markers O1 including TNF-a (22). Additional work using –

Table I. Influences of pro- and anti-inflammatory immune cell supernatants on cytokine and growth factor expression in human astrocytes

M1 M2

Factor Th1 Th17 Th2 MDM HAM MDM HAM TIMP-1 ↑(p = 0.0028) ↑(p = 0.0047) ↑(p = 0.0251) ↑(p = 0.0159) ↑(p = 0.0020) – – LIF ↑(p = 0.0001) ↑(p = 0.0026) ↑(p = 0.0129) ↑(p = 0.0598) – – ↓(p = 0.0029) IL-6 ↑(p = 0.0053) ↑(p = 0.0124) ↑(p = 0.0043) ↑(p = 0.0231) ↑(p = 0.0055) ↑(p = 0.0154) – IGF-1 – – ↑(p = 0.0452) ↓(p , 0.0001) ↓(p , 0.0001) – – CNTF ↓(p = 0.0152) ↓(p = 0.0005) ↓(p , 0.0001) ↓(p = 0.0051) ↓(p , 0.0001) ↓(p = 0.0018) ↓(p = 0.0003) FGF-2 – – – ↑(p = 0.0386) ↑(p = 0.0420) – – PDGF-A ↓(p = 0.0445) – – ↑(p = 0.0177) – ↓(p = 0.0167) – CXCL10 ↑(p = 0.0039) ↑(p = 0.0028) ↑(p = 0.0057) ↑(p = 0.0003) ↑(p = 0.0052) – – –, no change; HAM, human adult microglia. 768 IMMUNE CELL REGULATION OF OLIGODENDROCYTE DIFFERENTIATION Downloaded from http://www.jimmunol.org/

FIGURE 6. ACM collected from astrocytes pre-exposed to proinflammatory immune cell supernatants decrease OPC cell differentiation. (A) After a 48-h exposure to ACM, cells were fixed and stained via IHC for the OPC marker O4. TUNEL, Ki67, and DAPI staining also were performed (Supplemental Fig. 3). The numbers of O4+ cells were counted in each well. Compared with RPMI 1640 and MEM controls, a significant reduction in O4+ cells was noted in the ACM collected from astrocytes pre-exposed to Th1 media (A and B) and M1 human adult microglia (C). This effect was not due to factors such as cell by guest on October 1, 2021 death or proliferation as confirmed by TUNEL and Ki67 labeling (Supplemental Fig. 3). In OPCs pre-exposed for 1 h to a CXCL10 (IP-10) inhibitor, AMG487 (1 mM), the effect of the Th1 ACM was lost, thus implicating CXCL10 as a factor in the supernatants contributing to the decrease in O4+ cells. Scale bar, 20 mm. The bar graphs show the mean 6 SEM; n = 9/condition performed in duplicate. *p , 0.05, **p , 0.01. reactive T lymphocytes demonstrated that these cells increased ol- and diagnoses in other patient populations (28–30). Although the igodendrocyte numbers, with Th2-polarized lines more potent than mechanisms remain unclear, TNF-a has two known receptors Th1-polarized lines (23). In contrast to several reports using rodent (TNFR1 and TNFR2) that have pleiotropic effects in the context models, soluble products derived from immune cells polarized to- of neurodegenerative diseases (31–34). Although TNFR1 can ward an anti-inflammatory (Th2/M2) phenotype do not significantly mediate the negative effects of TNF-a, such as those observed in enhance human OPC differentiation. These results suggest that, SCI and MS, less is known about TNFR2 signaling. It is important within the active MS lesion, the nonpermissive and inflammatory to note that, despite the complicated biology associated with environment restricting the differentiation of neural progenitor cells TNF-a signaling, there are multiple factors within the active MS is due to both infiltrated and resident CNS cells. Furthermore, our lesion that contribute to the dearth of OPC differentiation, such as results also highlight important differences that exist between spe- increased expression of additional proinflammatory cytokines/ cies in the context of immune cells (both lymphocytes and myeloid chemokines, reactive , and the presence of myelin debris. cells) and their roles in stimulating OPC differentiation. The precise role of activated macrophages and microglia within In the current study, we identified TNF-a as a critical factor MS lesions is currently a highly debated topic. Macrophages/ released by activated M1-polarized myeloid cells that decreases microglia can serve as effective APCs and secrete several solu- OPC survival. Indeed, TNF-a was demonstrated to cause degen- ble factors (e.g., TNF-a, IL-12, IL-6) that promote a proin- eration of oligodendrocytes in the MS brain (4, 24). Previous flammatory immune environment. In contrast, myeloid cells also reports suggested that TNF-a inhibits survival and differentiation can secrete neural growth factors (16) and enhance for- of OPCs in a model of injury (SCI) (25), enhances mation (35), and they are the primary cells responsible for the oxidative –induced oligodendrocyte apoptosis (26), and phagocytosis of myelin debris, which is necessary for OPC mi- mediates oligodendrocyte cell death through a TNFR1-dependent gration and differentiation (36, 37). This concept of myeloid cell mechanism (27). Although our data support the previous work plasticity has led researchers to model distinct phenotypes in vitro, implicating TNF-a as a key cytokine that can dramatically influ- which has been termed macrophage/microglia polarization—M1 ence oligodendrocyte biology (Fig. 4), we know that anti–TNF-a (classical, proinflammatory activation) and M2 (alternative, anti- therapies are not a clinically relevant reality in MS because they inflammatory activation) (38). In situ validation of these phenotypic can exacerbate the disease and even initiate MS–like symptoms and functionally distinct myeloid cell populations was demon- The Journal of Immunology 769 Downloaded from http://www.jimmunol.org/

FIGURE 7. ACM collected from astrocytes pre-exposed to proinflammatory immune cell supernatants influences O4+ cell morphology, ramification, and process extension. (A)O4+ cells were categorized into four stages based on the complexity of their process formation and branching. Stage 1, cells with round to bipolar morphology; stage 2, cells with few and short processes; stage 3, cells with ramified, longer, and multiple processes; and stage 4, cells with membrane sheets and/or lacy processes. Scale bar, 200 microns. (B) Cultures treated with ACM pre-exposed to Th1 supernatants had decreased stage 3 O4+ cells compared with stimulated, nonpolarized T cell or Th2 supernatants; no differences were measured in the other stages. (C) Cultures treated with ACM pre-exposed to M1 supernatants had significantly decreased numbers of stage 4 OPCs compared with M2 ACM, and the numbers of stage 3 OPCs also were decreased; however, these differences were not statistically significant. The bar graphs show the mean 6 SEM; n = 6/condition performed in duplicate. *p , 0.05, ***p , 0.001. by guest on October 1, 2021 strated, including within different MS lesions (active versus chronic Previous studies highlighted the shift in the Th1/Th2 balance versus remyelinating) (15, 39), stroke models (40), and SCI (41). as a mechanism by which DMTs can promote a permissive Our previous work on this topic demonstrated that both human environment in which OPCs can differentiate and subsequently macrophages and microglia (fetal and adult) can be successfully remyelinate. Specifically, supernatants collected from human polarized to either the M1 or M2 phenotypes, and they exhibit glatiramer acetate–reactive T lymphocytes and subsequently po- distinct genotypic, phenotypic, and functional features (11, 12, 42). larized to the Th2 phenotype were demonstrated to enhance num- Given the substantial plasticity that may be retained by myeloid bers of both rodent and human OPCs in vitro (23). In the ex- cells, even after activation, we previously validated the polarization perimental autoimmune encephalomyelitis model, human bone properties of our human myeloid cells, including the MDMs and marrow–derived mesenchymal stem cells induced a Th2-polarized microglia used in our current experiments (11, 19). In rodents, response and promoted endogenous repair by increasing the supernatants collected from LPS-activated microglia (M1) were numbers of OPCs in lesions (44). In the current study, we did not shown to attenuate proliferation of OPCs through a TNF-a/IL-6– observe a significant effect of human Th2 supernatants on the dependent mechanism; no effect on OPC survival was observed survival or differentiation of human OPCs. The cytotoxic effects (43). Interestingly, the same supernatants enhanced survival of that we observed using Th1 supernatants were mediated, at least in mature oligodendrocytes. Also in rodent models, M2-conditioned part, via a caspase-dependent mechanism, although they could not media enhanced OPC differentiation in vitro, and M2 polarization be directly attributed to TNF-a because removal of this cytokine was essential for efficient remyelination in a model of focal de- did not reverse the cytotoxicity. The Th1 supernatants likely myelination (16). In our experiments, we exclusively used human contain one or more additional injury mediators. cells and observed that application of both macrophages and In our experiments using human astrocytes, we demonstrated microglia M1-conditioned media decreased OPC differentiation, as that pro- and anti-inflammatory immune cell subsets have opposing reflected by the decrease in O4+ and GalC+ cells (Figs. 2, 3). In OPC effects on astrocyte expression of factors relevant for OPC biology cultures treated with M2-conditioned media (both macrophages and and myelin repair. Application of proinflammatory Th1-, Th17-, microglia), an increase in total O4+ and GalC+ cells was measured; and M1-polarized supernatants increased the expression of several however, the number of OPCs relative to total cell number astrocytic factors, including PDGF-A, TIMP-1, IL-6, FGF-2, (i.e., percent O4+/GalC+) did not differ compared with control CXCL10, and LIF. Similar results were observed using both hu- conditions. These findings help to confirm that the M2-conditioned man and rodent astrocytes following IL-1b activation (45–47). media induced the proliferation of A2B5+ neural progenitors cells These results are of particular significance given that factors se- but not the cells belonging to the OPC lineage (Supplemental Fig. 3). creted from proinflammatory immune cells had a negative impact 770 IMMUNE CELL REGULATION OF OLIGODENDROCYTE DIFFERENTIATION Downloaded from http://www.jimmunol.org/

FIGURE 8. ACM collected from astrocytes pre-exposed to proinflammatory immune cell supernatants does not influence O1+ and MBP+ cell numbers. (A) Following a 48-h application of polarized T cell and myeloid cell supernatants to astrocytes, plates were thoroughly washed, and N1 media were added. Astrocytes were grown in N1 media for an additional 48 h, and the media were collected and stored at 280˚C. The supernatants were then applied to human PDGFRa+ cells isolated from the human fetal brain using immunomagnetic bead selection. After 48 h, cells were fixed and stained via IHC for the mature oligodendrocyte markers O1 and MBP. Scale bar, 200 mm. No significant effects of ACM collected from M1 or M2 supernatant treatment were noted following the counting of O1+ cells (B) or MBP+ cells (C) or on the percentage of MBP+/O1+ cells (D). All bar graphs show the mean 6 SEM; n = 3/ by guest on October 1, 2021 condition performed in duplicate. on OPC differentiation (Fig. 6, 7). However, in line with our cyte-derived factors could mediate this biology. CXCL10 (IP-10) observations, clinical data demonstrated that expression of several is a potent chemoattractant for activated Th1 and NK cells, and it growth factors also is increased within active MS lesions (48–50), can promote neuronal apoptosis through its receptor CXCR3 (54, yet there is still the failure of the OPCs to differentiate. A possible 55). Recent work also demonstrated that CXCL10 is expressed by explanation for this phenomenon is that OPCs cannot overcome astrocytes in MS lesions (56–59) CXCL10 is toxic to rodent OPCs the presence of restrictive factors, such as increased cytokines, (60). Given that the concentration of CXCL10 in our ACM col- chemokines, radicals, and myelin debris and, therefore, lected from astrocytes pre-exposed to inflammatory immune cell fail to differentiate. A second possibility is that individual growth supernatants was within the range of 100–150 ng/ml, we used factors have pleiotropic roles and can exert different effects under a CXCR3 receptor antagonist to block any biological effect that basal and/or inflammatory conditions. For example, PDGF-A and may be mediated by its ligand. Indeed, pre-exposing human OPCs FGF-2 increase proliferation of OPCs, but their presence in the to the CXCR3 antagonist completely blocked the negative effect culture media also can restrict OPC differentiation (51–53), sug- of the Th1-preconditioned astrocytes on OPC differentiation gesting that OPCs are very sensitive and receptive to individual (Fig. 6). This novel finding demonstrates that the protective effects growth factors, depending on their state of maturation. Indeed, our of CXCR3 blockade in rodent cells also can be observed using data support this hypothesis, given that increased expression levels human OPCs. Although macrophages/microglia also secrete of PDGF-A and FGF-2 were measured in astrocytes exposed to CXCL10 upon activation, the levels of this chemokine are sig- proinflammatory myeloid cell supernatants (Table I) and, thus, nificantly lower compared with astrocytes (data not shown) and resulted in fewer differentiated OPCs (Fig. 6). Furthermore, we likely are not responsible for the direct negative effects that were did not observe differences in O4+ cell differentiation until late observed on the OPCs. stage 3–4 (Fig. 7), and our mature oligodendrocytes (O1+/MBP+) In summary, our results suggest that the failure of OPCs to were not sensitive to astrocyte supernatants pre-exposed to proin- differentiate within MS lesions may be mediated by actively se- flammatory supernatants (Fig. 8). This observation is also reflected creted factors that are expressed by both infiltrated proinflam- both in vitro and in situ; immature human OPCs have increased matory immune cell subsets and CNS-resident cells, including susceptibility to TNF-a compared with mature oligodendrocytes microglia and astrocytes. Our findings also indicate that human and are more apoptotic within MS lesions. OPCs may exhibit a greater susceptibility to such factors compared Given that supernatants derived from astrocytes significantly with mature human oligodendrocytes. In this regard, we note that decreased OPC differentiation, we explored how additional astro- our prior array studies comparing genomic profiles of the human The Journal of Immunology 771

OPCs with the adult human oligodendrocytes identified no differences 2013. M2 microglia and macrophages drive oligodendrocyte differentiation during CNS remyelination. Nat. Neurosci. 16: 1211–1218. in the levels of TNFR1, TNFR2, and CXCR3 expression (3), sug- 17. Williams, A., G. Piaton, and C. Lubetzki. 2007. Astrocytes—friends or foes in gesting that the increased susceptibility of OPCs to inflammatory/ multiple sclerosis? Glia 55: 1300–1312. apoptotic stimuli may not be exclusively mediated by either TNF-a 18. Moore, C. S., S. L. Abdullah, A. Brown, A. Arulpragasam, and S. J. Crocker. 2011. How factors secreted from astrocytes impact myelin repair. J. Neurosci. or CXCL10. Although there are distinct cellular mechanisms by Res. 89: 13–21. which these cells can contribute to a nonpermissive environment for 19. Durafourt, B. A., C. S. Moore, M. Blain, and J. P. Antel. 2013. Isolating, cul- the differentiating OPC, it may be possible to selectively target key turing, and polarizing primary human adult and fetal microglia. Methods Mol. Biol. 1041: 199–211. molecules that are responsible for mediating these effects. Given that 20. Wu, C., S. Y. Leong, C. S. Moore, Q. L. Cui, P. Gris, L. P. Bernier, T. A. Johnson, astrocytes secrete a variety of factors that regulate OPC biology, it P. Se´gue´la, T. E. Kennedy, A. Bar-Or, and J. P. Antel. 2013. Dual effects of daily may be possible to direct the myelin repair process through modu- FTY720 on human astrocytes in vitro: relevance for neuroinflammation. J. Neuroinflammation 10: 41. lating astrocytes. Our work suggests that future therapeutics in MS, 21. Livak, K. J., and T. D. Schmittgen. 2001. Analysis of relative gene expression particularly those that aim to improve the remyelination process in data using real-time quantitative PCR and the 2(2Delta Delta C(T)) Method. Methods 25: 402–408. progressive MS, could be complementary in targeting astrocytes, 22. Lisak, R. P., J. A. Benjamins, L. Nedelkoska, J. L. Barger, S. Ragheb, B. Fan, microglia, and infiltrated immune cells. N. Ouamara, T. A. Johnson, S. Rajasekharan, and A. Bar-Or. 2012. Secretory products of multiple sclerosis B cells are cytotoxic to oligodendroglia in vitro. J. Neuroimmunol. 246: 85–95. Acknowledgments 23. Zhang, Y., F. Jalili, N. Ouamara, A. Zameer, G. Cosentino, M. Mayne, We thank Manon Blain, Ellie McCrea, and Farah Jalili for technical assis- L. Hayardeny, J. P. Antel, A. Bar-Or, and G. R. John. 2010. Glatiramer acetate- tance; Dr. Bradford Poulos, director of the Human Fetal Tissue Repository reactive T lymphocytes regulate oligodendrocyte progenitor cell number in vitro: role of IGF-2. J. Neuroimmunol. 227: 71–79. at Albert Einstein College of Medicine, for providing human fetal tissue; 24. Selmaj, K., C. S. Raine, B. Cannella, and C. F. Brosnan. 1991. Identification of Downloaded from and Dr. Samuel Ludwin for reviewing and discussing the manuscript. lymphotoxin and in multiple sclerosis lesions. J. Clin. We also acknowledge Drs. Andre´ Olivier, Jeffery Hall, and Kevin Petrecca Invest. 87: 949–954. for providing adult tissue. 25. Su, Z., Y. Yuan, J. Chen, Y. Zhu, Y. Qiu, F. Zhu, A. Huang, and C. He. 2011. Reactive astrocytes inhibit the survival and differentiation of oligodendrocyte precursor cells by secreted TNF-a. J. Neurotrauma 28: 1089–1100. Disclosures 26. Paintlia, M. K., A. S. Paintlia, A. K. Singh, and I. Singh. 2011. Synergistic activity of -17 and tumor necrosis factor-a enhances oxidative stress- The authors have no financial conflicts of interest. mediated oligodendrocyte apoptosis. J. Neurochem. 116: 508–521. 27. Kim, S., A. J. Steelman, H. Koito, and J. Li. 2011. Astrocytes promote TNF- http://www.jimmunol.org/ mediated to oligodendrocyte precursors. J. Neurochem. 116: 53–66. References 28. Theibich, A., L. Dreyer, M. Magyari, and H. Locht. 2014. Demyelinizing neu- rological disease after treatment with tumor necrosis factor alpha-inhibiting 1. Lassmann, H., and J. van Horssen. 2011. The molecular basis of neuro- agents in a rheumatological outpatient clinic: description of six cases. Clin. degeneration in multiple sclerosis. FEBS Lett. 585: 3715–3723. Rheumatol. 33: 719–723. 2. Oreja-Guevara, C., J. Ramos-Cejudo, L. S. Aroeira, B. Chamorro, and E. Diez- 29. Seror, R., C. Richez, C. Sordet, S. Rist, L. Gossec, G. Direz, E. Houvenagel, Tejedor. 2012. TH1/TH2 Cytokine profile in relapsing-remitting multiple sclerosis J. M. Berthelot, C. Pagnoux, E. Dernis, et al; Club Rhumatismes et Inflammation patients treated with Glatiramer acetate or . BMC Neurol. 12: 95. Section of the SFR. 2013. Pattern of demyelination occurring during anti-TNF-a 3. Cui, Q. L., T. Kuhlmann, V. E. Miron, S. Y. Leong, J. Fang, P. Gris, therapy: a French national survey. Rheumatology (Oxford) 52: 868–874. T. E. Kennedy, G. Almazan, and J. Antel. 2013. Oligodendrocyte progenitor cell 30. Matsumoto, T., I. Nakamura, A. Miura, G. Momoyama, and K. Ito. 2013. New- susceptibility to injury in multiple sclerosis. Am. J. Pathol. 183: 516–525. onset multiple sclerosis associated with treatment in rheumatoid

4. D’Souza, S., K. Alinauskas, E. McCrea, C. Goodyer, and J. P. Antel. 1995. by guest on October 1, 2021 arthritis: a case report and literature review. Clin. Rheumatol. 32: 271–275. Differential susceptibility of human CNS-derived cell populations to TNF- 31. Maier, O., R. Fischer, C. Agresti, and K. Pfizenmaier. 2013. TNF receptor 2 J. Neurosci. dependent and independent immune-mediated injury. 15: 7293– protects oligodendrocyte progenitor cells against oxidative stress. Biochem. 7300. Biophys. Res. Commun. 440: 336–341. 5. Selmaj, K., and C. S. Raine. 1988. Tumor necrosis factor mediates myelin 32. Arnett, H. A., J. Mason, M. Marino, K. Suzuki, G. K. Matsushima, and J. P. Ting. damage in organotypic cultures of . Ann. N. Y. Acad. Sci. 540: 2001. TNF alpha promotes proliferation of oligodendrocyte progenitors and 568–570. remyelination. Nat. Neurosci. 4: 1116–1122. 6. Selmaj, K. W., and C. S. Raine. 1988. Tumor necrosis factor mediates myelin 33. Fischer, R., H. Wajant, R. Kontermann, K. Pfizenmaier, and O. Maier. 2014. and oligodendrocyte damage in vitro. Ann. Neurol. 23: 339–346. Astrocyte-specific activation of TNFR2 promotes oligodendrocyte maturation by 7. Louis, J. C., E. Magal, S. Takayama, and S. Varon. 1993. CNTF protection of secretion of leukemia inhibitory factor. Glia 62: 272–283. oligodendrocytes against natural and tumor necrosis factor-induced death. Sci- 34. Patel, J. R., J. L. Williams, M. M. Muccigrosso, L. Liu, T. Sun, J. B. Rubin, and ence 259: 689–692. R. S. Klein. 2012. Astrocyte TNFR2 is required for CXCL12-mediated regula- € 8. Bennett, J. L., and O. Stuve. 2009. Update on inflammation, , tion of oligodendrocyte progenitor proliferation and differentiation within the and immunoregulation in multiple sclerosis: therapeutic implications. Clin. adult CNS. Acta Neuropathol. 124: 847–860. Neuropharmacol. 32: 121–132. 35. Schafer, D. P., E. K. Lehrman, A. G. Kautzman, R. Koyama, A. R. Mardinly, € 9. Kuhlmann, T., V. Miron, Q. Cui, C. Wegner, J. Antel, and W. Bruck. 2008. R. Yamasaki, R. M. Ransohoff, M. E. Greenberg, B. A. Barres, and B. Stevens. Differentiation block of oligodendroglial progenitor cells as a cause for 2012. Microglia sculpt postnatal neural circuits in an activity and complement- remyelination failure in chronic multiple sclerosis. [Published erratum appears in dependent manner. 74: 691–705. 2009 Brain 132: 1118.] Brain 131: 1749–1758. 36. Plemel, J. R., S. B. Manesh, J. S. Sparling, and W. Tetzlaff. 2013. Myelin inhibits 10. Sica, A., and A. Mantovani. 2012. Macrophage plasticity and polarization: oligodendroglial maturation and regulates oligodendrocytic transcription factor in vivo veritas. J. Clin. Invest. 122: 787–795. expression. Glia 61: 1471–1487. 11. Durafourt, B. A., C. S. Moore, D. A. Zammit, T. A. Johnson, F. Zaguia, 37. Nielsen, H. H., R. Ladeby, C. Fenger, H. Toft-Hansen, A. A. Babcock, T. Owens, M. C. Guiot, A. Bar-Or, and J. P. Antel. 2012. Comparison of polarization and B. Finsen. 2009. Enhanced microglial clearance of myelin debris in T cell- properties of human adult microglia and blood-derived macrophages. Glia 60: infiltrated central nervous system. J. Neuropathol. Exp. Neurol. 68: 845–856. 717–727. 38. Mantovani, A., S. K. Biswas, M. R. Galdiero, A. Sica, and M. Locati. 2013. 12. Moore, C. S., V. T. Rao, B. A. Durafourt, B. J. Bedell, S. K. Ludwin, A. Bar-Or, Macrophage plasticity and polarization in tissue repair and remodelling. J. and J. P. Antel. 2013. miR-155 as a multiple sclerosis-relevant regulator of Pathol. 229: 176–185. myeloid cell polarization. Ann. Neurol. 74: 709–720. 39. Zhang, Z., Z. Y. Zhang, J. Schittenhelm, Y. Wu, R. Meyermann, and 13. Se´guin, R., K. Biernacki, A. Prat, K. Wosik, H. J. Kim, M. Blain, E. McCrea, H. J. Schluesener. 2011. Parenchymal accumulation of CD163+ macrophages/ A. Bar-Or, and J. P. Antel. 2003. Differential effects of Th1 and Th2 lymphocyte microglia in multiple sclerosis . J. Neuroimmunol. 237: 73–79. supernatants on human microglia. Glia 42: 36–45. 40. Hu, X., P. Li, Y. Guo, H. Wang, R. K. Leak, S. Chen, Y. Gao, and J. Chen. 2012. 14. Kim, H. J., I. Ifergan, J. P. Antel, R. Seguin, M. Duddy, Y. Lapierre, F. Jalili, and Microglia/macrophage polarization dynamics reveal novel mechanism of injury A. Bar-Or. 2004. Type 2 monocyte and microglia differentiation mediated by expansion after focal cerebral ischemia. Stroke 43: 3063–3070. glatiramer acetate therapy in patients with multiple sclerosis. J. Immunol. 172: 41. Shechter, R., O. Miller, G. Yovel, N. Rosenzweig, A. London, J. Ruckh, 7144–7153. K. W. Kim, E. Klein, V. Kalchenko, P. Bendel, et al. 2013. Recruitment of 15. Vogel, D. Y., E. J. Vereyken, J. E. Glim, P. D. Heijnen, M. Moeton, P. van der beneficial M2 macrophages to injured spinal cord is orchestrated by remote brain Valk, S. Amor, C. E. Teunissen, J. van Horssen, and C. D. Dijkstra. 2013. choroid . Immunity 38: 555–569. Macrophages in inflammatory multiple sclerosis lesions have an intermediate 42. Butovsky, O., M. P. Jedrychowski, C. S. Moore, R. Cialic, A. J. Lanser, activation status. J. Neuroinflammation 10: 35. G. Gabriely, T. Koeglsperger, B. Dake, P. M. Wu, C. E. Doykan, et al. 2014. 16. Miron, V. E., A. Boyd, J. W. Zhao, T. J. Yuen, J. M. Ruckh, J. L. Shadrach, P. van Identification of a unique TGF-b-dependent molecular and functional signature Wijngaarden, A. J. Wagers, A. Williams, R. J. Franklin, and C. ffrench-Constant. in microglia. Nat. Neurosci. 17: 131–143. 772 IMMUNE CELL REGULATION OF OLIGODENDROCYTE DIFFERENTIATION

43. Taylor, D. L., G. Pirianov, S. Holland, C. J. McGinnity, A. L. Norman, C. Reali, 52. McKinnon, R. D., C. Smith, T. Behar, T. Smith, and M. Dubois-Dalcq. 1993. L. T. Diemel, D. Gveric, D. Yeung, and H. Mehmet. 2010. Attenuation of pro- Distinct effects of bFGF and PDGF on oligodendrocyte progenitor cells. Glia 7: liferation in oligodendrocyte precursor cells by activated microglia. J. Neurosci. 245–254. Res. 88: 1632–1644. 53. Cui, Q. L., G. Fragoso, V. E. Miron, P. J. Darlington, W. E. Mushynski, J. Antel, 44. Bai, L., D. P. Lennon, V. Eaton, K. Maier, A. I. Caplan, S. D. Miller, and and G. Almazan. 2010. Response of human oligodendrocyte progenitors to R. H. Miller. 2009. Human bone marrow-derived mesenchymal stem cells induce growth factors and signals. J. Neuropathol. Exp. Neurol. 69: 930–944. Th2-polarized immune response and promote endogenous repair in animal 54. Sui, Y., R. Potula, N. Dhillon, D. Pinson, S. Li, A. Nath, C. Anderson, Glia models of multiple sclerosis. 57: 1192–1203. J. Turchan, D. Kolson, O. Narayan, and S. Buch. 2004. Neuronal apoptosis is 45. Moore, C. S., R. Milner, A. Nishiyama, R. F. Frausto, D. R. Serwanski, mediated by CXCL10 overexpression in simian human immunodeficiency virus R. R. Pagarigan, J. L. Whitton, R. H. Miller, and S. J. Crocker. 2011. Astrocytic tissue inhibitor of metalloproteinase-1 (TIMP-1) promotes oligodendrocyte . Am. J. Pathol. 164: 1557–1566. differentiation and enhances CNS myelination. J. Neurosci. 31: 6247–6254. 55. Sui, Y., L. Stehno-Bittel, S. Li, R. Loganathan, N. K. Dhillon, D. Pinson, 46. Silberstein, F. C., R. De Simone, G. Levi, and F. Aloisi. 1996. Cytokine- A. Nath, D. Kolson, O. Narayan, and S. Buch. 2006. CXCL10-induced cell death regulated expression of platelet-derived growth factor gene and in cul- in : role of calcium dysregulation. Eur. J. Neurosci. 23: 957–964. tured human astrocytes. J. Neurochem. 66: 1409–1417. 56. Balashov, K. E., J. B. Rottman, H. L. Weiner, and W. W. Hancock. 1999. CCR5 47. Aloisi, F., S. Rosa, U. Testa, P. Bonsi, G. Russo, C. Peschle, and G. Levi. 1994. (+) and CXCR3(+) T cells are increased in multiple sclerosis and their ligands Regulation of leukemia inhibitory factor synthesis in cultured human astrocytes. MIP-1alpha and IP-10 are expressed in demyelinating brain lesions. Proc. Natl. J. Immunol. 152: 5022–5031. Acad. Sci. USA 96: 6873–6878. 48. Su, J. J., M. Osoegawa, T. Matsuoka, M. Minohara, M. Tanaka, T. Ishizu, 57. Simpson, J. E., J. Newcombe, M. L. Cuzner, and M. N. Woodroofe. 2000. Ex- F. Mihara, T. Taniwaki, and J. Kira. 2006. Upregulation of vascular growth pression of the -gamma-inducible chemokines IP-10 and Mig and their factors in multiple sclerosis: correlation with MRI findings. J. Neurol. Sci. 243: receptor, CXCR3, in multiple sclerosis lesions. Neuropathol. Appl. Neurobiol. 21–30. 26: 133–142. 49. Cuzner, M. L., D. Gveric, C. Strand, A. J. Loughlin, L. Paemen, G. Opdenakker, 58. Sørensen, T. L., C. Trebst, P. Kivisa¨kk, K. L. Klaege, A. Majmudar, R. Ravid, and J. Newcombe. 1996. The expression of tissue-type plasminogen activator, H. Lassmann, D. B. Olsen, R. M. Strieter, R. M. Ransohoff, and F. Sellebjerg. matrix metalloproteases and endogenous inhibitors in the central nervous system 2002. Multiple sclerosis: a study of CXCL10 and CXCR3 co-localization in the in multiple sclerosis: comparison of stages in lesion evolution. J. Neuropathol. J. Neuroimmunol. Downloaded from Exp. Neurol. 55: 1194–1204. inflamed central nervous system. 127: 59–68. 59. Tanuma, N., H. Sakuma, A. Sasaki, and Y. Matsumoto. 2006. Chemokine ex- 50. Slaets, H., J. J. Hendriks, C. Van den Haute, F. Coun, V. Baekelandt, P. Stinissen, and N. Hellings. 2010. CNS-targeted LIF expression improves therapeutic effi- pression by astrocytes plays a role in microglia/macrophage activation and cacy and limits autoimmune-mediated demyelination in a model of multiple subsequent neurodegeneration in secondary progressive multiple sclerosis. Acta sclerosis. Mol. Ther. 18: 684–691. Neuropathol. 112: 195–204. 51. McKinnon, R. D., T. Matsui, M. Dubois-Dalcq, and S. A. Aaronson. 1990. FGF 60. Tirotta, E., R. M. Ransohoff, and T. E. Lane. 2011. CXCR2 signaling protects modulates the PDGF-driven pathway of oligodendrocyte development. Neuron oligodendrocyte progenitor cells from IFN-g/CXCL10-mediated apoptosis. Glia 5: 603–614. 59: 1518–1528. http://www.jimmunol.org/ by guest on October 1, 2021