| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print December 5, 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
- and CD40L-mediated costimulation
Department of Immunohematology and Blood Transfusion, Leiden University Medical Centre, the Netherlands
2Correspondence: Department of IHB, LUMC, Albinusdreef 2, Leiden NL-2333-ZA, the Netherlands. E-mail: t.h.m.ottenhoff{at}lumc.nl
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) comprise a heterogeneous population of cells with various immune and homeostatic functions. Recently, we have described type-1 and type-2 human monocyte-derived M subsets. Although both support outgrowth of intracellular mycobacteria, M-1 secretes interleukin (IL)-23/IL-12 and supports T helper cell type 1 (Th1) responses, whereas M-2 fails to produce IL-23/IL-12, predominantly secretes IL-10, and inhibits Th1 function. Here, we further describe the phenotypic and functional profiles of M-1 and M-2 in response to microbial antigens and interferon-
(IFN-) and CD40L as costimulatory T cell back-talk signals. Activated IL-23+/IL-12+ M-1 secreted IL-1ß, IL-18, IL-6, and tumor necrosis factor-
(TNF-), as well as IL-8, monocyte chemoattractant protein-1 (MCP-1), IFN-inducible protein 10 (IP-10), M inflammatory protein-1ß (MIP-1ß), regulated on activation, normal T expressed and secreted (RANTES), M-derived chemokine (MDC), and (low levels of) pulmonary and activation-regulated chemokine and thymus and activation-regulated chemokine (TARC), corroborating their proinflammatory function. Regardless of the stimulus, M-2 maintained their IL-10+ signature cytokine profile and produced no or relatively low levels of IL-12p40, IL-1ß, IL-6, TNF-, MDC, or TARC. It is remarkable that M-2 secreted high levels of IL-8, MCP-1, IP-10, MIP-1ß, and RANTES, suggesting an active role for these cells in regulating cellular immunity and homeostasis. M-1 and M-2 expressed similar levels of Toll-like receptor and dendritic cell-specific intercellular adhesion molecule-3-grabbing nonintegrin as microbial pattern recognition receptors. M-2, unlike M-1 but like other nonclassical M described previously, expressed CD163 and down-modulated human leukocyte antigen and costimulatory molecules specifically upon activation. These findings demonstrate how M-1/M-2 polarization can differentially skew the host response toward pro- or anti-inflammatory immune responses, respectively. This is likely to be relevant for host-pathogen interactions in chronic bacterial infections and provides a model for dissecting pro- and anti-inflammatory cascades.
Key Words: macrophage polarization • proinflammatory cyto-kines • chemokines • mycobacteria • immune escape
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) are the major targets of infection. In susceptible individuals, Mtb infection can lead to chronic pneumonia with serious lung pathology or cause disseminated disease in the immunocompromised host. In most cases, however, Mtb infection will lead to latent infection, being effectively controlled by the host’s immune response.
Protective immunity against intracellular pathogens requires the induction of type-1 cell-mediated immunity. This is initiated by innate recognition of pathogen-associated molecular patterns by the host’s pattern recognition receptors (PRR), such as Toll-like receptors (TLR), on mononuclear phagocytes. TLR ligation triggers the release of cytokines and chemokines, which activate neutrophils, phagocytes, and lymphocytes to respond to the infection [1
, 2
]. Activated T lymphocytes that are recruited to the site of infection subsequently activate mononuclear phagocytes further and modulate their secretory and effector function through the secretion of interferon- (IFN-) and the expression of CD 40 ligand (CD40L). Optimal protection against Mtb depends on optimal activation of T helper cell type 1 (Th1) cells and the induction of antimycobacterial activity [3
].
Proinflammatory immune mechanisms activated via TLR can be counter-acted by signaling through other PRR. Mononuclear phagocytes can express PRR, which can bind endogenous ("self") ligands, and play an important role in the homeostatic clearance of cellular debris in the absence of inflammation [4
]. Recently, the Mtb cell wall component mannosylated lipoarabinomannan (ManLAM) was identified as a ligand for the C-type lectin dendritic cell-specific intercellular adhesion molecule-3-grabbing nonintegrin (DC-SIGN) [5
, 6
]. Binding of ManLAM to DC-SIGN inhibited lipopolysaccharide (LPS)-induced DC maturation and production of interleukin (IL)-12 and enhanced IL-10 secretion. This illustrates how microbes can hijack anti-inflammatory/suppressive signaling cascades to evade immune recognition [7
]. Mtb is known to be capable of actively interfering with M-associated host defense mechanisms, thereby promoting immune escape and establishment of chronic infection. Mtb has an intrinsic capacity to block the maturation of phagosomes by prohibiting their fusion with lysosomal vesicles. This leads to diminished antigen presentation [8
], decreased innate (e.g., TLR-mediated) immune signaling [9
], and suppressed IFN- receptor (IFN-R)-mediated signaling in mononuclear phagocytes, an otherwise major route to activate bactericidal mechanisms in M [10
, 11
].
Various M subsets have been identified, which are dispersed throughout the body with different functional potential. For example, alveolar M have been shown to be immunosuppressive and display a poor antigen-presenting capacity [12
, 13
]. In support of the heterogeneity and plasticity of the M compartment, in vitro studies have shown that M can be modulated to express different functional patterns. Murine and human M subsets have been identified with a classical (proinflammatory) phenotype or a nonclassical phenotype (for reviews, see refs. [14
, 15
]). Classically, stimulation of M with microbial components and/or IFN- leads to the induction of antimicrobial effector mechanisms and the secretion of proinflammatory cytokines such as tumor necrosis factor (TNF-), IL-1, and IL-12(p40). Nonclassical M subsets, however, are hallmarked by a lack of microbicidal activity and IL-12(p40) production and express IL-10 as the signature cytokine. Despite the recognition of these distinct subsets, the consequences of such differently polarized M interacting with human pathogens, such as mycobacteria, remain unclear so far. The availability of well-defined pro- and anti-inflammatory human M subsets in vitro now allows a comprehensive analysis of their functional profiles and their responses induced by mycobacteria, cytokines, or T cell-mediated back-talk costimuli.
We have recently described human monocyte-derived M-1 and M-2, which strongly differed in their response to (myco)bacterial stimulation [16
]. Although M-1 and M-2 could be infected with mycobacteria, only M-1 (but not M-2) secreted high levels of IL-12p40 monomer and IL-23 (p40/p19) heterodimer upon mycobacterial activation. It is notable that exogenous IFN- was necessary as a costimulus to induce the production of the IL-12 (p40/p35) heterodimer by activated M-1. Whereas M-1 efficiently supported Th1 responses, M-2 displayed poor antigen-presenting capacity and upon activation, predominantly secreted IL-10 and suppressed Th1 function. To further investigate the functional capacity of these polarized human M-1 and M-2, we determined the secretion of a series of relevant cytokines and chemokines, which is associated with classical or alternative M function. In particular, we studied M-1 and M-2 response profiles to Mtb sonicate and LPS in the presence or absence of IFN- or CD40L as mediators of T lymphocyte (back-talk) stimulation and compared these to the response of monocyte-derived DC throughout. In line with their capacity to produce IL-23 and IL-12, M-1 secreted high levels of proinflammatory IL-1ß, IL-18, IL-6, and TNF-. Activated M-2 consistently failed to produce these cytokines and predominantly secreted IL-10, whether or not they were costimulated with IFN- or CD40L. It is notable that M-2 secreted considerable levels of IL-8, monocyte chemoattractant protein-1 (MCP-1), IFN-inducible protein 10 (IP-10), M inflammatory protein-1ß (MIP-1ß), and regulated on activation, normal T expressed and secreted (RANTES), suggesting a potential for these M-2 to recruit and interact with other immune cells. As phenotypical characteristics, we found that M-2, besides activation-induced down-modulation of antigen-presenting and costimulatory molecules, selectively expressed high levels of CD163. Altogether, M-1 appear as mononuclear phagocytes, which are distinct from DC but share a proinflammatory cytokine and chemokine secretion profile with mature DC. M-2, on the contrary, exhibit a suppressive/anti-inflammatory cytokine profile even in the face of T cell-associated, costimulatory signals. Based on their secretory and accessory response, M-2 may play an active role in attracting and suppressing adaptive immune cells upon (myco)bacterial stimulation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
were obtained after 6 days of culturing in medium (RPMI 1640) with 10% fetal calf serum, supplemented with 50 units/ml recombinant human granulocyte M-colony stimulating factor (rhGM-CSF) or with 50 ng/ml rhM-CSF, respectively [16
]. Monocyte-derived DC were obtained as described previously with 1000 units/ml GM-CSF and 500 U/ml IL-4 [17
]. The mouse plastocytoma cell line J558 transfected with murine CD40L (which cross-reacts with human CD40) was used for costimulation experiments [18
]. The tranfected cell line was maintained in RPMI-1640 culture medium with regular supplements and every month for 7–10 days with additional L-histidinol (Sigma-Aldrich, St. Louis, MO) to select for expression of the CD40L-containing vector. J558 (or untransfected control cells) were mixed with monocyte-derived M or DC in stimulation assays in a 1:1 ratio.
Reagents
A sonicate of heat-killed Mtb H37Rv was generously provided by Arend Kolk (Royal Tropical Institute, Amsterdam); cells were stimulated with 10 µg/ml (bacterial dry weight). LPS (Escherichia coli serotype 055:B5) and zymosan A (zym A) were obtained from Sigma-Aldrich and used at 10 ng/ml and 100 ng/ml, respectively, to stimulate cells. IFN- was obtained from Boehringer-Ingelheim (Germany) and used at 500 U/ml.
Flow cytometry
Cell surface marker expression was evaluated by standard flow cytometry using a FACSCalibur cytometer (Becton Dickinson, San Jose, CA) and CellQuest software. Precooled aliquots of 105 cells were incubated with antibody-phycoerythrin (PE) conjugates for 30 min at 0ºC in the dark, washed, and fixed with 1% paraformaldehyde prior to analysis. Appropriate isotype-specific controls were used accordingly. PE conjugates of antibodies against CD14, CD1a, CD83, CD11b, CD11c, CD33, CD54, human leukocyte antigen (HLA)-DR, CD40, CD80, CD86, CD23, CD16, CD32, CD64, and CD163 as well isotype controls were obtained from BD PharMingen (San Diego, CA). HLA class I was detected by using affinity-purified W6.32 in combination with goat-anti-mouse PE as a secondary reagent. PE-conjugated anti-TLR2 and anti-TLR4 were from Hycult (the Netherlands) and anti-IL-18R, from R&D Systems (Minneapolis, MN). Anti-DC-SIGN, clone DC-2, was kindly provided by Theo Geijtenbeek (Free University, Amsterdam).
Enzyme-linked immunosorbent assay (ELISA)
To measure cytokine/chemokine secretion, M and DC were harvested (M after trypsinization), counted, and seeded at 105 cells per 200 µl in triplicates and incubated for 8 or 24 h at 37ºC in the absence or presence of stimulatory agents (as indicated in the legends to the figures and in Results). Supernatants as pools of triplicates were collected and stored at –20ºC prior to use in ELISA. The concentration of secreted cytokines/chemokines was determined by using the cytometric bead assay (BD PharMingen) in case of IL-12(p70), TNF-, IL-6, IL-1ß, IL-10, and IL-8 and using ELISAs from BioSource (Camarillo, CA) for IL-12p40, MIP-1, and RANTES; from R&D Systems, for MCP-1, MIP-1, IP-10, M-derived chemokine (MDC), pulmonary and activation-regulated chemokine (PARC) and thymus and activation-regulated chemokine (TARC); and from Medical and Biological Laboratories (Worburn, MA), for IL-18.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-1 and M-2 were obtained by differentiation of purified human CD14+ monocytes in the presence of GM-CSF or M-CSF, respectively, as described previously [16
]. DC were generated along with M-1 and M-2 for reference purposes. After 6 days of culture, the majority of M-1 consistently had a classical adherent "fried egg" morphology (Fig. 1A
). M-2, on the contrary, primarily appeared as adherent cells with a stretched, spindle-like morphology. Both M subsets expressed CD14 on their cell surface. CD14, however, was absent from monocyte-derived DC, which were yielded as a nonadherent population of cells (Fig. 2A
). Of all cell surface molecules analyzed (see below), CD163, which is a cysteine-rich scavenger receptor family member, was the only distinctive marker expressed by M-2 but not M-1 (or DC; Fig. 2A
).
|
|
-1 and M-2 as pro- and anti-inflammatory M subsets, based on the dominant expression of IL-23/(IL-12p40) or IL-10, respectively, upon stimulation with microbial antigens [16
]. In Figure 1B
, the average levels of IL-12p40 and IL-10 secreted by DC, M-1, and M-2 after stimulation with LPS are depicted. It is notable that M-2 fail to secrete IL-12p40 and produce the highest levels of IL-10. Although the absolute levels of IL-10 differ per donor, without exception throughout our studies, M-2 compared with M-1 from the same donor yielded higher levels of IL-10 than any particular microbial stimulus tested. The average ratio of M-1-derived over M-2-derived IL-10 was 0.30 (95% confidence interval of 0.036–0.57). The IL-10 production by activated M-2 is significantly higher than by activated M-1 (t-test; P=0.0004). M-2 maintain their IL-10+IL-12p40– phenotype after infection with Mycobacterium bovis bacillus Calmette-Guerin [16
].
To further assess the stability of their secretory profiles in the face of major back-talk stimuli of activated T cells, we measured IL-12p40 and IL-10 production upon microbial stimulation in the presence or absence of IFN- or CD40L. Figure 1C
shows that anti-inflammatory M-2 again consistently failed to secrete IL-12p40 upon stimulation with LPS, mycobacterial sonicate, or zym A, with or without IFN- or CD40L as costimuli. Although zym A, like the sonicate of heat-killed mycobacteria, activates via TLR2, LPS is a TLR4 agonist [19
]. Under all these conditions, IL-10 remained the predominant cytokine released by M-2. IL-12p40 secretion by M-1, upon microbial stimulation, was further enhanced by IFN- or CD40L costimulation. CD40L alone, but not IFN- alone, was able to induce expression of IL-12p40 as well. Although activated M-1 secrete substantial amounts of IL-10, in all donors examined, these levels are always lower than those obtained with M-2 (Fig. 1 B
and 1C
; ref. [16
]). Microbial induction of IL-10 production is only weakly modified by costimulation: IFN- shows a mild trend of reduction of IL-10; stimulation with CD40L alone already induces the release of IL-10 from M-1 and M-2. Under any of the conditions tested, M-1 produce higher absolute levels of IL-12p40 than IL-10.
Altogether, these results indicate that polarized M-1 and M-2 are locked in their pro- versus anti-inflammatory response modus, respectively, which is independent of the stimulus applied.
Cell surface marker expression
To investigate possible differences in the expression of several relevant cell surface markers, M-1 and M-2 were left in control medium or activated for 24 h with LPS and stained with a series of phagocyte-marker-specific antibodies and appropriate isotype controls. A summary of the expression levels of these markers is listed in Table 1
. Although absolute expression levels differed among the cells obtained from various donors, the pattern and regulation of expression were consistent throughout.
|
-1 and M-2 were CD1a–CD14+ and expressed CD33, a monocytic cell marker, but lacked the DC-specific activation marker CD83 (Fig. 2A
, Table 1
). CD14 was expressed consistently higher on M-2 than M-1 but not expressed by DC. As mentioned earlier, only M-2 specifically expressed CD163 (Fig. 2A)
, which has previously been associated with nonclassical M phenotypes [20
]. Another distinctive feature between M-1 and M-2 was the expression of IL-18R, which was relatively high on M-1 and down-modulated upon activation but relatively low on M-2 and up-regulated upon activation (Table 1)
.
As the differential response of M-1 and M-2 to microbial stimuli might be the result of the differential expression of PRR, we measured the levels of TLR2, TLR4, and DC-SIGN on both M subsets and DC. Figure 2B
shows that M-1 and M-2 express similar levels of TLR2 and TLR4 and respond similarly to IFN- by up-regulation of both TLR. Expression of DC-SIGN was high on DC as expected. Low but significant levels of DC-SIGN were found on M-1 and M-2 with average MFI of 7.0 (±0.2 SD) and 7.3 (±0.6), respectively (Fig. 2B)
.
LPS-mediated activation of M-2 (for 24 h) resulted in down-modulation of HLA class I HLA-DR and costimulatory molecules CD40 and CD86 on the cell surface (Table 1)
. With the exception of HLA-DR expression, which was slightly down-regulated in some donors, this was not observed for M-1. These findings closely mirror those obtained with mycobacterial stimulation of these cells [16
]. Whereas activated DC showed their characteristic and profound up-regulation of HLA class I, HLA-DR, CD40, CD80, and CD86, as expected, this up-regulation was less prominent (in the case of HLA class I, CD40, CD80, or CD86) or absent (HLA-DR) for M-1. Intercellular adhesion molecule-1/(CD54), the ß2 integrin chain CD11b, and CD11c (gp155/95) were expressed on all cells without major differences between M-1 and M-2. Analysis of the immunoglobulin (Ig)-binding receptors CD23/Fc
R and CD64/FcRI also revealed no clear distinction between M-1 and M-2. CD16/FcRIII and CD32/FcRII, however, were consistently higher on M-2 than M-1.
Taken together, polarized M-1 and M-2 are clearly different from DC and are mutually distinctive as judged by the expression of CD163 and the differential regulation of the expression of HLA and costimulatory molecules IL-18R and FcRII and -III.
Secretion of proinflammatory cytokines
To further characterize the secretory profile of M-1 and M-2 and to compare that to DC, we measured the production of the proinflammatory cytokines IL-1ß, IL-18, IL-6, and TNF-. Cytokine levels were measured in supernatants of cells stimulated or not with mycobacterial lysate or LPS in the absence or presence of IFN- or CD40L. As a control for CD40-mediated stimulation by coculturing with CD40L-tranfected cells, nontransfected control cells had no effect on cytokine secretion in any of the experiments (not shown).
Cytokine levels were measured at 8 h and 24 h, and in general, levels of TNF- were highest after 8 h, and those for IL-1ß, IL-18, and IL-6 were highest after 24 h. Activated M-1 produced considerable levels of IL-1ß, IL-18, IL-6, and TNF- (Fig. 3
). This finding confirms and extends the proinflammatory nature of M-1. In contrast, M-2 displayed no, or a weak, secretory activity for these proinflammatory cytokines (Fig. 3)
. Of note: IL-18 was produced constitutively by M-2, and its production was higher than that observed for DC. M-1 showed highest constitutive levels of IL-18 and uniquely released increased levels of IL-18 upon IFN- costimulation. In comparison with DC, overall M-1 showed the strongest responses to mycobacterial and LPS stimulation with or without IFN-, except for IL-6, which was equally high in M-1 and DC in the presence of IFN-. CD40L-mediated activation alone, however, yielded higher IL-1ß, IL-6, and TNF- levels with DC than with M-1. Together, these data confirm the proinflammatory nature of M-1 and further indicate that IL-10 is the most prominent cytokine produced by activated M-2, irrespective of the mode of stimulation.
|
-1, M-2, and DC. As before, cells were cultured for 8 h and 24 h in the absence or presence of mycobacterial sonicate or LPS alone or with IFN- or CD40L as costimuli. Highest levels of chemokines were consistently obtained at 24 h, except for IL-8, which was secreted at slightly higher levels at 8 h after stimulation. The data depicted represent the highest yields of the respective chemokines (see the figure legends). Neutrophil-attracting IL-8 [CXC chemokine ligand 8 (CXCL8)] was produced at high constitutive levels already by unstimulated M-1 and M-2 but not DC (Fig. 4
). All three cell types yielded increased levels of IL-8 in response to mycobacteria, LPS, or CD40L alone. MCP-1 [CC chemokine ligand 2 (CCL2)] was produced at relatively high constitutive levels by unstimulated M-2 only (Fig. 4)
. In all three cell types, mycobacteria or IFN- alone increased the secretion of MCP-1; LPS enhanced the production in DC and M-1 only. A strong costimulatory effect of IFN- on Mycobacterium-primed MCP-1 was observed in DC only. In contrast, in all cases (except for LPS-activated M-2), (co)stimulation by CD40L inhibited the secretion of MCP-1 by DC, M-1, and M-2 (Fig. 4)
.
|
-1 and M-2, we determined the secreted levels of IFN--IP-10 (CXCL10), MIP-1ß (CCL4), and RANTES (CCL5; Fig. 4
). DC, M-1, and M-2 produced no, or only low levels of, IP-10 when cultured (for 24 h) in medium alone or with mycobacterial sonicate; activation with LPS showed enhanced IP-10 production in all cells, where IP-10 levels were also strongly elevated by the addition of IFN- (as expected) but not by CD40L. MIP-1ß and RANTES secretion was induced (albeit to different levels) by mycobacteria or LPS in DC, M-1, and M-2 and by CD40L stimulation in DC and M-1 (Fig. 4)
. In particular, costimulation by IFN- or by CD40L did not show any profound effects on the production of MIP-1ß or RANTES by M-2. Preliminary data suggest that MIP-1 (CCL3) and MIP-1ß show similar expression patterns (not shown).
Alternatively activated (aa.)M, as the archetype of nonclassical M, have been characterized by high expression of MDC (CCL22), TARC (CCL17), as well as PARC (also known as DC-casein kinase 1 or alternative macrophage activation-associated CC chemokine-1; CCL18 [14
, 21
]). Unlike these aa.M, M-2 failed to secrete high levels of MDC and TARC (Fig. 5
). PARC, conversely, was found at significant levels in the supernatant of resting M-2 or M-2, which were activated for 24 h (Fig. 5)
. M-1 secreted relatively high levels of MDC and lower levels of PARC and TARC when compared with DC, which constitutively produced high levels of all three chemokines (Fig. 5)
. IFN- (co)stimulation showed (modest) down-regulation of MDC secretion by resting or activated M-1 or resting DC, which is in agreement with data in literature [22
].
|
-2, although producing no, or relatively low levels of the, proinflammatory cytokines that we assessed and although producing high levels of IL-10, can also secrete significant levels of neutrophil-, monocyte-, and T lymphocyte-attracting chemokines (IL-8, MCP-1, IP-10, MIP-1ß, and RANTES). This suggests that M-2 have the capacity to actively recruit such cells and in contrast to M-1 and DC, may interact with these cells in an anti-inflammatory/regulatory mode. In contrast to aa.M in literature, the nonclassical M-2 in this study lack a profound secretion of MDC and TARC in particular. Although absolute levels are apparently different, the overall secretory potential of the activated M subsets is determined primarily by their polarization rather than the nature of the stimulus that is used. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-1 and M-2 with pro- versus anti-inflammatory phenotypes, respectively, which can support the intracellular outgrowth of mycobacteria [16
]. These polarized human M, however, show a striking difference in their response to (myco)bacterial stimulation: Whereas activated M-1 predominantly secrete IL-23 and IL-12 and support Th1 activation, activated M-2, by contrast, fail to produce type-1 cytokines, predominantly secrete IL-10, inhibit the response of Th1 cells, and thus, may promote immune evasion of intracellular pathogens. In line with their potential to produce the type-1 cytokines IL-23/IL-12, we now show that activated M-1 also secrete the proinflammatory cytokines IL-1ß, IL-18, TNF-, and IL-6, as well as various chemokines that are known to attract and/or activate various immune cells. M-2, conversely, secrete none of these proinflammatory cytokines (except for a low constitutive level of IL-18), even in the presence of IFN- or CD40L as costimulatory T cell back-talk signals. Despite these strong costimuli, M-2 stably maintain IL-10 secretion as their signature cytokine profile.
It is interesting that M-2 secreted significant and often high levels of various chemokines, which suggests an active role for M-2 in the interaction with other immune cells, possibly in a homeostatic rather than immunostimulatory manner. For example, the secretion of the inflammatory chemokines IP-10, RANTES, MIP-1ß, and MIP-1 suggests that M-2 can recruit effector Th1 cells, which typically express CXC chemokine receptor 3 and CC chemokine receptor 5 as receptors for IP-10 and RANTES/MIP-1ß/1, respectively [23
]. The capacity to produce PARC indicates that M-2 may also recruit memory and naive T cells [24
]. The prominent secretion of IL-8 and MCP-1 likely facilitates their interaction with neutrophils and monocytes, and although not unique to M-2, this may be of relevance for the formation of granulomas in response to mycobacterial infection [25
]. The specific anti-inflammatory response pattern of M-2 as compared with M-1 does not correlate with a differential expression of TLR2/4 and DC-SIGN, which are associated with proinflammatory versus anti-inflammatory signaling in mononuclear phagocytes. Although many receptors may exist that can modulate the innate response, it is tempting to speculate that upon ligation of the same innate receptors, intracellular signaling and subsequent gene expression are differentially regulated between M-1 versus M-2.
Together, our previous observations and the results presented herein suggest that even in the face of mycobacterial stimuli, M-2 fail to support immunogenic/inflammatory responses, as they lack a type-1/proinflammatory cytokine secretion pattern and are able to suppress cellular immunity by the production of IL-10 and the inhibition of antigen presentation. M-2 may attract leukocytes by the secretion of chemokines and are likely to deviate from (protective) type-1 immune effector responses.
Heterogeneity in the activation mode of mononuclear phagocytes is well recognized in vivo and in vitro [14
, 26
]. Originally, so-called "alternative activation" by IL-4 and later, also IL-13, has been found to induce phenotypic and functional changes in mouse and human M, which are distinct from the classical activation phenotype induced by IFN- [27
28
29
30
]. Classical M activation by IFN-, together with microbial stimuli, induces the release of proinflammatory cytokines including IL-12p40 and the expression of microbicidal activity. Instead, aa.M typically inhibits IL-12p40 production and microbicidal function and stimulates the expression of HLA, costimulatory molecules, scavenger receptors such as the mannose receptor and CD163, and the production of IL-10 [14
]. Moreover, aa.M show high and elevated levels of MDC, TARC, and PARC. Expression of these chemokines is specifically potentiated by type-2 cytokines and correspondingly, also found at high (constitutive or inducible) levels in monocyte-derived DC, which are polarized in the presence of IL-4 [21
22
23
24
, 30
]. Despite a large phenotypical overlap, the activated IL-12p40–IL-10+CD163+ M-2 in our study produce no MDC or TARC, secrete only low levels of PARC, and express reduced levels of major histocompatibility complex (MHC) class II and costimulatory molecules, which clearly distinguishes them from IL-4/IL-13-induced aa.M. Another nonclassical M subset (also designated as type-2 M) has been described in the mouse system [15
]. These mouse M have been obtained by classical stimulation in the presence of FcR-ligating IgG complexes and characterized by a high level of IL-10 expression and Th2-supporting activity. Whereas these mouse M also show an abrogated production of IL-12p40, they secrete, in contrast to the human M-2 in this study, relatively high levels of TNF- within 8 h after stimulation [31
]. It is unclear at present whether this is a reflection of differential M polarization or a species-specific response pattern. Our findings corroborate the previous notion by Gordon that various M subsets can be identified, which share a nonclassical IL10+IL-12(p40)– phenotype but yet are distinguishable as separate, nonclassical M subsets by unique response patterns [14
].
Although the present study describes an in vitro model to study human polarized M subsets, it is tempting to speculate that these cells reflect relevant members of the natural M spectrum. It is well established that M can display homeostatic/nonimmune functions, and CD163+ M have been identified as dispersed throughout the body [32
]. Of relevance to airborne infectious pathogens such as Mtb, alveolar M, which are likely the first targets for such pathogens, have been shown to be immunosuppressive and to express CD163 in concordance with the nonclassical M-2, which we describe [20
, 31
]. It is notable that M-CSF, which we have used to differentiate M-2, is a ubiquitous serum protein that may maintain the anti-inflammatory status of M under homeostatic conditions. Although polarized M-2 in this study seem to be locked in an anti-inflammatory mode, recruitement of monocytes by M-2-derived MCP-1 under "danger" conditions such as microbial infection may allow for fresh M to adapt a proinflammatory M-1 phenotype. Thus, the local M milieu could switch from an anti-inflammatory/homeostatic into a proinflammatory/immune mode, allowing the establishment of efficient antimicrobial immunity. Of note, RANTES, MIP-1, and MIP-1ß (which can be released by type-2 M) have been suggested to activate classical M effector mechanisms and type-1 immunity in a murine model of listeriosis [33
]. As the inflammatory response or danger signal wanes, the ubiquitous M-CSF, or possibly also other mechanisms such as neuroendocrine cascades [34
, 35
], may skew M polarization back to the anti-inflammatory phenotype, thereby establishing homeostasis again. Of note, high expression of M-CSF has been found to be associated with conditions of reduced immunity such as in pregnancy [36
, 37
].
It has been noted before that M-CSF and GM-CSF have differential effects on M. Specifically, murine bone marrow cells cultured in M-CSF express reduced levels of MHC class II, are less potent inducers of antigen-specific proliferation, and display enhanced phagocytosis [38
, 39
]. Besides a poor antigen-presenting capacity, M-CSF-derived human M are tolerogenic and support indoleamine 2,3-dehydrogenase-mediated T cell hyporesponsiveness [40
, 41
]. M-2 display a functional profile that fits with a role in homeostatic activity: They express various PRR, fail to produce proinflammatory signals, down-modulate antigen-presenting capacity upon activation, and are equipped to recruit and suppress the activity of lymphocytes. This suggests that M-2 may promote immune evasion of intracellular pathogens such as Mtb from optimal cell-mediated immunity (this study, ref. [16
]). Future studies should explore the possibility of whether M-2 have a potential to induce (pathogen-specific) T lymphocytes with a regulatory capacity. It is interesting that a recent publication has indicated that M-CSF treatment of core blood monocytes establishes tolerogenic antigen-presenting cells, which induce hyporesponsive CD4+ T cells with regulatory potential (Treg) in an alloreactivity model [42
]. In infections, such Tregs may be triggered to prevent inflammation-associated immunopathology at the expense of persistent (latent) infection. Therefore, it will be of interest to explore whether (myco)bacterial persistence is indeed correlated with the presence of nonclassical/anti-inflammatory M phenotypes. The present characterization of human M-1 and M-2 extends our knowledge on functional M heterogeneity and further establishes a powerful human model system to study proimmune versus homeostatic signaling and gene expression in M.
|
ACKNOWLEDGEMENTS |
|---|
|
FOOTNOTES |
|---|
Received January 11, 2005; revised September 2, 2005; accepted September 27, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
transcriptional responses without inhibiting activation of STAT1 J. Immunol. 163,3898-3906 through myeloid differentiation factor 88-dependent and -independent mechanisms J. Immunol. 172,6272-6280 J. Exp. Med. 179,1109-1118 with a Th2-associated expression pattern J. Immunol. 160,1411-1418 on macrophage-derived chemokine production: an amplification circuit of polarized T helper 2 responses Blood 92,2668-2671 receptors J. Immunol. 166,6861-6868, MIP-1ß, RANTES, and ATAC/lymphotactin function together with IFN- as type 1 cytokines Proc. Natl. Acad. Sci. USA 99,6181-6186This article has been cited by other articles:
|
|
 |
|
  A. M. Abdallah, N. D. L. Savage, M. van Zon, L. Wilson, C. M. J. E. Vandenbroucke-Grauls, N. N. van der Wel, T. H. M. Ottenhoff, and W. Bitter The ESX-5 Secretion System of Mycobacterium marinum Modulates the Macrophage Response J. Immunol., November 15, 2008; 181(10): 7166 - 7175. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. A. Kusi, B. A. Gyan, B. Q. Goka, D. Dodoo, G. Obeng-Adjei, M. Troye-Blomberg, B. D. Akanmori, and J. P. Adjimani Levels of Soluble CD163 and Severity of Malaria in Children in Ghana Clin. Vaccine Immunol., September 1, 2008; 15(9): 1456 - 1460. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. D. L. Savage, T. de Boer, K. V. Walburg, S. A. Joosten, K. van Meijgaarden, A. Geluk, and T. H. M. Ottenhoff Human Anti-Inflammatory Macrophages Induce Foxp3+GITR+CD25+ Regulatory T Cells, Which Suppress via Membrane-Bound TGF{beta}-1 J. Immunol., August 1, 2008; 181(3): 2220 - 2226. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. J. Fleetwood, T. Lawrence, J. A. Hamilton, and A. D. Cook Granulocyte-Macrophage Colony-Stimulating Factor (CSF) and Macrophage CSF-Dependent Macrophage Phenotypes Display Differences in Cytokine Profiles and Transcription Factor Activities: Implications for CSF Blockade in Inflammation J. Immunol., April 15, 2007; 178(8): 5245 - 5252. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Pestka and H.-R. Zhou Toll-Like Receptor Priming Sensitizes Macrophages to Proinflammatory Cytokine Gene Induction by Deoxynivalenol and Other Toxicants Toxicol. Sci., August 1, 2006; 92(2): 445 - 455. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||
