Down-regulation of Macrophage Immune Activity by Natural CD 8 + Regulatory T Cells

To evaluate the influence of natural regulatory CD8+ T cells on macrophages we investigated in vitro production of cytokines, reactive oxygen intermediates (ROIs) and expression of CD80 surface costimulatory molecules by macrophages (MF) of wild type (WT) B10PL and syngeneic knock-out (KO) strains, TCRá-/-, â2m-/and CD1d-/mice. MF of TCR á-/(CD4and CD8-) and â2m-/(CD8-) animals produced higher levels of TNF-á, IL-6, IL-12 and ROIs and showed increased expression of CD80 costimulatory molecules in comparison to MF of WT or CD1d-/(NKT-) mice. When MF of these strains were conjugated with TNP hapten and injected i.v. into WT mice to test either induction of contact sensitivity (CS) or tolerance, only TNP-MF of TCRá-/and â2m-/animals induced a significant CS reaction, while cells of WT and CD1d-/-strains were tolerogenic. MF of the tested strains can be classified functionally as resembling either proinflammatory (TCRá-/and â2m-/-mice) or immunosuppressive (WT and CD1d-/-) phenotypes. We suggest the presence of an in vivo regulatory loop in which innate CD8+ Treg cells control the transition between MF phenotypes and thus adjust the magnitude of the inflammatory response to strictly local requirements.

Skin painting with haptens such as trinitrophenyl (TNP) leads in mice to the development of a contact-sensitivity (CS) reaction mediated by Ag/MHC class II restricted Th1 CD4+ cells (ASHERSON & PTAK 1968) or class I restricted CD8 + Tc cells (SAINT-MEZARD et al. 2004).A similar form of CS reaction can be induced by subcutaneous injection of hapten-conjugated peritoneal macrophages (TNP-MF).However, when TNP-MF are administered intravenously (i.v.) they induce antigen-specific CD8+ suppressor T cells (Ts) that not only inactivate immune CD4+ lymphocytes that mediate CS (GREENE et al. 1978;PTAK et al. 1980;SZCZEPANIK et al. 1993) but also suppress the antigen-presenting activity of MF making them unable to induce immunity (PTAK & GERSHON 1982).This results in a long lasting state of tolerance (unresponsiveness).Similar antigen-specific adoptive regulatory T cells were described in several experimental systems (JIANG & CHESS 2004;VLAD et al. 2005).
In the absence of antigenic stimulation the immune system of naive mice is equipped with several populations of down-regulatory cells such as natural CD4+CD25+ FoxP3+ Treg cells and less well characterized natural CD8+ regulatory T cells (JIANG & CHESS 2004;RIFA'i et al. 2004), which inhibit immune responses in a antigen-nonspecific, MHC-non-restricted way and play a key role in the suppression of self-reactive T lymphocytes and in preservation of self-tolerance (RIFA'i et al. 2004).
However, the question of whether natural regulatory CD8+ T cells can influence MF immune activity, as adoptive antigen-induced CD8+ suppressor cells do, is still open (PTAK & GERSHON 1982).Thus, our experiments were aimed to test this possibility.
In this study we describe the outcome of i.v.immunization of B10PL (H-2 u ) wild-type mice with TNP-substituted peritoneal oil-induced MF isolated from syngeneic, wild-type (WT) mice or from their histocompatible deletion mutants (knock-out animals, KO) lacking certain subpopulations of T lymphocytes, or with MF fractions obtained by density gradient centrifugation.Our in vitro studies on cytokines and oxygen radical production and in vivo experiments on immunogenicity show that MF of WT and KO mice lacking CD8+ T cells differ functionally.

Material and Methods
All experiments were approved by the Ethics Committee for Animal Research of Jagiellonian University in Kraków (dated 16.03.2004).

Isolation of peritoneal oil-induced macrophages
Peritoneal macrophages (MF) were collected using 5 mL of DPBS containing 5U heparin/mL from the peritoneal cavity of mice which had been injected i.p. with 1 mL of Marcol 152 mineral oil 4 days earlier.Four to six mice were used per group.Peritoneal exudate cells contained over 90% macrophages (FcãR + and esterase + cells) and were not purified further (PTAK et al. 1980, SZCZEPA-NIK et al. 1993).To test the influence of CD8 + T cells on MF activity, B10PL wild type mice were injected intravenously (i.v.) with 0.25 mg of anti-CD8 mAbs one day before harvest of oil-induced MF.

Discontinuous gradient centrifugation
Macrophages were separated into fractions by centrifugation on a 9% Ficoll gradient as described previously (SZCZEPANIK et al. 1993).0/9% Ficoll interface forms the borderline between immunogenic and tolerogenic MF.Cells recovered from the 0/9% interface and pellet were examined for expression of several important surface markers or were TNP-labeled and injected i.v.into naïve recipients to test the induction of contact hypersensitivity.

Lucigenin-dependent chemiluminescence
Macrophages were resuspended at a concentration of 5 x 10 5 viable cells /well in 0.2 mL RPMI 1640 medium supplemented with 10% FCS in 96well flat bottom dark plates (Nunc, Roskilde, Denmark) and 1 FM lucigenin was added and cells were incubated for 15 min at 37 o C in dark adaptation (GYLLENHAMMER 1987).Then, some MF groups were stimulated by zymosan opsonized with mouse serum (in a ratio of 10 particles per cell) and plates were immediately transferred to a Lucy 1 luminometer (Anthos, Salzburg, Austria) and photon emission was measured for 75 to 100 min.Each experiment was run in duplicate.

Cytokine immunoassay
Macrophages were cultured in 24-flat bottom plates at a concentration of 5 x 10 5 /mL in RPMI 1640 medium supplemented with 5% FCS for 24 or 48h in an atmosphere of 5% CO 2 at 37 o C. Supernatants for TNF-á and IL-6 estimation were collected after 24 h but for IL-10, IL-12 and TGF-â after 48 h and frozen at -80 o C until further use.Concentrations of IL-6, IL-10, IL-12p40, TNF-á, TGF-â, were estimated as described previously in capture ELISA test (BRYNIARSKI et al. 2003(BRYNIARSKI et al. , 2004)).
Induction of contact sensitivity by TNP-derivatized MF and its measurement MF or their fractions were labeled with 10 mM TNBSA in PBS and then washed extensively in DPBS supplemented with 2% FCS (PTAK et al. 1980;SZCZEPANIK et al. 1993).Cell viability exceeded 95%.Hapten conjugated MF (1x10 6 ) in 0.2 mL of PBS were injected i.v.into mice.Seven days later ear thickness was measured with a digital engineer's micrometer (Mitutoyo, Tokyo, Japan) before the application of 10 FL of 0.4% PCL solution in a mixture of acetone and olive oil (1:1) to both sides of the ears.The increment of ear thickness was measured 24 h later and expressed in units of 10 -2 mm ± SD (ASHERSON & PTAK 1968).In each experiment a separate group of nonimmunized controls was challenged simultaneously with hapten and their resulting 24 h ear swelling responses were subtracted from that of the immunized group to obtain the net increase.To test the tolerogenic capacity of TNP-MF, mice which received i.v.cell injection 7 days earlier were skin sensitized by topical application of 150 FL of 5% PCL in an acetone:ethanol mixture (1:3) to the shaved abdomen (ASHERSON & PTAK 1968).Four days later animals were tested for the CS reaction as described above.Mice that did not receive TNP-MF and were only actively immunized with PCL were used as a positive control.

Statistical analysis
All experiments were carried out three to four times.A double-tailed Student's t-test was used to assess the significance of the differences between positive control and the experimental groups, with p<0.05 accepted as a minimum level of significance.

Results
Macrophages of TCRá-/-and â 2 m-/-mice produce higher levels of proinflammatory cytokines and express more CD80 molecules than MF of other strains The data presented in Table 1 shows that MF of TCR á-/and â 2 m -/-mice produce a significantly higher level of proinflammatory cytokines such as TNF-á, IL-6 and IL-12 (p40) and slightly more anti-inflammatory cytokine IL-10 than WT or CD1d-/-MF.Production of inhibitory cytokine TGF-â was similar in all tested groups.
Simultaneously, MF of TCRá-/-and â 2 m-/-mice showed significantly increased expression of CD80 co-stimulatory molecules which play a central role in the induction of the immune response by activation of naive T cells.In TCRá-/-and â 2 m-/-mice almost 70% of MF constitutively expressed CD80 while in WT and CD1d-/-animals the expression of this molecule was 27.4% and 38.8%, respectively (Fig. 5).Other surface molecules such as CD86, CD14 or CD54 were only slightly and insignificantly increased in TCRá-/and â 2 m-/-mice.
Lack of all T lymphocytes or only CD8 + lymphocytes results in higher production of ROIs by peritoneal macrophages Fig. 1 shows that MF of TCRá-/-and â 2 m -/mice stimulated with zymosan produced significantly more ROIs than MF of WT and CD1d-/animals.The data presented in Fig. 2 shows that MF of WT mice depleted of CD8 + cells by prior in vivo treatment with anti-CD8 mAb produced sig-nificantly more ROIs after zymosan treatment than MF of control mice that received the isotype control.Anti-CD8 mAb by itself had no direct influence on MF activity.
MF of TCRá-/-and â 2 m -/-mice induce a CS response while MF of WT and CD1d-/-mice are tolerogenic The data presented in Fig. 3 shows that TNP-MF from TCRá-/-and â 2 m-/-induce a significant CS response after i.v.injection into naïve syngeneic WT recipient mice.In contrast, TNP-MF from WT and CD1d-/-donors did not induce a CS immune response in WT recipients which may result from Table 1 Production of cytokines by MF of B10PL WT mice and their syngeneic deletion mutants (KO).5x10 5 MF in 1 ml RPMI medium supplemented with 5% FCS were cultured at 37 o C in a 5% CO 2 atmosphere.For TNF-á and IL-6 measurements supernatants were collected after 24h culture for other cytokines after 48h.The concentrations of cytokines were measured in a capture ELISA test (see Materials and Methods)   their low immunogenicity or alternatively on their ability to induce tolerance rather than immunity.
To determine which of these two possibilities is true, WT mice were i.v.injected with TNP-MF isolated from different knock out or WT mice and subsequently sensitized with homologous TNP hapten by the highly immunogenic epicutaneous route.Figure 4 shows that the recipient mice which  Macrophages from B10PL WT and â2m-/animals were separated on Ficoll gradient into light (0/9% interface) and heavy (pellet) fractions.Unseparated cells and their fractions were labeled with TNP and 1x10  received CD1d-/-or WT TNP-MF before skin sensitization failed to produce any significant CS reaction upon antigen challenge, indicating that MF of CD1d-/-and WT mice are not only nonimmunogenic but in fact highly tolerogenic.In contrast, animals injected with TCRá-/-or â 2 m -/-TNP-MF prior to epicutaneous hapten sensitization developed a significant CS reaction, although lower than in mice only PCL sensitized (compare groups C, D and E).
Next, using Ficoll gradient centrifugation we compared the Ag-presenting activity of TNPsubstituted light (0/9% Ficoll interface) and heavy (pellet) MF fractions from WT and â2m-/mice with non-separated TNP-MF populations of both strains.Figure 5 shows that light MF fractions of both strains were equally immunogenic while only heavy MF fraction of â2m-/animals upon TNPlabeling induced a significant CS reaction.It is necessary to add that cells in the light fraction of both strains formed only a minority of the total cell gradient load, not exceeding 10 per cent.Immunogenicity of the indicated MF fraction was paralleled by a high percentage of cells expressing the CD80+ marker.

Discussion
Macrophages localized in different tissues and body cavities are heterogenous and differ morphologically and functionally.Their heterogeneity is additionally enhanced by the actual microenvironmental influences such as non-specific inflammatory stimuli ("danger" signals) e.g.PAMPs or cytokines.These signals recognized by cell surface recognition molecules lead to reprogramming of the cell machinery, dependent on the quality of signal, to alternative MF polarization.Polarized macrophages have been classified into two groups, pro-and anti inflammatory (M1 and M2 respectively) although intermediate forms also exist (i.e."mixed" or "overlapping" phenotypes) (MARTINEZ et al. 2009;BENOIT et al. 2008;MOSSER & EDWARDS 2009;MURRAY & WYNN 2011a;MURRAY & WYNN 2011b).
The main goal of our study was to characterize in vitro the possible functional differences between MF isolated from WT and several strains of KO mice, deficient in certain T cell subsets and compare the ability of TNP-substituted MF to induce an immune response or tolerance when given i.v.into naïve syngeneic WT recipients.As MF donors we used TCRá-/-mice missing all Táâ+ cells (CD4+, CD8+ and NKT), â 2 m -/-animals defective in expression of MHC class I antigens and also class Ib, such as CD1d and Qa antigens which results in the lack of CD8+ and NKT cells and finally CD1d-/-mice with absent NKT cells.In all in vivo experiments B10.PL (H-2 u ) wild-type histocompatible animals were used as MF recipients and their MF were used as control antigen presenting cells.
Our experiments show that MF of mice lacking CD8+ lymphocytes: TCRá-/-and â 2 m-/-differed in many aspects from Mf of WT and CD1d-/-animals.They produced higher levels of proinflammatory cytokines (TNF-á, IL-6 and IL-12) (Table 1) as well as reactive oxygen intermediates (ROIs) (Fig. 1) and showed increased expression of CD80 costimulatory molecules.Moreover, MF of TCRá-/-and â 2 m-/-donors conjugated with TNPhapten and injected into WT recipients induced a significant contact sensitivity reaction; under these conditions WT and CD1d-/-MF not only were non-immunogenic but instead induced tolerance (Fig. 3).It is well known that IL-12 plays an important role in the induction of Th1 lymphocytes that mediate the CS reaction.This cytokine upregulates the expression of CD80/86 costimulatory molecules and in cooperation with TNF-á and IL-6 enhances induction of CS (USHIO et al. 1998).
Other studies showed that IL-6 makes T cells refractory to suppression mediated by Treg lymphocytes (PASARE & MEDZHITOW 2003).
Our in vitro and in vivo studies might suggest that MF of the tested mouse strains belong to two different activation phenotypes (MARTINEZ et al. 2009;BENOIT et al. 2008;MOSSER & EDWARDS 2008).However, our further experiments using cells separated by gradient centrifugation indicate that MF of the tested strains were most likely a mixture of different proportions of two MF phenotypes.Macrophages of WT mice contained low numbers of light fraction MF which contain a high proportion of CD80-positive cells able to induce CS, while MF present in the heavy fraction were tolerogenic in WT animals and contained low numbers of CD80 positive cells.In contrast, both gradient fractions of â2m-/-MF induced CS and showed a high proportion of CD80-positive cells.Despite this, as shown in Fig. 3, MF of these two strains still induced a marginal suppression apparent when comparing the magnitude of the CS reaction in groups of mice (C and D) which received TNP -MF and then were skin-sensitized.
The high immunogenicity of MF, apparent in the absence of CD 8 + lymphocytes is functionally determined by their ability to produce a potpourri of proinflammatory cytokines (IL-6, IL-12, TNFá) creating a favourable milieu for the formation of immunological synapse between TCR of responding cells and MHC class II antigen on APC and on the presence of CD80 marker.In contrast, the mechanism of action of natural CD8 + cells on MF activity and on their phenotype is unclear at pres-ent.Whether it depends on reciprocal recognition by receptors on cell membranes or on soluble mediators released by Treg cells can be at present only a matter of speculations.Our former experiments in which contacts of T cells with isolated, metabolically inactive MF cell membranes led to down-regulation of T cell activity may suggest that in a reverse situation surface mechanisms are responsible (PTAK & GERSHON 1975).
In summary, the presence of an in vivo regulatory mechanism by which innate regulatory CD8 + cells supervise the activity of MF may have an important biological sense, providing adjustment of the size and character of MF activity to local requirements.
Naturally present CD8+ lymphocytes described here, controlling the MF phenotype in a nonspecific manner, are only a fraction of the total CD8+ population and are certainly different from antigen-specific CD8+ T cells generated by i.v.administration of hapten-substituted MF and responsible for tolerance (PTAK et al. 1980).Multiple subtypes of naturally occurring regulatory CD8+ T cells with different mechanisms of action and attributed functions have been described by several authors.This seems to be a heterogeneous cell group since some express CD25 and FoxP3 (BIENVENU et al. 2005) or alternatively CD122 (RIFA'i et al. 2004) or CTLA 4 antigens.Some show a low level of CD45RC (XYSTRAKIS et al. 2004) or are either CD28 positive or negative (VLAD et al. 2005(VLAD et al. , 2008)).It was suggested that the target of natural CD8+ regulatory cells may be the APCs (JIANG & CHESS 2004;XYSTRAKIS et al. 2004;BIENVENU et al. 2005;VLAD et al. 2008) but the experimental proof was lacking.Our experiments demonstrate that this indeed might be the case.Future experiments are aimed at establishing the phenotype of natural-regulatory CD8+ T cells that control MF activity and the mode of their action.

Fig. 1 .
Fig. 1.Production of reactive oxygen intermediates (ROIs) by MF of B10PL WT mice and their syngeneic deletion mutants (TCRá-/-, â2m-/and CD1d-/-KO mice).To 5x10 5 MF in 220 microlitres of nutrient medium lucigenin and zymosan were added and photon emission was measured over 75-100 min (abscissa).For details see Materials and Methods.The results of one out of three experiments are demonstrated.Chemiluminescence (ordinate) is expressed in arbitrary units (RULE).

Fig. 2 .
Fig. 2. Production of reactive oxygen intermediates (ROIs) by macrophages of wild type B10PL mice treated in vivo with rat anti-mouse CD8 mAb compared with MF of â 2 m-/-mice.To 5x10 5 Oil-MF in 220 microlitres nutrient medium lucigenin and zymosan were added and photon emission was measured over 75-100 min (abscissa).For details see Materials and Methods.The results of one out of three experiments are demonstrated.Chemiluminescence (ordinate) is expressed in arbitrary units (RULE).

Fig. 3 .Fig. 4 .
Fig.3.Induction of contact sensitivity by TNP-macrophages isolated from B10PL wild-type and knock-out mice.Macrophages of from wild type B10PL mice and KO animals of syngeneic strains: CD1d-/-, TCRá-/-and â 2 m-/-were labeled with TNP (see Materials and Methods) and used as antigen-presenting cells in passive transfer of TNP-CS into naïve recipients.and 1x10 6 viable cells were injected i.v.into B10PL recipients (6 mice per group).(CS reaction was elicited seven days later by TNP-Cl challenge and manifested as ear swelling, measured with a micrometer and expressed in units x10 -2 mm ± SE.) Statistical significance: group A vs groups C and D P<0.01; group B vs groups C and D P<0.01.
and are expressed in pg/ml