IgG immune complexes break immune tolerance of human microglia

Graphic abstract | IgG immune complexes break immune tolerance of primary human microglia. In healthy individuals, microglia are immune tolerant to viral stimuli. Here, we identified that my - elin structures of the majority (8/11) of MS patients is bound by IgG. Moreover, we identified that the combination of these immune complexes with a viral stimulus breaks the immune tolerance of primary human microglia. Abstract Microglia are phagocytic cells involved in homeostasis of the brain and are key players in the pathogenesis of multiple sclerosis (MS). A hallmark of MS diagnosis is the presence of immunoglobulin G (IgG) antibodies, which appear as oligoclonal bands in the cerebrospinal fluid. Here, we demonstrate that myelin obtained post-mortem from 8 out of 11 MS brain donors is bound by IgG antibodies. Importantly, we show that IgG immune complexes strongly potentiate activation of primary human microglia by breaking their tolerance for microbial stimuli, such as LPS and Poly I:C, resulting in increased production of key pro-inflammatory cytokines, such as TNF and IL-1β. We identified Fc γ RI and Fc γ RIIa as the two main responsible IgG receptors for breaking of immune tolerance of microg - lia. Combined, these data indicate that IgG immune complexes potentiate inflammation by human microglia, which may play an important role in MS-associated inflammation and the formation of demyelinating lesions.

M icroglia are phagocytic cells of the CNS that play an important role in brain homeostasis but also have been implicated in neuroinflammatory diseases (1). In the healthy CNS, microglia are kept in a homeostatic state by their microenvironment, whereas during inflammation, they can be activated to secrete a wide range of cytokines and chemokines (1, 2).
Microglia are known as central players in multiple sclerosis (MS) because they are involved in demyelination and may trigger the adaptive immune response by interacting with infiltrating lymphocytes (3). In active and mixed active/inactive MS lesions, microglia highly express HLA-DR and contain myelin (4). We have previously shown that microglia in normal-appearing MS tissue are in a homeostatic state, identified by RNA sequencing (5). We further demonstrated that microglia isolated from postmortem brain tissue do not respond to common TLR ligands, such as LPS (6,7). This immune tolerance of microglia to microbial stimuli may serve an important physiological purpose by maintaining homeostasis to prevent collateral neuronal damage caused by inflammation. Considered the contribution of proinflammatory microglia to axonal or myelin damage in MS (8,9), there likely is an additional stimulus that converts microglial cells from immune tolerant into proinflammatory cells. Yet, the nature of this additional stimulus is still undefined for microglia in MS.
Oligoclonal bands (OCBs) detected in the cerebrospinal fluid are a diagnostic marker for MS (10,11). Abs isolated from MS serum or cerebrospinal fluid have been identified to specifically target myelin lipids (12) or myelin proteins (13), suggesting the existence of myelin-specific IgG autoantibodies. Interestingly, microglial cells are equipped with various FcgRs that can recognize IgG (6). Therefore, if antimyelin IgG Abs are indeed present in the CNS of MS patients, their binding to myelin and the subsequent formation of large insoluble IgG immune complexes (IgG-ICs) could promote immune activation of microglial cells and phagocytosis of myelin through these FcƴRs.
In the current study, we have taken into account that in the CNS of MS brain donors, microglial cells are likely to be exposed to IgG-ICs, which thereby provides an additional stimulus for microglia activation. We report that myelin obtained from postmortem tissue of the majority of MS brain donors is indeed bound by IgG Abs, whereas low or no IgG binding was observed in the myelin of controls. Moreover, we demonstrate that exposure of human microglia to IgG-ICs breaks their immune tolerance for microbial stimuli such as Poly I:C or LPS, leading to a high expression of proinflammatory cytokines and chemokines such as TNF, IL-1b, IL-8, and IL-12. Finally, we identified FcgRI and FcgRIIa as the two main responsible IgG receptors for this effect. informed consent to perform an autopsy and to use tissue, clinical, and pathological information for research purposes, approved by the medical ethics committee of the Vrije Universiteit Medical Center (Amsterdam, the Netherlands). The donor diagnosis was based on clinical and neuropathological information, which is provided in Supplemental Table I. For microglia isolations, we used tissue from subsequent brain autopsies, regardless of the diagnosis (Supplemental Table I). For myelin isolation, we used tissue from other donors: 9 progressive MS, 1 relapsing-remitting and 1 progressive-relapsing MS brain donors, and 11 nonneurologic control donors were included (Supplemental Table I). MS brain donors had an average disease duration of 29.5 y, and average years until expanded disability status scale (EDSS) score 6 was 15 (Table I). Control brain donors had an average age of 72 y, which is significantly higher than the age of MS donors (63 y). The average postmortem delay (PMD) for MS brain donors was also significantly higher (8.37 h) as compared with control donors (6 h) (Table I).

Myelin isolation
Myelin was isolated from postmortem normal-appearing white matter (WM) (NAWM) tissue of nonneurologic control (n = 11) and MS (n = 11) donors, as previously described by our group. Briefly, after Percoll (GE Healthcare, Little Chalfont, U.K.) density centrifugation during isolation protocol of microglia, the top layer containing myelin was collected and purified by a sucrose gradient (Sigma-Aldrich, St. Louis, MO). After centrifugation, myelin was collected from the interface and washed in water to remove any remaining cells. Purified myelin was labeled with pHrodo dye (1:100; Invitrogen, Carlsbad, CA). Myelin concentration was measured with the Bicinchoninic Acid Protein Assay Kit (Pierce, Thermo Fisher Scientific, Rockford, IL). Detailed donor information and lesion loads for MS donors are provided in Table II and Supplemental Table I.

Microglia isolation
Microglia were isolated from postmortem subcortical WM tissue of brain donors with different clinical backgrounds, as described previously (5,14). Briefly, 6-8 g of tissue was collected during autopsy and stored in Hibernate-A Medium (Invitrogen) at 4˚C. After tissue homogenization, Percoll density centrifugation was performed, and the middle layer containing glial cells was collected, followed by MACS (Miltenyi Biotec, Bergisch Gladbach, Germany) using CD11b beads (Miltenyi Biotec). CD11b-positive cells were collected in beads buffer (PBS with 0.5% BSA and 2 mM EDTA), and viable cells were counted. The purity of microglia isolation was assessed by analyzing CD45

Stimulation
Predestined wells in a 96-wells MaxiSorp plate (Nunc, Thermo Fisher Scientific, Waltham, MA) were coated with 2 mg/ml IgG Abs (Nanogam; Sanquin, Amsterdam, the Netherlands) for 1 h at room temperature (RT), followed by blocking with PBS containing 10% FBS for 0.5 h at 37˚C.
Directly after isolation, microglia (40,000-60,000 cells per well) were divided over wells coated with or without Nanogam, together with or without the TLR ligands LPS (100 ng/ml, from Escherichia coli 0111:B4; Sigma-Aldrich) or Poly I:C (20 mg/ml; Sigma-Aldrich), and cultured in medium (RPMI 1640 [Invitrogen] supplemented with 10% FBS and 1% penicillin-streptomycin) at 37˚C in an incubator. After 3, 6, or 24 h in culture, microglia were lysed for RNA collection in 800 ml TRIsure (Bioline, London, U.K.) and stored at 280˚C. After 24 h, supernatant (200 ml) was collected for ELISA and stored at 220˚C.

RNA isolation
RNA was isolated according to manufacturer's instructions (Bioline). Briefly, chloroform was added to TRIsure samples, and after centrifugation, aqueous phase was collected, mixed with ice-cold isopropanol, and incubated with 1 mg glycogen (Roche, Basel, Switzerland) for 30 min at 220˚C. Precipitated RNA was washed twice in ice-cold 75% ethanol and diluted in 10 ml deionized water.
cDNA synthesis and RT-qPCR cDNA synthesis was performed according to manufacturer's instructions (QuantiTect Reverse Transcription Kit; QIAGEN). Purified RNA was incubated with genomic DNA Wipeout buffer and subsequently incubated with QuantiTect Buffer, RT Primer Mix, and QuantiTect Reverse Transcriptase for 30 min at 42˚C, followed by incubation at 95˚C for 3 min.
Microglial gene expression levels of cytokines, chemokines, and receptors were determined by RT-qPCR. Primer pairs were designed using the Integrated DNA Technologies Web site (https://eu.idtdna.com), and primer specificity was examined using cDNA derived from pooled brain tissue of MS and control donors. Optimal primers were selected based on dissociation curve, and gene expression was normalized to the mean of housekeeping genes GAPDH and elongation factor-1 a. Gene expression values were calculated using the 2 2DD cycle threshold (CT) method. Primers used to determine gene expression are listed in Supplemental Table II.

ELISA
To assess the presence of IgG Abs on myelin, 5 mg/ml myelin isolated from control WM or MS NAWM tissue were coated overnight on a 96-well high-affinity MaxiSorp plate (Nunc). Plates were gently washed using PBS 0.05% Tween and blocked for 1 h at 37˚C with PBS 0.05% Tween containing 1% BSA. After removing the block, plates were incubated for 1 h at RT with polyclonal rabbit anti-mouse IgG HRP (1:1500; DAKO, Jena, Germany) to detect nonspecific binding and with goat anti-human IgG HRP (1:2500; Jackson ImmunoResearch Laboratories, Cambridge, U.K.) to detect IgG present on myelin.
Supernatants of stimulated microglia were analyzed for TNF (eBioscience) and IL-1b (U-CyTech Biosciences, Utrecht, the Netherlands) protein expression, using the indicated Ab pairs.

Myelin phagocytosis
For phagocytosis experiments, microglia (100,000 cells per well) were stimulated with or without Poly I:C in a 96-well MaxiSorp plates (Nunc), coated with or without Nanogam, and cultured at 37˚C. After culturing for 20 h, dead cells and debris were washed away, and adhered microglia were incubated for 24 h with pooled pHrodo-labeled myelin (10 mg/ml) from seven control donors. Information for control donors is provided in Supplemental Table I. After 24-h culturing with myelin, adhering microglia were deattached with TrypLE Select (Thermo Fisher Scientific), incubated for 8 min at 37˚C, and transferred to a 96-well plate. Uptake of pHrodo-labeled myelin by microglia was measured with flow cytometry.

Flow cytometry
Flow cytometry was performed to determine FcgRI, FcgRIIa, and FcgRIII protein expression on isolated WM control microglia. Furthermore, microglial CD45 and CD11b expression was determined for each donor we used in this study. Finally, we used flow cytometry to determine pHrodolabeled myelin uptake by microglia in vitro.
Microglia were incubated for 10 min with FcR Blocking Reagent (Miltenyi Biotec), followed by incubation with conjugated Abs for 30 min on ice and measurement on BD FACSCanto II (BD Biosciences). Viable cells were detected using Viability Dye eFluor 780 (1:1500; eBioscience), and pHrodo-conjugated myelin was detected in the PE channel.

Statistical analysis
Statistical analysis was performed on data obtained from ELISA, RT-qPCR, and flow cytometry measurements. After testing for normality using Shapiro-Wilk normality test, parametric or nonparametric tests were performed to define p values using GraphPad Prism software version 7.03 (GraphPad, La Jolla, CA). Statistical tests performed for each experiment are indicated in the figure legends.

Myelin of MS donors is bound by IgG Abs
The presence of OCBs in the cerebrospinal fluid is one of the key criteria for MS diagnosis (10, 11). It has previously been shown that Abs derived from the serum or cerebrospinal fluid of MS patients specifically target lipids or proteins of MS myelin (12,13). Yet, the majority of these studies have only indirectly demonstrated myelin reactivity by isolating IgG from serum or cerebrospinal fluid, and they have not directly assessed whether IgG is actually bound to the myelin of MS patients. In this study, we investigated whether the myelin of MS brain donors is indeed bound by IgG through an ELISA-based setup, using postmortem myelin isolated from NAWM of MS donors while using nonneurologic donors as controls. Importantly, we found that myelin from the majority of MS brain donors (8/11) is bound by IgG Abs (Fig. 1A), thereby corroborating the idea that antimyelin Abs are present in the CNS and indeed bind to myelin structures. In contrast, almost no IgG binding was observed in myelin that was isolated from nonneurologic controls (Fig. 1A), which is in line with the general absence of OCBs in the cerebrospinal fluid of control individuals (15). MS brain donors had a significantly lower age (62.6 6 11.1 y) and higher PMD (8:37 6 1:21 h) as compared with control donors (age: 72.4 6 10.4 y; PMD: 6:04 6 1:24 h) ( Table I), but when we matched age and PMD for control and MS brain donors, by removing in the control donor group the two eldest donors or two donors with lowest PMD and in the MS donor group the two youngest donors or two donors with highest PMD, the number of donors with IgG-bound myelin was still higher in MS brain donors as compared with control brain donors (data not shown).
Because we did not detect IgG presence on myelin in all MS brain donors (Table II), we determined whether IgG-bound myelin is associated with particular disease characteristics.
Interestingly, MS brain donors that showed IgG bound to myelin displayed a trend toward higher lesion load as compared with MS brain donors without IgG presence on myelin (Fig. 1B). No difference was observed for reactive lesions [which are characterized by the absence of demyelination based on PLP and accumulation of HLA-DR + microglia/ macrophages, defined by immunohistochemistry (16)] in a standard location in the brainstem, disease duration, disease severity (defined as years until patient reached EDSS 6), or proportions of active and mixed active/inactive WM lesions (16) (Fig. 1B). Combined, these data demonstrate that, in contrast to control brain donors, the myelin of the majority of MS brain donors is bound by IgG, which correlates with a trend toward higher lesion load.

IgG-ICs break immune tolerance of human microglia for TLR ligands
Previously, we showed that human microglia are nonresponsive to classic proinflammatory stimuli, which includes TLR ligands such as LPS (7). In addition to testing TLR4 ligand LPS, which enabled the comparison with previous studies, we also used TLR3 ligand Poly I:C as a viral mimic because it has been suggested that viruses can play a role in the pathogenesis of MS (17,18). In addition, both TLR4 and TLR3 are well-known for recognizing nonmicrobial endogenous danger signals that become available during stress, damage, and cell death (19), which are also present under neuropathological conditions, including MS. Because OCBs are a hallmark of MS diagnosis, the presence of IgG Abs that bind to myelin in MS brain donors may act as an additional danger signal, next to microbial/endogenous stimuli, to stimulate primary microglia. To test whether IgG Abs have an effect on the immune activation profile of microglia, we isolated microglia from postmortem WM tissue, regardless of donor clinical diagnosis, and immediately stimulated them with either plate-bound IgG-ICs, TLR ligands Poly I:C or LPS, or a combination. Purity of microglia isolations was defined based on CD45 expression and negative selection for CD15 (marker for granulocytes). The cell samples we collected contained on average 94-98% microglial cells (Supplemental Fig. 2B). Confirming previous results (7), individual stimulation with Poly I:C or LPS induced very little transcription of proinflammatory cytokine and chemokine genes, such as TNF, IL1B, IL6, IL23A, IFNB, IL12A, and IL8 (encoding the proteins TNF, IL-1b, IL-6, IL-23p19, IFN-b, IL-12p35, and IL-8, respectively) ( Fig. 2A, 2B). Similarly, individual stimulation with IgG-ICs also hardly induced any cytokine or chemokine production ( Fig. 2A,  2B). However, strikingly, the combination of Poly I:C and IgG-ICs strongly and synergistically amplified transcription of the proinflammatory cytokines TNF, IL1B, IL23A, IFNB, and IL12A and the chemokine IL8, whereas the production of IL6 was hardly increased ( Fig. 2A). A similar response was observed upon costimulation with LPS and IgG-IC, although in general, the effect was less pronounced and/or the amplification appeared to occur at a later time point (Fig. 2B). To verify that the transcription data correlated with protein secretion, we determined cytokine levels in the supernatant of cultured human microglia, which showed a similar pattern for TNF and IL-1b after costimulation (Fig. 2C, 2D).
In addition to proinflammatory cytokines, we studied the gene expression of additional chemokines (CCL2 and CXCL9), microglia activation markers (ITGAX and SPP1), and T cell costimulatory molecules (CD40, CD86, and CD274) after costimulation (Supplemental Fig. 1). We observed a trend toward increased expression of the chemokine CCL2 and a significantly higher expression of T cell-interacting receptor CD274 upon IgG-IC costimulation. Combined, these data indicate that although microglia are tolerogenic to stimulation with individual TLR ligands, costimulation with IgG-ICs breaks this tolerance by strongly promoting the production of various proinflammatory genes.

IgG-IC stimulation of microglia does not affect myelin uptake
Cross-talk between FcgRs and TLRs has recently been established to be important to potentiate proinflammatory cytokine production by myeloid immune cells (20), but it is less clear whether FcgR-TLR cross-talk also affects other important immune functions, such as phagocytosis. Therefore, we next set out to investigate whether FcgR-TLR cross-talk also has an impact on the general phagocytic capacity of microglia, determined by uptake of nonopsonized control myelin by microglia. In MS, microglia play a central role in demyelination by taking up myelin (4), which may be affected by prior FcgR activation. Therefore, we stimulated microglia with IgG-ICs, Poly I:C, or a combination and subsequently measured the uptake of pHrodo-labeled myelin, isolated from control donors, in lysosomes by flow cytometry (gating strategy is visualized in Supplemental Fig. 2A). We verified that the microglia that were used for the phagocytosis experiments indeed showed IgG-IC-induced immune activation by measuring increased TNF protein production (Fig. 3A). However, whereas stimulation with Poly I:C reduced the uptake of control myelin by microglia, stimulation with plate-bound IgG-ICs had no effect on phagocytosis, neither alone nor in combination with Poly I:C (Fig. 3B, 3C). Please note that the decreased uptake upon Poly I:C stimulation may partly be related to cell viability because the number of viable cells was lower after Poly I:C stimulation (Supplemental Fig. 2A). In conclusion, these data indicate that (co)stimulation of human microglia with IgG-ICs does not affect the uptake of control myelin.

Breaking of tolerance by IgG-ICs is dependent on FcgRI and FcgRIIa
Next, we set out to investigate which receptor on human microglia is responsible for IgG-IC-induced breakdown of tolerance. The main receptors for IgG on human myeloid immune cells belong to the family of FcgRs. As shown in Fig. 4A [and previously by our group (6)], human microglia highly expressed FcgRIa (encoded by FCGR1A), FcgRIIa (FCGR2A), and FcgRIIIa (FCGR3A). Subsequently, we confirmed expression of FcgRI and IIa on protein level using flow cytometry ( Fig. 4B; gating strategy in Supplemental Fig. 2B).
To determine whether FcgRs are responsible for the synergistic inflammatory response upon costimulation with IgG-ICs, we blocked the different FcgRs that are expressed by microglia with specific Abs during (co)stimulation and assessed proinflammatory cytokine gene expression. Blocking of FcgRI and FcgRIIa completely blocked IgG-IC-induced TNF expression, whereas blocking of FcgRIII had no effect (Fig. 4C). Combined, these data indicate that the binding of IgG is essential for the breaking of immunological tolerance, which is mediated by FcgRI and FcgRIIa.

Discussion
Human microglia are generally nonresponsive to microbial stimuli, thereby maintaining brain homeostasis and preventing neuronal damage by inflammation. Yet, considering their contribution to axonal or myelin damage in MS, there likely is an additional stimulus that converts tolerogenic microglia into proinflammatory cells. In this study, we provide evidence that IgG-ICs may act as such an additional stimulus, which breaks microglial immune tolerance for microbial stimuli, leading to increased expression of key proinflammatory cytokines such as TNF and IL-1b. We identified FcgRI and FcgRIIa as the two responsible IgG receptors for this effect. Moreover, we identified that the majority of the MS brain donors displays IgG-bound myelin, which strengthens the concept of antimyelin Ab immune complexes as a relevant secondary stimulus that promotes inflammation in MS brain tissue.
In this study, we identified that myelin in 8 out of 11 MS brain donors is bound by IgG Abs, whereas myelin of most control brain donors showed very little IgG binding. Although this could to some extent be caused by nonspecific binding, the observed IgG binding most likely reflects the presence of Ag-specific interactions of IgG Abs with myelin lipids or/and proteins in MS donors (12,13). Interestingly, a number of studies have shown a worse disease course for MS patients with OCBs present in the cerebrospinal fluid as compared with MS patients without OCBs (11,21,22), thereby corroborating the idea that clonal IgG Abs play an important role in MS disease progression.
Our data suggest that IgG-ICs contribute to MS-associated inflammation by breaking the tolerance of microglia to microbial stimuli. In this study, we used plate-bound IgG as a standardized approach to simulate IgG-ICs. Previously, we and others have  extensively compared plate-bound IgG-ICs to IgG-opsonized bacteria (23), viruses (24), beads (24), fibrinogen-bound Abs (25), or heat-aggregated IgG-ICs (26) and have shown that they all elicit a very similar response by myeloid immune cells. Although these data indicate that plate-bound IgG closely mimics other IgG-ICs, in future studies, it would be valuable to further validate these findings by testing IgG-opsonized MS myelin. Although microglia are generally nonresponsive to stimulation with individual TLR ligands, costimulation with IgG-ICs strongly potentiated the expression of various proinflammatory genes. These include key proinflammatory cytokines and chemokines, such as TNF, IL-1b, IL-8, IL-12, IL-23, and type I IFNs, several of which have been implicated in MS pathogenesis. For example, IL-23 is produced in active MS lesions and is elevated in the serum and cerebrospinal fluid of relapsingremitting MS patients (27,28). In addition, many of the upregulated cytokines and chemokines are involved in the activation of CD8 + T cells, which also play an important role in MS because they are found in higher numbers in active MS lesions (29,30). However, surprisingly, IgG-ICs also strongly amplified the production of IFN-b, which is one of the most commonly used therapies to treat relapsing-remitting MS by reducing the number of active lesions (31,32). Yet, because the majority of the amplified genes strongly promote inflammation, the net response induced by IgG-ICs will most likely promote pathology by activating the local tissue and through recruitment of additional immune cells.
For the stimulation of primary human microglia, we selected two classical TLR ligands, LPS (TLR4) and Poly I:C (TLR3). Although FcgR stimulation amplified the response to both ligands, the effect was most pronounced for Poly I:C. Because TLR4 signals through both MyD88 and TRIF and TLR3 only uses TRIF, these data may indicate that FcgR signaling particularly enables or amplifies the TRIF pathway in human microglia. In addition to recognizing microbial structures, TLR3 and TLR4 also recognize various nonmicrobial danger signals (19), indicating that FcgR-TLR cross-talk may not only occur during infection but also during cell damage and death, as observed under neuropathological conditions including MS. Because Fc receptors have been shown to engage in cross-talk with various different receptor families present on myeloid cells, including RIG-I-like receptors, NOD-like receptors, and C-type lectin receptors (23, 26,33), cross-talk with other receptors is also likely to occur in microglia.
We show that tolerance breakdown by FcgR-TLR cross-talk does not alter the general phagocytic capacity of microglia, as determined by the uptake of nonopsonized control myelin. Yet, FcgRs are very likely to be involved in the enhanced uptake of IgG-opsonized particles such as IgG-bound myelin (34,35). Previous studies indicate that the two different FcgR effector functions (cytokine production and phagocytosis) are controlled by distinct signaling pathways (20,26,36). Nevertheless, when microglia recognize IgG-bound MS myelin, the two FcgR effector functions will likely be activated simultaneously, leading to both phagocytosis and cytokine production. Indeed, we have previously shown that myelin of MS brain donors is phagocytosed more efficiently by microglia than control myelin (37). Notably, in the current study, we used the same control and MS myelin as in the study by Hendrickx et al. (37) to determine the presence of bound IgGs. Therefore, retrospectively, our current finding that MS myelin is IgG bound, whereas control myelin is not, may indicate that the increased uptake of MS myelin in the Hendrickx study was related to IgG opsonization of the myelin, resulting in increased uptake via FcgRs.
Because immune-activated microglia contribute to MS-associated inflammation by secreting proinflammatory mediators and phagocytosing myelin, it would be of high interest to study if microglial tolerance breakdown also occurs in situ in MS brain tissue and could trigger MS lesion initiation, for instance by identifying the presence of viral genes together with myelin-bound IgG-ICs in MS tissue. Activated microglia can appear in normalappearing MS tissue as clusters (38), and future studies should focus on the presence of the two activating stimuli near these microglial clusters because they have been suggested to be the first stage of lesion formation and microglial immune activation might start at these locations.
The molecular mechanisms underlying microglial activation are poorly understood. Our data clearly demonstrate that immune activation upon costimulation with IgG-ICs is regulated at the level of gene transcription. Interestingly, IgG-IC-induced microglial tolerance breakdown is reminiscent of previous studies on other tolerogenic immune cells, in which immune complexes of (auto) antibodies have been shown to break the tolerance of intestinal dendritic cells and synovial "M2" macrophages (33,39). In this regard, microglia seem to more closely mimic synovial macrophages, which also show amplification of gene transcription (33), than intestinal dendritic cells, in which IgG or IgA costimulation specifically amplifies gene translation (39). The cytokine profile of microglia cells upon IgG costimulation partially overlaps with that of the other myeloid immune cells, as illustrated by the increased production of TNF, IL-1b, and IL-23 but also shows microglia-specific responses, such as the strongly increased production of IFN-b, which (in contrast) is strongly suppressed in human macrophages, monocytes, dendritic cells, and Langerhans cells (24). This supports the concept that Fc receptor activation by IgG Abs contributes to the generation of tissue-specific immunity (26).
In addition to modulating gene transcription and gene translation, IgG-ICs have also been shown to activate caspase-1, which is required for the production of IL-1b by cleaving pro-IL-1b into its functional form (23). Interestingly, we observed an increased expression of IL-1b protein (but not mRNA) upon individual stimulation with IgG-ICs, which may suggest that IgG can also activate caspase-1 in human microglia. The signaling molecules involved in IgGinduced gene transcription and/or caspase-1 activation in human microglia will be an important topic for future investigations.
Taken together, these data indicate that the tolerogenic phenotype of primary human microglia is converted into a proinflammatory phenotype upon costimulation with IgG-ICs. Although this could serve a physiological purpose during viral infections of the CNS, the presence of IgG Abs on the myelin of the majority of MS donors suggests that this inflammatory response is activated undesirably in MS patients, thereby promoting chronic inflammation. Interfering with this mechanism may provide new tools to attenuate inflammation in patients suffering from MS.