doi:10.1093/brain/awn271
Brain (2008), 131, 3256 ^3265
A novel non-transcriptional pathway mediates the proconvulsive effects of interleukin-1b
Silvia Balosso,1, Mattia Maroso,1, Manuel Sanchez-Alavez,2 Teresa Ravizza,1 Angelisa Frasca,1 Tamas Bartfai2 and Annamaria Vezzani1
Laboratory of Experimental Neurology, Department of Neuroscience, Mario Negri Institute for Pharmacological Research, Milan, Italy and 2Molecular and Integrative Neurosciences Department (MIND), The Scripps Research Institute, La Jolla, CA, USA
These authors contributed equally to this work.
Correspondence to: A. Vezzani, PhD, Laboratory of Experimental Neurology, Department of Neuroscience, Mario Negri Institute for Pharmacological Research, Via G. La Masa 19, Milan, Italy E-mail: vezzani@marionegri.it Interleukin-1b (IL-1b) is overproduced in human and rodent epileptogenic tissue and it exacerbates seizures upon brain application in rodents. Moreover, pharmacological prevention of IL-1b endogenous synthesis, or IL-1 receptor blockade, mediates powerful anticonvulsive actions indicating a significant role of this cytokine in ictogenesis. The molecular mechanisms of the proconvulsive actions of IL-1b are not known. We show here that EEG seizures induced by intrahippocampal injection of kainic acid in C57BL6 adult mice were increased by 2-fold on average by pre-exposure to IL-1b and this effect was blocked by 3 -O-methylsphingomyelin (3 -O-MS), a selective inhibitor of the ceramide-producing enzyme sphingomyelinase. C2-ceramide, a cell permeable analog of ceramide, mimicked IL-1b action suggesting that ceramide may be the second messenger of the proconvulsive effect of IL-1b. The seizure exacerbating effects of either IL-1b or C2-ceramide were dependent on activation of the Src family of tyrosine kinases since they were prevented by CGP76030, an inhibitor of this enzyme family. The proconvulsive IL-1b effect was associated with increased T 418 phosphorylation of Src-family of yr kinases indicative of its activation, and T 1472 phosphorylation of one of its substrate, the NR2B subunit of the yr N-methyl-D -aspartate receptor, which were prevented by 3 -O-MS and CGP76030. Finally, the proconvulsive effect of IL-1b was blocked by ifenprodil, a selective NR2B receptor antagonist. These results indicate that the proconvulsive actions of IL-1b depend on the activation of a sphingomyelinase- and Src-family of kinasesdependent pathway in the hippocampus which leads to the phosphorylation of the NR2B subunit, thus highlighting a novel, non-transcriptional mechanism underlying seizure exacerbation in inflammatory conditions. Keywords: experimental epilepsy; glia activation; cytokines; NMDA receptor; inflammation Abbreviations: 3-O-MS = 3-O-methylsphingomyelin; BSA = bovine serum albumin; NMDA = N-methyl-D-aspartate; DMSO = dimethylsulfoxide; ICE = interleukin-1 converting enzyme; icv = intracerebroventricular; IL-1R1 = IL-1 receptor type 1; IL-1b = Interleukin-1b; PBS = phosphate-buffered saline Received June 13, 2008. Revised August 5, 2008. Accepted September 21 2008. Advance Access publication October 24, 2008 ,
Downloaded from brain.oxfordjournals.org by guest on June 7, 2011

Introduction

Prominent inflammatory processes have been described in epileptogenic brain tissue surgically resected from patients with chronic drug-resistant epilepsies (Vezzani and Granata, 2005), including clinical cases which do not feature a typical inflammatory pathophysiology, such as temporal lobe epilepsy and epilepsy-associated with malformations of cortical development (Crespel et al., 2002; Maldonado et al., 2003; Ravizza et al., 2006a; Ravizza et al., 2008). Moreover, brain
injuries associated with CNS inflammation lead to the early occurrence of seizures and can result in epilepsy (Pitkanen and Sutula, 2002; Vezzani and Granata, 2005). These observations, together with the clinical evidence that anti-inflammatory treatments provide seizure control in some cases of drug-resistant epilepsies (Vezzani and Granata, 2005), support the possibility that inflammation in the brain is implicated in the development of seizures. Accordingly, the induction of brain inflammation by

Bock et al., 2003; Salter and Kalia, 2004; Davis et al., 2006b). In particular, the ceramide-producing enzyme neutral sphingomyelinase (N-Smase) is activated in mouse forebrain upon application of IL-1b via the IL-1R1MyD88 complex (Nalivaeva et al., 2000). We tested here whether this pathway, which starts with IL-1b-mediated activation of N-Smase, the production of ceramide and the subsequent phosphorylation of Src-family of tyrosine kinases and the target receptor protein NR2B subunit, mediates the proconvulsive effects of IL-1b. The detailed understanding of the IL-1b-mediated signaling in conditions of elevated cytokine levels is instrumental for understanding the mechanisms underlying seizure exacerbation in inflammatory conditions.

Materials and Methods

Experimental animals
Animals were housed at a constant temperature (23 C) and relative humidity (60%) with free access to food and water and a fixed 12 h light/dark cycle. Procedures involving animals and their care were conducted in conformity with the institutional guidelines that are in compliance with national and international laws and policies.
Cannula and electrodes implantation
Male C57BL6 mice (60-day-old, 25 g, Charles River, Calco, Italy) were surgically implanted with an injection guide cannula and recording electrodes under deep Equithesin anestesia and stereotaxic guidance (Vezzani et al., 2000; Balosso et al., 2005). Two nichrome-insulated bipolar depth electrodes (60 mm OD) were implanted bilaterally into the dorsal hippocampus [from bregma (mm): nose bar 0; anteroposterior 1.9, lateral 1.5 and 1.8 below dura mater]. A 23-gauge cannula was unilaterally positioned on top of the dura mater and glued to one of the depth electrodes for the intrahippocampal infusion of drugs (see below). In some mice, an additional guide cannula was positioned on dura mater ipsilateral to the hippocampal cannula for intracerebroventricular (icv) injection of drugs [from bregma (mm): nose bar 0; anteroposterior 0.0, lateral 1.0 and 3.0 below dura mater]. The electrodes were connected to a multipin socket and, together with the cannula of injection, secured to the skull by acrylic dental cement. The correct position of the electrodes and injection needle was evaluated in the sections used for immunohistochemical analysis of brain tissue (see below).

Pharmacological treatments
Intracerebral injection of drugs in freely moving mice have been previously described (Vezzani et al., 2000; Balosso et al., 2005). Kainic acid (7 ng/0.5 ml; Sigma, Saint Louis, MI, USA) was dissolved in 0.1 M phosphate-buffered saline (PBS, pH 7.4) and injected unilaterally in the dorsal hippocampus by using a needle protruding 1.8 mm from the bottom of the guide cannula. This dose of kainic acid was proven to induce EEG ictal episodes and spiking activity in the hippocampus in 100% of mice without mortality (Balosso et al., 2005). Human recombinant (hr)IL-1b (R&D System, Minneapolis, USA) was dissolved in 0.1 M PBS supplemented with 0.1%

S. Balosso et al.

90 min from their onset. Interictal activity was reckoned by summing up the time spent in EEG spiking.
bovine serum albumin (BSA) and injected intrahippocampally (1 ng/0.5 ml) 10 min before kainic acid. This is the same dose causing proconvulsive effects in kainate-injected rats (Vezzani et al., 1999, 2002). C2-ceramide, a membrane-permeable ceramide analogue, and dihydroceramide, the biologically inactive form of C2-ceramide (Sigma) (Obeid et al., 1993) were dissolved in 10% dimethylsulfoxide (DMSO) in PBS and injected intrahippocampally (0.251.02.0 mg in 0.5 ml), 10 min before kainic acid. 3-O-Methylsphingomyelin (3-O-MS, N-Smase inhibitor; BioMol Research Laboratories Inc., PA, USA) (Zeng et al., 2005; Tsakiri et al., 2008) was dissolved in 10% DMSO in PBS and injected intrahippocampally (3 mg/0.5 ml) or ICV (15 mg/1 ml), 20 min before kainic acid. These doses were shown to block the rapid phase of febrile response to IL-1b after injection into the mouse preoptic area/anterior hypothalamus (pilot study; see Sanchez-Alavez et al., 2006). CGP76030, a selective Src-family of tyrosine kinase inhibitor which binds to the catalytic SH1 domain, thus preventing substrate phosphorylation (Susa et al., 2000; Susa et al., 2005; Rucci et al., 2006) was provided as a research tool by Novartis AG, Basel, Switzerland. It was dissolved in 2% DMSO in PBS and injected intrahippocampally (65 ng/0.5 ml) or icv (130 ng/1 ml), 20 min before kainic acid. The dosage of CGP76030 corresponds to the one achieved in mouse brain after oral administration of 30 mmol/kg, as assessed measuring its brain/plasma ratio at 1 and 3 h, reaching a highest ratio at 3 h. This dose blocked the electrophysiological effects of IL-1b on the activity of warm-sensitive hypothalamic neurons in vivo (pilot study; see also Sanchez-Alavez et al., 2006). Ifenprodil [NR2B-selective NMDA antagonist (Chenard and Menniti, 1999), Sigma] was dissolved in 5% DMSO with 9% Tween 80 in PBS and injected intraperitoneally (1 mg/kg), 15 min before kainic acid. We choose this dose since it was shown to block NR2B receptors in vivo during tPA facilitation of ethanol withdrawal seizures, although not affecting seizures per se (Pawlak et al., 2005). Control mice were injected with the corresponding volume of vehicles before kainic acid. Pharmacological experiments were carried between 9.00 am and 2.00 pm.

Immunocytochemistry

At the end of the EEG analysis, mice injected with kainic acid (n = 5) and their controls (n = 5), were deeply anaesthetized using Equithesin and perfused via ascending aorta with 50 mM cold PBS, pH 7.4 followed by chilled 4% paraformaldehyde in 0.1 M PBS. The brains were post-fixed for 90 min at 4 C, and then transferred to 20% sucrose in PBS for 24 h at 4 C. The brains were rapidly frozen in 50 C isopentane for 3 min and stored at 80 C until assayed. Serial cryostat coronal sections (40 mm) were cut from all brains throughout the septo-temporal extension of the hippocampus (Franklin and Paxinos, 1997) and collected in 0.1 M PBS. Primary and secondary antibodies and experimental procedures were chosen to determine a specific immunohistochemical signal of the protein of interest in rodent brain slices (Ravizza and Vezzani, 2006; Ravizza et al., 2008). IL-1. Slices were incubated at 4 C for 10 min in 70% methanol and 2% H2O2 in TrisHCl-buffered saline (TBS, pH 7.4) followed by 30 min incubation in 10% foetal calf serum (FCS) in 1% Triton X-100 in TBS. Then, slices were incubated overnight with the primary antibody against IL-1b (1:200, Santa Cruz Bio., CA, USA) at 4 C in 10% FCS in 1% Triton X-100 in TBS. Immunoreactivity was tested by the avidin-biotin-peroxidase technique (Vector Labs, USA), the sections were reacted using diaminobenzydine (DAB) and the signal was amplified by nickel ammonium. IL-1R1. IL-1R1 immunostaining was carried out as previously described (Ravizza and Vezzani, 2006). Briefly, sections were incubated at 4 C for 10 min in 0.3% H2O2 in 0.3% Triton X-100 in PBS. After three 5 min washes in 0.3% Triton X-100 in PBS, slices were incubated at 4 C for 60 min in 10% FCS in 0.3% Triton X-100 in PBS. Then, slices were incubated with the primary antibody against IL-1R1 (6 mg/ml, R&D System) for 72 h at 4 C in 4% FCS in 0.3% Triton X-100 in PBS. Immunorectivity was tested as described for IL-1b.

Double-immunostaining

Two brain slices in each mouse brain for each cell type marker were randomly chosen to identify the cells expressing IL-1b or IL-1R1. After incubation with the primary IL-1b and IL-1R1 antibodies, slices were incubated in biotinylated secondary anti-goat and anti-hamster antibodies respectively (1:200, Vector Labs), then in streptavidinHRP and the signal was revealed with tyramide conjugated to Fluorescein using TSA amplification kit (NEN Life Science Products, Boston, MA, USA). Sections were subsequently incubated with the following primary antibodies: mouse anti-glial fibrillary acidic protein (1:2500; GFAP, Chemicon, Temecula, CA, USA), a selective marker of astrocytes, or rat anti-mouse CD11b (1:1000; MAC-1, Serotec, Oxford, UK), a marker of microglia-like cells, or mouse anti-neuronal specific nuclear protein (1:1000; NeuN, Chemicon), a selective neuronal marker. Fluorescence was detected using anti-mouse or anti-rat secondary antibody conjugated with Alexa546 (Molecular Probes, Leiden, The Netherlands). Slide-mounted sections were examined with an Olympus Fluorview laser scanning confocal microscope (microscope BX61 and confocal system FV500; Hamburg, Germany) using dual excitation of 488 nm (Laser Ar) and 546 nm (Laser HeNe green) for Fluorescein and Alexa546, respectively. The emission of fluorescent probes was collected on separate detectors.

Seizures assessment and quantification
EEG seizures induced by intrahippocampal injection of kainic acid in mice have been extensively described (Balosso et al., 2005). Briefly, a 30 min recording was done before kainic acid injection to assess the basal EEG pattern and for 180 min thereafter. At least a 30 min recording similar to baseline was required before ending the experiment. Ictal episodes are characterized by high-frequency and/or multispike complexes and/or high-voltage synchronized spikes simultaneously occurring in the injected and contralateral hippocampi. Spiking activity is typically observed between seizures and after seizures subside. The EEG recording of each animal was analysed visually by two independent investigators unaware of the treatments to detect any activity different from baseline. Seizure activity was quantified by reckoning the time elapsed from kainic acid injection to the occurrence of the first EEG seizure (onset) and the total number and total duration of seizures (reckoned by summing up the duration of every ictal episode during the EEG recording period). Seizures occurred with an average latency of about 10 min from kainic acid injection, then recurred for about
Mechanism of IL-1b proconvulsive effect
To eliminate the possibility of bleed-through between channels, the sections were scanned in a sequential mode.
Evaluation of neuronal damage
Neuronal loss was evaluated using Nissl-stained slices as previously described (Hoffman et al., 2003). A subset of mice treated with kainic acid, or ceramide + kainate, or IL-1b + kainate and their respective vehicle- injected controls, were killed one week after treatment by transcardial perfusion as described above (n = 5 mice in each experimental group). Serial cryostat sections (40 mm) were cut coronally from plate 41 to 50 according to Franklin and Paxinos (1997), collected on slides and Nissl stained. Two sections in each brain were analysed, namely at 1.46 mm (corresponding to plate 43) and 2.06 mm (corresponding to plate 48) from bregma. In each section, neuronal cell loss was quantified by measuring the area occupied by Nissl-stained neurons in CA3 pyramidal cell layers (Hoffman et al., 2003). An image of the whole hippocampus was captured with a 4 objective using an Olympus light microscope (BX61; Hamburg, Germany) connected to a ColorView digital camera (Olympus). Using an AnalySIS Image Processing software (Soft Imaging System, Olympus), we divided the areas of interest in non-overlapping fields (each of 700 mm width 400 mm high) using a 20 magnification, and in these fields we measured the area (in mm2) occupied by Nissl-stained neurons. By summing up these values, we determined the total area in CA3 occupied by Nissl-stained neurons which reflects neuronal density in each region. The quantification of neuronal loss was performed by two investigators blind to the identification code of the samples.

was done to quantify the changes in protein levels (AIS image analyzer, Imaging Research Inc., Ontario, Canada) using film exposures with maximal signals below the photographic saturation point. Optical density values in each sample were normalized using the corresponding amount of b-actin.
Statistical analysis of data
Data are the mean SEM (n = number of individual samples). The effects of treatments were analysed by two-way ANOVA followed by Tukey or KruskalWallis test, or by Student t-test. Data reporting quantification of cell loss or depicted in Figs 2 and 3 are expressed as % of vehicle-injected mice; however, statistical analysis was carried out using absolute values.
Results Seizure-mediated induction of IL-1b and IL-1R1 in the mouse hippocampus
Immunohistochemical analysis of hippocampal sections 3 h after intrahippocampal injection of kainic acid, showed increased IL-1b immunoreactivity in astrocytes (Fig. 1B) while IL-1R1 was enhanced both in neurons and astrocytes (Fig. 1D); this activation involved the dorsal and temporal poles of the hippocampus bilaterally. This is the first evidence of increased IL-1b and IL-1R1 in mice after kainate-induced seizures which is in accordance with our previous findings showing induction of IL-1b and IL-1R1 by chemically- or electrically-induced seizures in rats (Vezzani et al., 1999, Ravizza and Vezzani, 2006; Ravizza et al., 2008) or bicuculline-induced seizures in mice (Vezzani et al., 2000). IL-1b and

Western blot

Different groups of mice (n = 1014 in each group) were implanted with electrodes and cannula as described before and injected with hrIL-1b, kainic acid, or their combination 3-OMS or CGP76030 (see before for injection protocol). Control mice (n = 14) were implanted with electrodes and cannula and injected with the corresponding vehicles of the various treatments at the appropriate time interval. One hour after the onset of EEG seizures, experimental mice and their controls were decapitated. The injected dorsal hippocampus was dissected out at 4 C and two hippocampi obtained from different mice within the same experimental group were pooled and homogenized in 20 mM TrisHCl buffer (pH 7.4) containing 1 mM EDTA, 5 mM EGTA, 1 mM Na-vanadate, 2 mg/ml aprotinin, 1 mg/ml pepstatin, 2 mg/ml leupeptin (30 mg tissue/150 ml homogenization buffer), thus obtaining five to seven individual samples in the various treatment groups. Total proteins (70 mg per lane; Bio-Rad Protein Assay) were separated using SDS-PAGE, 10% acrylamide and each sample was run in duplicate. Proteins were transferred to Hybond nitrocellulose membranes by electroblotting. For immunoblotting, we used antipTyr418-Src which is located in the catalytic domain therefore indicative of Src-family of tyrosine kinases activation (1:750; Sigma) or anti-pTyr1472-NR2B (1:1000; Affinity Bioreagents Golden, CO, USA) rabbit polyclonal antibodies. Total NR2B levels were assessed using goat polyclonal anti-NR2B antibody (1:1000; Santa Cruz). Immunoreactivity was visualized with enhanced chemiluminescence (ECL, Amersham, UK) using peroxydase-conjugated goat antirabbit (1:2000; Sigma) or rabbit anti-goat (1:10 000; Sigma) IgGs as secondary antibodies. Densitometric analysis of immunoblots

Fig. 1 IL-1b and IL-1R1 expression in the mouse hippocampus after kainic acid-induced seizures. Representative photomicrographs of IL-1b (A^B) and IL-1R1 (C^D) immunoreactivity in the CA3 area of the hippocampus, 3 h after seizures induced by intrahippocampal injection of kainic acid (7 ng/0.5 ml; B and D) and in vehicle-injected C57BL6 mice (A, C). IL-1b (A) and IL-1R1 (C) immunostaining was not detectable in control hippocampus. After seizures, IL-1bimmunoreactivity is strongly enhanced in GFAP-positive astrocytes (B, yellow signal in inset); IL-1R1staining was enhanced both in neurons (D, yellow signal in d1) and astrocytes (D, yellow signal in d2). Scale bar: A^D100 mm; insets, 25 mm.
S. Balosso et al. seizures was observed using 0.25 mg C2-ceramide, 1 mg significantly increased seizures duration by $2-fold (P50.05) while 2 mg significantly increased the number of seizures by 2-fold and their duration by 2.5-fold on average (P 5 0.01) versus mice injected with dihydroceramide, a membraneimpermeable ceramide analog (Fig. 2A; see Table 2). Thus, C2-ceramide mimics IL-1b proconvulsive effects inducing also a similar EEG pattern of seizures (Fig. 2B). IL-1b or C2-ceramide alone did not induce seizures. Either C2-ceramide or IL-1b did not affect the pattern or extent of neuronal cell loss induced by kainate in mice, which consists of degeneration of CA3 pyramidal neurons in the injected hippocampus as assessed by Nissl staining in a subset of mice (n = 5 each group) killed 7 days after seizure induction [1.46 mm from bregma; data are expressed as percentage of vehicle (n = 5); vehicle, 100 3.7, kainate, 86.2 5.4; ceramide+kainate, 81.1 8.1; IL-1b+kainate, 81.5 7.0 (P50.05 for each treatment groups versus vehicle); 2.06 mm from bregma; vehicle, 100 2.0, kainate, 67.1 9.7; ceramide+kainate, 69.5 9.0; IL-1b+kainate, 63.3 5.4 (P50.05 for each treatment groups versus vehicle)].
IL-1R1 (Fig. 1A and C) were not detectable in control hippocampus, as previously shown (Vezzani et al., 1999, 2000; Ravizza and Vezzani, 2006; Ravizza et al., 2008).
N-Smase-Src kinase-NR2B pathway mediates the proconvulsive activity of IL-1b

3-O-MS

We established the role of N-Smase in the proconvulsive effect of IL-1b using the selective N-Smase inhibitor 3-O-MS (Zeng et al., 2005; Tsakiri et al., 2008) (Table 1). The intrahippocampal injection of hrIL-1b, 10 min before kainate, increased by $1.8-fold the number of seizures and by $2-fold their duration as compared with vehicle-treated mice (P50.01; Table 1). Three micrograms of 3-O-MS blocked the proconvulsive effect of hrIL-1b when injected into the hippocampus, 10 min before the cytokine (i.e. 20 min before kainate) since it abolished the increase in the number of seizures [F(1,62) = 2.2; P50.05] and time spent in ictal activity [F(1,62) = 5.9; P50.01 by two-way ANOVA) induced by hrIL-1b. The latency to the first seizure or the time spent in interictal activity were not affected by these treatments. 3-O-MS injected intrahippocampally (3 mg/0.5 ml) (Table 1) or icv (15 mg/1 ml) (data not shown) 20 min before kainic acid, did not affect seizure parameters; higher doses could not be used because of the limit of solubility of this compound in the volume of injection.

Fig. 2 Effect of C2-ceramide on seizures. (A) Bargrams represent the mean SE (n = 12).Dihydroceramide, the inactive analogue of ceramide, (2 mg/0.5 ml) or C2-ceramide (the cell-permeable analogue of ceramide; 0.25^1.0 ^2.0 mg/0.5 ml) were injected intrahippocampally,10 min before kainic acid. Significant increases in seizure parameters were observed at1and 2 mg C2-ceramide. P50.05; P50.01versus dihydroceramide by one way ANOVA followed byTukey test. (B) Representative EEG tracings of freely moving C57BL6 mice injected unilaterally in the hippocampus with kainic acid IL-1b (1ng/0.5 ml) or C2-ceramide (2 mg/0.5 ml).Treatments or vehicles were given10 min before kainic acid. (a) Baseline recording before kainic acid injection; arrowheads in (b) and (c) include representative ictal episodes recorded in the EEG during 90 min after kainic acid injection IL-1bor C2-ceramide; tracings in (d) depict spiking activity in the EEG after termination of seizures. Either IL-1bor C2-ceramide alone did not induce seizures. RHP and LHP are right and left (injected) hippocampus, respectively.
T able 2 Effect of inhibition of Src-family of tyrosine kinases on the proconvulsive effects of C2-ceramide
Dose Dihydroceramide C2-Ceramide CGP076030 CGP076030 + C2-Ceramide 2 mg 2 mg 65 ng Onset (min) 8.6 0.7 8.9 1.4 8.1 1.0 8.4 1.6 Number of seizures 8.0 1.0 18.0 2.0 10.0 2.0 8.0 1.0yy Time in ictal activity (min) 5.4 0.5 13.4 1.0 6.3 1.1 5.5 0.4yy Time in spikes (min) 44.0 3.4 51.1 3.7 53.3 7.5 45.0 3.3
Data are the mean SE (n = 7^9 mice in each experimental group). C57BL6 mice were injected intrahippocampally with C2-ceramide (2 mg/0.5 ml) or CGP076030 (65 ng/0.5 ml), or their combination before intrahippocampal kainic acid injection (7 ng/0.5 ml). P 5 0.01 versus the inactive analog dihydroceramide; yyP50.01 versus C2-ceramide by two-way ANOVA followed by Tukeys test [F(DF) for each treatment group are reported in the Results section].
seizures were decreased by $1.8- and 2-fold, respectively, as compared with vehicle-treated mice [number of seizures: vehicle, n = 8, 16.0 2.0; CGP, n = 8, 9.0 1; P50.01; time in ictal activity (min): vehicle, 9.9 0.5; CGP, 4.8 0.4; P50.01 by Students t-test].
Effect of pharmacological treatments on Src-family of tyrosine kinases activation and NR2B phosphorylation
We assessed by western blot analysis of hippocampal homogenates, the level of tyrosine phosphorylated (p) forms of Src-family of kinases and the NR2B subunit of NMDA receptor, 60 min after seizures onset in kainate IL-1b injected mice, as well as in mice receiving IL-1b alone (Fig. 3). Densitometric analysis of the specific protein bands showed that either hrIL-1b alone (no seizures) or seizures per se increased phosphorylation (p) of Src-family of kinases on Tyr418 in the catalytic domain denoting their activation (Papp et al., 2008), by 40% on average as compared to vehicle-treated mice (P50.05 and P50.01, respectevely) while in the same hippocampi the Tyr1472 phosphorylation

Ifenprodil

Ifenprodil (1 mg/kg), a NR2B-specific NMDA antagonist (Chenard and Menniti, 1999), injected 5 min before hrIL-1b blocked the cytokine-mediated increase in the number [F(1,61) = 2.2; P50.05] and total time in seizures [F(1,61) = 9.0; P50.01 versus hrIL-1b by two-way ANOVA] (Table 1). This dose of ifenprodil did not affect seizures per se (Pawlak et al., 2005), although higher doses are known to provide anticonvulsive effects (Kohl and Dannhardt, 2001; Yen et al., 2004).
S. Balosso et al. by reducing (p)NR2B to the levels of vehicle-injected mice [3-O-MS: F(1,17) = 21.9; P50.01; CGP76030: F(1,17) = 10.7; P50.01]. In IL-1b+kainate-treated mice, inhibition of N-Smase by 3-O-MS similarly reversed Src-family of kinases phosphorylation [F(1,17) = 30.5; P50.01], which was not affected by CGP76030 treatment. 3-O-MS or CGP76030 alone did not change the basal level of (p)Src and (p)NR2B. The total levels of NR2B were not changed by the various treatments (not shown).

Discussion

This study shows a novel IL-1b-IL-1R1 activated cell signaling that mediates the proconvulsive actions of IL-1b in the mouse hippocampus. IL-1b is a pluripotent proinflammatory cytokine which binds to IL-1R1, a Toll receptor family member, and induces via an NFkB-dependent mechanism, the transcription of various genes encoding several downstream mediators of inflammation, including IL-6 or TNF-a, inducible NO and COX-2 (Dinarello, 2004); the time scale of these effects is included between 30 and 90 min. The activation of this Toll-like pathway requires transcription and translation of the encoded proteins with their subsequent release, therefore it is unlikely to account for the recently described rapid effects of IL-1b on NMDAdependent neuronal Ca2+ influx and seizure activity (Vezzani et al., 1999, 2000; Viviani et al., 2003; Ravizza et al., 2006b). This consideration raises the hypothesis that the proconvulsant action of IL-1b involves an additional, rapid, non-transcriptional intracellular neuronal pathway leading to fast changes in ion channels. Previous reports have shown neuronal effects of IL-1b in hippocampal and hypothalamic neurons involving synaptic plasticity and thermoregulation which also appear to be independent of transcriptional events (Davis et al., 2006b; Sanchez-Alavez et al., 2006; Pickering and OConnor, 2007; Viviani et al., 2007). In particular, the activation of the IL-1b-N-Smase-Src kinase pathway has been shown to mediate the fast actions of IL-1b on preoptical/anterior warm sensitive hypothalamic neurons underlying the rapid phase of the febrile response to IL-1b (Sanchez-Alavez et al., 2006). When studying these IL-1b effects in hypothalamic neurons in vivo and in vitro, we found that the effects of IL-1b can be mimicked by C2-ceramide, and that inhibitors of N-Smase in the brain lower or prevent the response to IL-1b (Sanchez-Alavez et al., 2006). Furthermore, the formed ceramide activates the Src-family of tyrosine kinases in hypothalamic neurons, but not in glia (Davis et al., 2006a, b). We show here, using a detailed pharmacological approach, that inhibition of N-Smase using 3-O-MS blocks the increase in the frequency and duration of seizures induced by IL-1b and that C2-ceramide faithfully mimics the IL-1b effect on seizures, thus suggesting that the activation of the N-Smaseceramide pathway mediates the proconvulsive activity of this cytokine. To elucidate the downstream events which follow N-Smase activation, we addressed the possible involvement

Fig. 3 IL-1b and seizure induced tyrosine phosphorylation of Srcfamily of kinases and the NR2B subunit of the NMDA receptor: effect of pharmacological treatments. Bargrams show densitometry analysis of the Src kinases and NR2B bands corresponding to their phosphorylated (p) forms (Src-Tyr418; NR2B-Tyr1472) in the various experimental groups, 60 min after the onset of kainateinduced seizures ($70 min after kainate injection), as assessed by western blot analysis of hippocampal homogenates. Vehicle- or IL-1b alone- treated mice were killed 70 min after the injection. Data (means SE, n = 5^7) are optical density (O.D.) values of the relevant bands (as depicted in the representative western blot), divided by the corresponding b-actin value (internal standard). Data are expressed as percentage of values measured in corresponding vehicle-treated mice. IL-1b (1ng/0.5 ml) was injected alone or 10 min before kainic acid; 3-O-MS (3 mg/0.5 ml) and CGP76030 (65 ng/0.5 ml) were injected 20 min before kainic acid (i.e. 10 min before IL-1b). 3-O-MS or CGP76030 alone did not change protein phosphorylation levels as compared to vehicle-injected mice; total NR2B levels were not changed by the various treatments (not shown). P50.05; P50.01 versus vehicle; ##P50.01 versus 3-O-MS+IL-1+KA and versus CGP76030+IL-1+KA (for pNR2B only) by two-way ANOVA followed by Kruskal ^Wallis test. Representative western blot bands corresponding to the specific proteins are depicted in B.
of the NR2B form, causing upregulation of channel gating properties (Salter and Kaila, 2004), was increased by 26% and 47% on average, respectively (P50.01). When hrIL-1b was coinjected with kainic acid, thus producing proconvulsive effects, (p)Src and (p)NR2B levels were increased by 127% and 82% respectively as compared to vehicle-treated mice (P50.01), denoting additive effects of the single treatments. We also evaluated whether inhibition of N-Smase or Src-family of kinases activity affects the increased (p)Src and (p)NR2B levels observed in proconvulsive conditions. 3-O-MS or CGP76030 at the doses which blocked the proconvulsive effects of IL-1b, reversed NR2B phosphorylation
Mechanism of IL-1b proconvulsive effect of the Src-family of kinases. Thus, this family of enzymes is known to be activated by ceramide (Kolesnick and Golde, 1994; Saklatvala, 1995), and a similar cascade of events was previosly shown to mediate the electrophysiological effects of IL-1b on the activity of warm-sensitive hypothalamic neurons (Davis et al., 2006b; Sanchez-Alavez et al., 2006). Our data show that the inhibition of the Src-family of kinases activity prevents the IL-1b and C2-ceramide proconvulsive effects similarly, thus suggesting that Src-family of tyrosine kinases activation is a downstream event subsequent to IL-1b activation of N-Smase. This possibility is also supported by the reversal of IL-1b-induced tyrosine phosphorylation of Src-family of kinases during seizures using the N-Smase inhibitor. Src-family of tyrosine kinases is abundantly expressed in neurons, and one main function of the activated forms of two members of this family, namely c-Src and Fyn, is to upregulate the activity of NMDA receptors via Tyr1472phosphorylation of the NR2B subunit (Ali and Salter, 2001). We found that ifenprodil, a selective antagonist of NMDA receptors containing the NR2B subunit, prevented the proconvulsive effects of IL-1b, thus indicating that this subunit is critically involved in IL-1b action on seizures. Our previous findings in primary cultures of hippocampal neurons showed that IL-1b induces the Tyr1472-phosphorylation of the NR2B subunit via Src-family of kinases activation (Viviani et al., 2003), therefore we assessed whether these events occurred in vivo when seizures were potentiated by IL-1b. The levels of the phosphorylated forms of Src-family of kinases and the NR2B subunit in the hippocampus were indeed enhanced by IL-1b or seizures, and these effects were additive when seizures were increased by IL-1b. Moreover, NR2B Tyr1472-phosphorylation induced in proconvulsive conditions was reversed to baseline level in mice treated with N-Smase and Src-family of kinases inhibitors, thus supporting the involvement of these enzymes in the activation of this receptor subunit (Yu et al., 1997). Tyr1472-phosphorylation of the NR2B subunit promotes Ca2+ influx into neurons (Yu et al., 1997; Viviani et al., 2003), thus resulting in potentiation of NMDA function which is pivotal for evoking neuronal hyperexcitability (Meador, 2007). This molecular event may be the ultimate step of this novel IL-1b-N-Smase-Src kinase pathway responsible for the increased neuronal excitability and subsequent proconvulsive effects of this cytokine. Although CGP76030 selectively inhibits Src-family of tyrosine kinases activation, and it was shown to have higher affinity for c-Src versus other Src-family of tyrosine kinase members (Susa et al., 2000, 2005), no data about the inhibitor IC50 on Fyn activity are available, therefore both c-Src and Fyn may be involved in the IL-1b signaling underlying its proconvulsive effects. The possible specific involvement of c-Src versus Fyn should await the possibility to unequivocally distinguish the involvement of these two kinase activities using pharmacological approaches which are not yet available.

Interestingly, Src-family of tyrosine kinases appears to contribute to seizures also in the absence of proinflammatory conditions since Src-family of kinases inhibition per se after icv, but not intrahippocampal injection of CGP76030, reduced seizures number and duration in mice. We could not affect seizures by inhibiting the activity of N-Smase using 3-O-MS either intrahippocampally or icv injected. The lack of effect of intrahippocampal application of these drugs on kainate-induced seizures may be due to the local spread of the inhibitors around the injection site as opposed to seizures-induced IL-1b expression throughout the whole hippocampus. Moreover, we should also consider the possibility that the dose of 3-O-MS allowed to be injected by the limited solubility of this compound is not enough to protect from kainate seizures. Although hippocampal EEG seizure activity was increased by IL-1b or C2-ceramide, cell death in the hippocampus was not affected. This observation indicates that the increase in seizures was insufficient to induced additional cell loss which was restricted in this model to CA3 pyramidal cells in the injected hippocampus. Moreover, CA3 degeneration has been reported to depend on a direct neurotoxic action of kainate rather than on seizures per se (Fariello et al, 1989; Jarrard, 2002), therefore suggesting that kainate-induced cell loss in this model is insensitive to the proneurotoxic actions of IL-1b. The relevance of our findings for the etiopathogenesis of seizures is based on clinical and experimental evidence showing that the IL-1b system is activated in chronic human epilepsy (Vezzani and Granata, 2005; Ravizza et al., 2006a, 2008) and that anti-IL-1b pharmacological treatments result in powerful anticonvulsant effects (De Simoni et al., 2000; Vezzani et al., 2000, 2002; Ravizza et al., 2006b). Moreover, several brain injuries in humans are associated with brain inflammation, result in early occurrence of seizures and present a high risk of developing epilepsy (Pitkanen and Sutula, 2002; Vezzani and Granata, 2005). Therefore, the elucidation of this novel IL-1bactivated pathway in the hippocampus adds important insights into the mechanisms of ictogenesis in inflammatory conditions and may allow the development of innovative strategies to block the activation of IL-1b signaling in disease conditions, thus highlighting potential new targets of therapeutic intervention.

Funding

EPICURE (LSH-CT-2006-037315 to AV), Negri Weizmann Programme (to A.V.) and Fondazione Monzino (to A.V.) and the NIH grant NS43501-04 (to T.B.)

References Ali DW, Salter MW. NMDA receptor regulation by Src kinase signalling in excitatory synaptic transmission and plasticity. Curr Opin Neurobiol 2001; 11: 33642.
Obeid LM, Linardic CM, Karolak LA, Hannun YA. Programmed cell death induced by ceramide. Science 1993; 259: 176971. Papp S, Szabo E, Kim H, McCullock CA, Opas M. Kinase-dependent adhesion to fibronectin: regulation by calreticulin. Exp Cell Res 2008; 314: 131326. Pawlak R, Melchor JP, Matys T, Skrzypiec AE, Strickland S. Ethanolwithdrawal seizures are controlled by tissue plasminogen activator via modulation of NR2B-containing NMDA receptors. Proc Natl Acad Sci USA 2005; 102: 4438. Pickering M, OConnor JJ. Pro-inflammatory cytokines and their effects in the dentate gyrus. Prog Brain Res 2007; 163: 33954. Pitkanen A, Sutula TP. Is epilepsy a progressive disorder? Prospects for new therapeutic approaches in temporal-lobe epilepsy. Lancet Neurol 2002; 1: 17381. Ravizza T, Boer K, Redeker S, Spliet WG, van Rijen PC, Troost D, et al. The IL-1beta system in epilepsy-associated malformations of cortical development. Neurobiol Dis 2006a; 24: 12843. Ravizza T, Gagliardi B, Noe F, Boer K, Aronica E, Vezzani A. Innate and adaptive immunity during epileptogenesis and spontaneous seizures: evidence from experimental models and human temporal lobe epilepsy. Neurobiol Dis 2008; 29: 14260. Ravizza T, Lucas SM, Balosso S, Bernardino L, Ku G, Noe F, et al. Inactivation of caspase-1 in rodent brain: a novel anticonvulsive strategy. Epilepsia 2006b; 47: 11608. Ravizza T, Vezzani A. Status epilepticus induces time-dependent neuronal and astrocytic expression of interleukin-1 receptor type I in the rat limbic system. Neuroscience 2006; 137: 3018. Rucci N, Recchia I, Angelucci A, Alamanou M, Del Fattore A, Fortunati D, et al. Inhibition of protein kinase c-Src reduces the incidence of breast cancer metastases and increases survival in mice: implications for therapy. J Pharmacol Exp Ther 2006; 318: 16172. Saklatvala J. Intracellular signalling mechanisms of interleukin 1 and tumour necrosis factor: possible targets for therapy. Br Med Bull 1995; 51: 40218. Salter MW, Kalia LV. Src kinases: a hub for NMDA receptor regulation. Nat Rev Neurosci 2004; 5: 31728. Sanchez-Alavez M, Tabarean IV, Behrens MM, Bartfai T. Ceramide mediates the rapid phase of febrile response to IL-1beta. Proc Natl Acad Sci USA 2006; 103: 29048. Sayyah M, Javad-Pour M, Ghazi-Khansari M. The bacterial endotoxin lipopolysaccharide enhances seizure susceptibility in mice: involvement of proinflammatory factors: nitric oxide and prostaglandins. Neuroscience 2003; 122: 107380. Susa M, Missbach M, Green J. Src inhibitors: drugs for the treatment of osteoporosis, cancer or both? TiPS 2000; 21: 48995. Susa M, Missbach M, Gamse R, Kneissel M, Buhl T, Gasser JA et al. Src as a target for pharmaceutical intervention. In: Fabbro D, McCormick F, editors. Cancer drug discovery and development. Totowa: Humana Press Inc.; 2005. p. 7192. Tsakiri N, Kimber I, Rothwell NJ, Pinteaux E. Interleukin-1-induced interleukin-6 synthesis is mediated by the neutral sphingomyelinase/Src kinase pathway in neurones. Br J Pharmacol 2008; 153: 77583. Turrin NP, Rivest S. Innate immune reaction in response to seizures: implications for the neuropathology associated with epilepsy. Neurobiol Dis 2004; 16: 32134. Vezzani A, Conti M, De Luigi A, Ravizza T, Moneta D, Marchesi F, et al. Interleukin-1beta immunoreactivity and microglia are enhanced in the rat hippocampus by focal kainate application: functional evidence for enhancement of electrographic seizures. J Neurosci 1999; 19: 505465. Vezzani A, Granata T. Brain inflammation in epilepsy: experimental and clinical evidence. Epilepsia 2005; 46: 172443. Vezzani A, Moneta D, Conti M, Richichi C, Ravizza T, De Luigi A, et al. Powerful anticonvulsant action of IL-1 receptor antagonist on intracerebral injection and astrocytic overexpression in mice. Proc Natl Acad Sci USA 2000; 97: 115349.

Vezzani A, Moneta D, Richichi C, Aliprandi M, Burrows SJ, Ravizza T, et al. Functional role of inflammatory cytokines and antiinflammatory molecules in seizures and epileptogenesis. Epilepsia 2002; 43 (Suppl 5): 305. Viviani B, Bartesaghi S, Gardoni F, Vezzani A, Behrens MM, Bartfai T, et al. Interleukin-1beta enhances NMDA receptor-mediated intracellular calcium increase through activation of the Src family of kinases. J Neurosci 2003; 23: 8692700. Viviani B, Gardoni F, Marinovich M. Cytokines and neuronal ion channels in health and disease. Int Rev Neurobiol 2007; 82: 24763.
Yen W, Williamson J, Bertram EH, Kapur J. A comparison of three NMDA receptor antagonists in the treatment of prolonged status epilepticus. Epilepsy Res 2004; 59: 4350. Yu XM, Askalan R, Keil GJ II, Salter MW. NMDA channel regulation by channel-associated protein tyrosine kinase Src. Science 1997; 275: 6748. Zeng C, Lee JT, Chen H, Chen S, Hsu CY, Xu J. Amyloid-beta peptide enhances tumor necrosis factor-alpha-induced iNOS through neutral sphingomyelinase/ceramide pathway in oligodendrocytes. J Neurochem 2005; 94: 70312.

HIV-Specific IL-10-Positive CD8+ T Cells Suppress Cytolysis and IL-2 Production by CD8 + T Cells
This information is current as of June 7, 2011 Mohamed Elrefaei, Florence L. Ventura, Chris A. R. Baker, Richard Clark, David R. Bangsberg and Huyen Cao J Immunol 2007;178;3265-3271

References

This article cites 56 articles, 25 of which can be accessed free at: http://www.jimmunol.org/content/178/5/3265.full.html#ref-list-1
Downloaded from www.jimmunol.org on June 7, 2011
Article cited in: http://www.jimmunol.org/content/178/5/3265.full.html#related-urls Subscriptions Permissions Email Alerts Information about subscribing to The Journal of Immunology is online at http://www.jimmunol.org/subscriptions Submit copyright permission requests at http://www.aai.org/ji/copyright.html Receive free email-alerts when new articles cite this article. Sign up at http://www.jimmunol.org/etoc/subscriptions.shtml/
The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 9650 Rockville Pike, Bethesda, MD 20814-3994. Copyright 2007 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606.
The Journal of Immunology
HIV-Specic IL-10-Positive CD8 T Cells Suppress Cytolysis and IL-2 Production by CD8 T Cells1
Mohamed Elrefaei,2* Florence L. Ventura,* Chris A. R. Baker,* Richard Clark, David R. Bangsberg, and Huyen Cao*
IL-10 producing T cells inhibit Ag-specic CD8 T cell responses and may play a role in the immune dysregulation observed in HIV infection. We have previously observed the presence of HIV-specic IL-10-positive CD8 T cells in advanced HIV disease. In this study, we examined the suppressive function of the Gag-specic IL-10-positive CD8 T cells. Removal of these IL-10positive CD8 T cells resulted in increased cytolysis and IL-2, but not IFN-, production by both HIV- and human CMV-specic CD8 T cells. In addition, these IL-10-positive CD8 T cells mediated suppression through direct cell-cell contact, and had a distinct immunophenotypic prole compared with other regulatory T cells. We describe a new suppressor CD8 T cell population in advanced HIV infection that may contribute to the immune dysfunction observed in HIV infection. The Journal of Immunology, 2007, 178: 32653271. istinct CD8 T cell populations with regulatory properties produce IL-10 (13). Suppressor CD8 T cells have been shown to inhibit both CD4 and CD8 T cell effector functions (2, 4), although the mechanism of this inhibition is not well dened. Suppressor CD8 T cells inhibit T cell proliferation and cytotoxicity via contact-dependent mechanisms and/or by using cytokines such as IL-10 (2, 5, 6). Suppressor CD8 T cells have also been shown to decrease the expression of MHC class I on target cells through IL-10 secretion (3). Both CD4 and CD8 T cells have been shown to express high IL-10 levels in HIV infection (7, 8). However, the suppressor role of the IL-10-positive CD8 T cells is not fully understood (9 13). IL-10 producing HIV-specic CD4 regulatory T cells suppressed both CD4 and CD8 HIV-specic T cell function in vitro (14 16). Previous studies have found that the proliferation of CD4 T cells from HIV-infected individuals was inversely associated with the levels of IL-10 and that responses were partially restored by anti-IL-10 Abs (17, 18). In addition, a higher frequency of CD4 T cells producing IL-10 was associated with a more progressive HIV disease (19). We have previously demonstrated the association of HIV-specic, IL-10-positive CD8 T cells with advanced HIV disease and higher plasma HIV RNA (20). We postulated that HIV-specic IL-10-positive CD8 T cells exert an inhibitory effect on distinct populations of effector CD8 T cells. In this study, we describe the novel suppressive effect of the HIV-specic IL-10-positive CD8 T cells in regulating CD8 T cell responses.

Materials and Methods

Study subjects and samples
HIV-positive volunteers (n 10) were recruited from the Research in Access to Care in the Homeless (REACH) cohort in San Francisco, CA, as previously described (21). Demographic information and the CD4 T cell count were obtained at the time of enrollment and blood draw. Institutional Review Board approvals were obtained from the California Department of Health Services and University of California San Francisco Committee on Human Research, and all study participants gave written informed consent. None of the study participants received antiretroviral therapy at the time of the study (6 mo before enrollment). The HIV-1 RNA level was determined from the plasma using the Roche Amplicor 1.5 as per the manufacturers recommendations. PBMC were separated and cryopreserved in liquid nitrogen until assay time.

Antigens

Peptides corresponding to the sequences of the clade B consensus sequences of HIV-1 for Gag (www.hiv.lanl.gov/content/hiv-db/ CONSENSUS/M_GROUP/Consensus.html) were synthesized as 15 aa overlapping by 11 aa (Mitochor Mimotopes). The Gag synthetic peptides used for all T cell assays were pooled into one single pool of peptides (total 123) with a nal concentration of 1 g/ml per peptide (22). A single pool of overlapping peptides corresponding to the amino acid sequence of the PP65 protein (BD Biosciences) was used to detect a human CMV (HCMV)3-specic response (23, 24).
Flow-based intracellular IL-10 staining
The detection of Gag-specic IL-10 production was performed using PBMC (1 106) incubated with Gag peptide pools for 2 h at 37C in 5% CO2 in the presence of costimulatory anti-CD49d and anti-CD28 (1 g/ml; BD Biosciences) followed by GolgiStop (BD Pharmingen) for 1214 h as previously described to be the optimal incubation period for the detection of IL-10 without affecting the viability of the cells (25). LPS (1 ng/ml; Sigma-Aldrich) and medium alone without Ag stimulation were used as positive and negative controls, respectively. Cells were then stained with the following Abs: IL-10 PE, CD4 PE Cy7, CD3 PerCP Cy5.5, and CD8 Pacic Blue (BD Pharmingen). A minimum of 30,000 CD3 cells per sample was acquired using a 6-color ow cytometer (LSR II; BD Biosciences) and analysis was performed by FlowJo software (TreeStar). The results were expressed as the percentage of IL-10-positive CD8 T cells (percentage positive percentage of Ag-specic percentage of negative control). Responses 0.1% and two times the background were considered
*California Department of Health Services, Richmond, CA 94804; and Department of Medicine, University of California, San Francisco, CA 94143 Received for publication September 27, 2006. Accepted for publication December 1, 2006. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1 This work was supported by National Institutes of Health Grants AI43885 and MH54907. 2 Address correspondence and reprint requests to Dr. Mohamed Elrefaei, California Department of Health Services, 850 Marina Bay Parkway, Viral and Rickettsial Disease Laboratory, Richmond, CA, 94804. E-mail address: melrefae@dhs.ca.gov
Abbreviations used in this paper: HCMV, human CMV; ILT, Ig-like transcript.
Copyright 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00

www.jimmunol.org

EFFECT OF IL-10-POSITIVE CD8 T CELLS IN HIV INFECTION
FIGURE 1. The removal of HIV-specic, IL-10-positive CD8 T cells is associated with increased HIV-specic cytolysis. PBMC were stimulated with Gag B peptides and then IL-10-positive cells were isolated. Eluted IL-10-negative PBMC were stimulated for an additional 6 h with Gag B peptides in the presence of anti-CD107a/b FITC and then stained with anti-IFN- PE, anti-CD3 PerCP Cy5.5, anti-CD4 PE Cy7, and anti-CD8 Pacic Blue and analyzed by ow cytometry. Samples were rst gated on the CD3CD8 lymphocyte population and then the percentages of CD107a/b- and IFN--positive cells were determined. The results were expressed as the percentage of Gag-specic CD8 T cells expressing CD107a/b or IFN- after subtraction of the background values. A, Representative plots of Gag-specic CD8 T cells expressing CD107a/b or IFN- in the presence or absence of IL-10-positive cells. B, Dashed line represents the cutoff for signicant CD107a/b and IFN- expression. The numbers in parentheses are median values. The two circles joined by a line represent the values obtained from the same individual, and analysis was performed by a paired t test. Signicant differences were observed in the percent of Gag-specic CD8 T cells expressing CD107a/b (p 0.0006), but not IFN- (p 0.28), between the two groups.
positive. All volunteers demonstrated signicant IL-10 production following LPS stimulation. Background expression was 0.05%.

MACS cytokine secretion/separation assay
IL-10-positive cells were isolated using the MACS cytokine secretion/ separation kit following the manufacturers protocol (Miltenyi Biotec). Briey, PBMC were stimulated with Gag B peptides for 6 h and then labeled with anti-IL-10 PE detection Ab followed by magnetic labeling with anti-PE microbeads. IL-10-positive cells were isolated by magnetic separation using the MACS columns. Unfractionated and eluted IL-10negative PBMC were stimulated for an additional 6 h with either Gag B or CMV PP65 peptides in a CD107a/b degranulation and intracellular cytokine staining assays as described below.
Flow-based CD107a/b degranulation and intracellular cytokine staining assay
A degranulation assay was performed as previously described (26). Cytotoxic CD8 T cell function has been shown to correlate directly with T cell degranulation, which is a prerequisite process of perforin/granzyme-mediated lytic function (27) and can be measured by the increased expression of
surface CD107 (28). Briey, PBMC (1 106) were incubated with 1 g/ml anti-CD49d and anti-CD28 (BD Biosciences), anti-CD107a and antiCD107b FITC (CD107a/b) (BD Biosciences), and Gag B peptide pools in a 0.2-ml nal volume for 6 h in the presence of GolgiStop (BD Pharmingen). Staphylococcus enterotoxin B (1 g/ml; Sigma-Aldrich) and medium alone were used as positive and negative controls respectively. Cells were then stained with the following Abs: IFN- PE, CD4 PE Cy7, CD3 PerCP Cy5.5, IL-2 allophycocyanin, and CD8 Pacic Blue (BD Pharmingen). Sample acquisition and analysis was performed as described above. Results were expressed as the percentage of CD107a/b-, IFN--, or IL-2-positive CD8 T cells (percentage of positive percentage of Ag-specic percentage of negative control). Responses 0.1% and two times the background were considered positive. All volunteers demonstrated signicant CD107a/b, IFN-, and IL-2 expression following Staphylococcus enterotoxin B stimulation. Background expression was 0.1%.

Transwell experiments

PBMC were cultured with or without IL-10-positive cells in the same well or separated by a microporous membrane (0.4-m pore size; Costar, Corning) as previously described (29). Briey, PBMC were placed into the lower chamber of a 24-well plate and the IL-10-positive cells were added

FIGURE 2. The removal of HIV-specic IL-10-positive CD8 T cells is associated with increased HIV-specic IL-2-positive CD8 T cells. PBMC were stimulated with Gag B peptides and then IL-10-positive cells were isolated. Eluted IL-10-negative PBMC were stimulated for an additional 6 h with Gag B peptides and then stained with anti-IL-2 PE, anti-CD3 PerCP Cy5.5, anti-CD4 PE Cy7, and anti-CD8 Pacic Blue and analyzed by ow cytometry. Samples were rst gated on the CD3CD8 lymphocyte population and then the percentage of IL-2-positive CD8 T cells was determined. Results were expressed as the percentage of Gag-specic CD8 T cells expressing IL-2 after subtraction of the background values. A, Representative plots of Gagspecic CD8 T cells expressing IL-2 in the presence or absence of IL-10-positive cells. B, Dashed line represents the cutoff for signicant IL-2 expression. The numbers in parentheses are median values. The two circles joined by a line represent the values obtained from the same individual, and analysis was performed by paired t test. Signicant differences were observed in the percentage of Gag-specic CD8 T cells expressing IL-2 (p 0.005) between the two groups.
into the same chamber or the upper chamber. Culture, stimulation, and staining procedures were performed as described above.
Flow-based FOXP3 staining assay
PBMC were stained for FOXP3 expression following the manufacturers protocol (eBioscience) with some modications. Briey, PBMC (1 106) were incubated with Gag B peptide pools for 1214 h as described above for intracellular IL-10 staining. Cells were then stained with different combinations of the Abs IL-10 PE, CD4 PE Cy7, CD3 PerCP Cy5.5, CD25 allophycocyanin, FOXP3 allophycocyanin, and CD8 Pacic Blue (BD Pharmingen) and analysis was performed by ow cytometry as described above. The percentage of IL-10-positive cells was determined and the extent of FOXP3 and CD25 expression was also assessed. Results were expressed as the fraction of IL-10-positive cells that expressed FOXP3 or CD25 over the total number of IL-10-positive cells (equivalent to 100%).

creased frequency of HIV-specic CD107a/b-positive ( p 0.0006), but not IFN--positive ( p 0.28), CD8 T cells compared with unfractionated PBMC (Fig. 1B). Adding back these autologous Gag-specic IL-10-positive CD8 T cells to the eluted IL-10-negative PBMC signicantly reduced the frequency of CD107a/b-positive CD8 T cells compared with IL-10-negative

Statistical analysis

Groups were compared using the paired t test, and analysis was performed with PRISM software version 4.02 (GraphPad). Statistical signicance was dened as p 0.05.

Results

Removal of HIV-specic IL-10-positive CD8 T cells results in increased HIV-specic cytolysis We have previously observed the association of HIV-specic, IL10-positive CD8 T cells with advanced disease and decreased HIV-specic cytolysis (20). We examined the effect of in vitro removal of the Gag-specic IL-10-positive CD8 T cells from HIV-positive volunteers (n 10). All volunteers had a CD4 T cell count of 200 cells/mm3 (median, 139; range, 11199) and HIV plasma RNA of 75,000 copies/ml (median, 227,000; range, 79,000 690,000). In addition, all volunteers demonstrated significant Gag-specic, IL-10-positive CD8 T cell responses (median, 0.4%; range, 0.21.7%). No Gag-specic IL-10-positive CD4 T cell responses were detected (data not shown). CD107a/b expression (26) and IL-2 and IFN- production by Gag-specic CD8 T cells were also determined concurrently. All volunteers demonstrated signicant responses to at least one of the three markers examined. Eluted IL-10-negative PBMC were evaluated for HIVspecic cytolysis and IFN- production using the CD107a/b degranulation and intracellular cytokine staining assay. Representative plots of the effect of the removal of HIV-specic, IL-10-positive CD8 T cells on HIV-specic cytolysis are shown in Fig. 1A. Removal of the IL-10-positive CD8 T cells resulted in a signicantly in-
FIGURE 3. The removal of HIV-specic, IL-10-positive CD8 T cells is associated with increased CMV-specic cytolysis and increased frequency of CMV-specic IL-2-positive CD8 T cells. PBMC were stimulated with Gag B peptides and then IL-10-positive cells were isolated. Eluted IL-10-negative PBMC were stimulated for an additional 6 h with CMV PP65 peptides in the presence of anti-CD107a/b FITC and then stained with anti-IFN- PE, anti-CD3 PerCP Cy5.5, anti-CD4 PE Cy7, anti-IL-2 allophycocyanin, and anti-CD8 Pacic Blue and analyzed by ow cytometry. Samples were rst gated on the CD3CD8 lymphocyte population and then the percentages of CD107a/b-, IL-2-, and IFN--positive cells were determined. The results were expressed as percentage of HCMV-specic CD8 T cells expressing CD107a/b, IL-2, or IFN- after subtraction of the background values. The dashed line represents the cutoff for signicant CD107a/b, IL-2, and IFN- expression. The percentages in parentheses are median values. The two circles joined by a line represent the values obtained from the same individual, and analysis was performed by a paired t test. Signicant differences were observed in the percentages of HCMV-specic CD8 T cells expressing CD107a/b (p 0.006) and IL-2 (p 0.0002), but not IFN- (p 0.52), between the two groups.

FIGURE 4. IL-10-positive CD8 T cells decrease HIV-specic CD107a/b expression and IL-2 production by direct cell-cell contact. PBMC were stimulated with Gag B peptides then IL-10-positive cells were isolated. Eluted IL-10-negative PBMC were cultured for an additional 6 h with Gag B peptides and anti-CD107a/b FITC, with or without IL-10-positive cells in the same well or separated by a microporous membrane. PBMC were then stained with anti-IFN- PE, anti-CD3 PerCP Cy5.5, anti-CD4 PE Cy7, anti-IL-2 allophycocyanin, and anti-CD8 Pacic Blue and analyzed by ow cytometry. Samples were rst gated on the CD3CD8 lymphocyte population and then the percentage of CD107a/b-, IFN--, and IL-2-positive cells were determined. Results were expressed as the percentages of Gag-specic CD8 T cells expressing CD107a/b, IFN-, or IL-2 after subtraction of the background values. Plots are representative of four independent experiments yielding similar results.
PBMC alone ( p 0.005; data not shown), comparable to the frequency observed in unfractionated PBMC ( p 0.4; data not shown). These results suggest that the Gag-specic, IL-10-positive CD8 T cells may have an inhibitory effect on HIV-specic cytolytic function. Removal of HIV-specic IL-10-positive CD8 T cells resulted in increased frequency of HIV-specic, IL-2-positive CD8 T cells Regulatory T cells have been shown to modulate IL-2 production by T cells in vitro (30). We next examined the effect of in vitro removal of the Gag-specic IL-10-positive CD8 T cells on HIVspecic IL-2-positive CD8 T cells. No Gag-specic IL-2-positive CD4 T cell responses were detected in our study volunteers (data not shown). Representative plots are shown in Fig. 2A. Removal of the IL-10-positive CD8 T cells resulted in signicantly increased frequency of HIV-specic IL-2-positive CD8 T cells compared with the unfractionated PBMC ( p 0.005; Fig. 2B). Adding back these autologous Gag-specic IL-10-positive CD8 T cells significantly decreased the frequency of IL-2-positive CD8 T cells compared with IL-10-negative PBMC alone ( p 0.007; data not shown), comparable to the frequency observed in unfractionated PBMC ( p 0.7; data not shown). These results suggest that the Gag-specic IL-10-positive CD8 T cells may have an inhibitory effect on HIV-specic IL-2 production by CD8 T cells. Removal of HIV-specic, IL-10-positive CD8 T cells resulted in increased cytolysis and IL-2 production by HCMV-specic CD8 T cells Suppressor CD8 T cells have been shown to mediate non-Ag specic suppression of T cell responses (2, 3, 31, 32). To determine whether Gag-specic, IL-10-positive CD8 T cells can also suppress non-HIV-specic T cell responses, we analyzed HCMVspecic responses in the same study population. PP65-specic CD107a/b, IL-2, and IFN- CD8 T cell responses were detected in all 10 study participants (Fig. 3). No HCMV-specic, IL-10positive CD4 or CD8 T cell responses were detected (data not shown). Interestingly, removal of the Gag-specic IL-10-positive

CD8 T cells resulted in a signicantly increased frequency of PP65-specic CD107a/b-positive and IL-2-positive, but not IFN-positive, CD8 T cells (see Fig. 5; p 0.006, p 0.0002, and p 0.52, respectively). Adding back autologous Gag-specic, IL10-positive CD8 T cells resulted in a signicantly decreased frequency of CD107a/b-positive and IL-2-positive, HCMV-specic CD8 T cells compared with IL-10-negative PBMC alone ( p 0.02, and p 0.001 for CD107a/b and IL-2, respectively; data not shown), comparable to the frequencies observed in unfractionated PBMC ( p 0.2, and p 0.7 for CD107a/b and IL-2, respectively; data not shown). These results suggest that the Gag-specic IL10-positive CD8 T cells do not require Ag specicity to inhibit cytolysis and IL-2 production. IL-10-positive CD8 T cells suppress CD107a/b and IL-2 expression by direct cell-cell contact Suppressor CD8 T cells exert their inhibitory function on T cell proliferation and CTL via contact-dependent mechanisms and/or by using cytokines such as IL-10 and TGF- (2, 5, 6). To determine the mechanism by which Gag-specic, IL-10-positive CD8 T cells mediated the suppression of HIV-specic responses, we performed transwell experiments on four HIV-positive volunteers. Representative plots are shown in Fig. 4. Gag-specic IL10-positive CD8 T cells, when separated from the IL-10-negative PBMC by a microporous membrane, had no effect on HIV-specic CD107a/b expression and IL-2 production compared with IL-10-negative PBMC alone. In contrast, the coculture of IL-10-positive CD8 T cells and IL-10-negative PBMC in the same well led to a strong suppression of HIV-specic CD107a/b-positive and IL-2-positive, but not IFN--positive, CD8 T cells compared with IL-10-negative PBMC alone. Analysis of the HCMV-specic responses in these four HIVpositive volunteers was also performed and the results were similar to the observed suppressive effects of the IL-10-positive CD8 T cells on the HIV-specic responses. Representative plots
FIGURE 5. IL-10-positive CD8 T cells decrease HCMV-specic CD107a/b expression and IL-2 production by direct cell-cell contact. PBMC were stimulated with Gag B peptides and then IL-10-positive cells were isolated. Eluted IL-10-negative PBMC were cultured for an additional 6 h with HCMV peptides and anti-CD107a/b FITC, with or without IL-10-positive cells in the same well or separated by a microporous membrane. PBMC were then stained with anti-IFN- PE, anti-CD3 PerCP Cy5.5, anti-CD4 PE Cy7, anti-IL-2 allophycocyanin, and anti-CD8 Pacic Blue and analyzed by ow cytometry. Samples were rst gated on the CD3CD8 lymphocyte population then the percentages of CD107a/ b-, IFN--, and IL-2-positive cells were determined. Results were expressed as percent of HCMV-specic CD8 T cells expressing CD107a/b, IFN-, or IL-2 after subtraction of the background values. Plots are representative of four independent experiments yielding similar results.

are shown in Fig. 5. These results indicate that Gag-specic IL10-positive CD8 T cells suppress HIV- and HCMV-specic cytolysis and IL-2 production by direct cell-cell contact. HIV-specic IL-10-positive CD8 T cells do not express FOXP3 or CD25 Suppressor CD8 T cells in HIV-negative individuals are CD28negative (5, 6) and express CD25, CD103, FOXP3, TNF receptorrelated protein (GITR), and CTLA-4 (33 40). We determined whether the HIV-specic IL-10-positive CD8 T cells express a similar immunophenotypic prole. PBMC were stimulated with Gag B peptides and the extent of FOXP3 and CD25 expression by the Gag-specic IL-10-positive CD8 T cells was assessed. Representative plots of the phenotype of the HIV-specic, IL-10-positive CD8 T cells are shown in Fig. 6. IL-10-positive CD8 T
cells were mainly FOXP3-negative and CD25-negative (medians of 9 and 8.5% for FOXP3 and CD25, respectively; data not shown). These results suggest that the Gag-specic, IL-10-positive CD8 T cells have a distinct regulatory T cell immunophenotypic prole compared with other non-HIV suppressor CD8 T cells and support our current use of functional characterization by IL-10 production to identify this suppressor T cell population in HIV infection.

Discussion

Regulatory T cells are implicated in modulating immune responses to chronic viral infections including HIV (41). CD4 regulatory T cells inhibit both CD4 and CD8 T cell-mediated HIV-specic immune responses in vitro and may have HIV specicity (14 16). However, the role of suppressor CD8 T cells in HIV infection has not been fully addressed. In this study, we observed signicantly increased HIV- and HCMV-specic cytolysis following in vitro removal of the Gag-specic, IL-10-positive CD8 T cells. These results led us to postulate that the Gag-specic, IL-10-positive CD8 T cells impair the ability of both HIV- and non-HIV-specic CTL to kill infected target cells by suppressing HIV-specic, CD8 T cell degranulation. In addition, the Gag-specic IL-10positive CD8 T cells displayed a suppressive effect on more than one CD8 T cell function, including the inhibition of IL-2 production. The apparent lack of effect on IFN--positive CD8 T cells suggest that inhibition by the Gag-specic IL-10-positive CD8 T cells may be very selective. Our data support a mechanism of Ag-specic induction of suppressor IL-10-positive CD8 T cells, which then exert suppression in a nonspecic manner (42). We have described the presence of suppressor IL-10-positive CD8 T cells that are Gag specic. However, suppressor IL-10positive CD8 T cells specic to other HIV Ags may also exist. Whether distinct populations of HIV-specic IL-10-positive CD8 T cells display similar inhibition remains to be determined. We have demonstrated that the suppression by Gag-specic IL-10-positive CD8 T cells appeared to be mediated by direct

FIGURE 6. HIV-specic IL-10-positive CD8 T cells do not express FOXP3 or CD25. PBMC were stimulated with Gag B peptides and then stained with anti-IL-10 PE, anti-CD3 PerCP Cy5.5, anti-CD4 PE Cy7, anti-CD8 Pacic Blue, and either anti-FOXP3 allophycocyanin or antiCD25 allophycocyanin and then analyzed by ow cytometry. Samples were rst gated on the CD3CD8 lymphocyte populations and then the percentages of IL-10-positive cells were determined and the extent of FOXP3 and CD25 expression was also assessed. Representative plots of Gag-specic IL-10-positive CD8 T cells expressing FOXP3 and CD25 are shown. The values marked with an asterisk represent the fraction of IL-10-positive cells that express FOXP3 or CD25 over the total number of IL-10-positive cells (equivalent to 100%).
5. Kumar, V. 2004. Homeostatic control of immunity by TCR peptide-specic Tregs. J. Clin. Invest. 114: 12221226. 6. Jiang, H., and L. Chess. 2004. An integrated view of suppressor T cell subsets in immunoregulation. J. Clin. Invest. 114: 1198 1208. 7. Graziosi, C., G. Pantaleo, K. R. Gantt, J. P. Fortin, J. F. Demarest, O. J. Cohen, R. P. Sekaly, and A. S. Fauci. 1994. Lack of evidence for the dichotomy of TH1 and TH2 predominance in HIV-infected individuals. Science 265: 248 252. 8. Zanussi, S., C. Simonelli, M. DAndrea, C. Caffau, M. Clerici, U. Tirelli, and P. DePaoli. 1996. CD8 lymphocyte phenotype and cytokine production in longterm non-progressor and in progressor patients with HIV-1 infection. Clin. Exp. Immunol. 105: 220 224. 9. Chen, W. F., and A. Zlotnik. 1991. IL-10: a novel cytotoxic T cell differentiation factor. J. Immunol. 147: 528 534. 10. Rohrer, J. W., and J. H. Coggin, Jr. 1995. CD8 T cell clones inhibit antitumor T cell function by secreting IL-10. J. Immunol. 155: 5719 5727. 11. Taga, K., and G. Tosato. 1992. IL-10 inhibits human T cell proliferation and IL-2 production. J. Immunol. 148: 11431148. 12. Ebert, E. C. 2000. IL-10 enhances IL-2-induced proliferation and cytotoxicity by human intestinal lymphocytes. Clin. Exp. Immunol. 119: 426 432. 13. Kang, S. S., and P. M. Allen. 2005. Priming in the presence of IL-10 results in direct enhancement of CD8 T cell primary responses and inhibition of secondary responses. J. Immunol. 174: 53825389. 14. Kinter, A. L., M. Hennessey, A. Bell, S. Kern, Y. Lin, M. Daucher, M. Planta, M. McGlaughlin, R. Jackson, S. F. Ziegler, and A. S. Fauci. 2004. CD25CD4 regulatory T cells from the peripheral blood of asymptomatic HIV-infected individuals regulate CD4 and CD8 HIV-specic T cell immune responses in vitro and are associated with favorable clinical markers of disease status. J. Exp. Med. 200: 331343. 15. Weiss, L., V. Donkova-Petrini, L. Caccavelli, M. Balbo, C. Carbonneil, and Y. Levy. 2004. Human immunodeciency virus-driven expansion of CD4CD25 regulatory T cells, which suppress HIV-specic CD4 T-cell responses in HIV-infected patients. Blood 104: 3249 3256. 16. Aandahl, E. M., J. Michaelsson, W. J. Moretto, F. M. Hecht, and D. F. Nixon. 2004. Human CD4 CD25 regulatory T cells control T-cell responses to human immunodeciency virus and cytomegalovirus antigens. J. Virol. 78: 2454 2459. 17. Kumar, A., J. B. Angel, M. P. Daftarian, K. Parato, W. D. Cameron, L. Filion, and F. Diaz-Mitoma. 1998. Differential production of IL-10 by T cells and monocytes of HIV-infected individuals: association of IL-10 production with CD28mediated immune responsiveness. Clin. Exp. Immunol. 114: 78 86. 18. Clerici, M., T. A. Wynn, J. A. Berzofsky, S. P. Blatt, C. W. Hendrix, A. Sher, R. L. Coffman, and G. M. Shearer. 1994. Role of interleukin-10 in T helper cell dysfunction in asymptomatic individuals infected with the human immunodeciency virus. J. Clin. Invest. 93: 768 775. 19. Ostrowski, M. A., J. X. Gu, C. Kovacs, J. Freedman, M. A. Luscher, and K. S. MacDonald. 2001. Quantitative and qualitative assessment of human immunodeciency virus type 1 (HIV-1)-specic CD4 T cell immunity to gag in HIV-1-infected individuals with differential disease progression: reciprocal interferon- and interleukin-10 responses. J. Infect. Dis. 184: 1268 1278. 20. Elrefaei, M., B. Barugahare, F. Ssali, P. Mugyenyi, and H. Cao. 2006. HIVspecic IL-10-positive CD8 T cells are increased in advanced disease and are associated with decreased HIV-specic cytolysis. J. Immunol. 176: 1274 1280. 21. Moss, A. R., J. A. Hahn, S. Perry, E. D. Charlebois, D. Guzman, R. A. Clark, and D. R. Bangsberg. 2004. Adherence to highly active antiretroviral therapy in the homeless population in San Francisco: a prospective study. Clin. Infect. Dis. 39: 1190 1198. 22. McEvers, K., M. Elrefaei, P. Norris, S. Deeks, J. Martin, Y. Lu, and H. Cao. 2005. Modied anthrax fusion proteins deliver HIV antigens through MHC class I and II pathways. Vaccine 23: 4128 4135. 23. Maecker, H. T., H. S. Dunn, M. A. Suni, E. Khatamzas, C. J. Pitcher, T. Bunde, N. Persaud, W. Trigona, T. M. Fu, E. Sinclair, et al. 2001. Use of overlapping peptide mixtures as antigens for cytokine ow cytometry. J. Immunol. Methods 255: 27 40. 24. Elrefaei, M., N. El-Sheikh, K. Kamal, and H. Cao. 2003. HCV-specic CD27 CD28 memory T cells are depleted in hepatitis C virus and Schistosoma mansoni co-infection. Immunology 110: 513518. 25. McElroy, M. D., M. Elrefaei, N. Jones, F. Ssali, P. Mugyenyi, B. Barugahare, and H. Cao. 2005. Coinfection with Schistosoma mansoni is associated with decreased HIV-specic cytolysis and increased IL-10 production. J. Immunol. 174: 5119 5123. 26. Betts, M. R., J. M. Brenchley, D. A. Price, S. C. De Rosa, D. C. Douek, M. Roederer, and R. A. Koup. 2003. Sensitive and viable identication of antigen-specic CD8 T cells by a ow cytometric assay for degranulation. J. Immunol. Methods 281: 6578. 27. Barry, M., and R. C. Bleackley. 2002. Cytotoxic T lymphocytes: all roads lead to death. Nat. Rev. Immunol. 2: 401 409. 28. Betts, M. R., D. A. Price, J. M. Brenchley, K. Lore, F. J. Guenaga, A. Smed-Sorensen, D. R. Ambrozak, S. A. Migueles, M. Connors, M. Roederer, et al. 2004. The functional prole of primary human antiviral CD8 T cell effector activity is dictated by cognate peptide concentration. J. Immunol. 172: 6407 6417. 29. Boettler, T., H. C. Spangenberg, C. Neumann-Haefelin, E. Panther, S. Urbani, C. Ferrari, H. E. Blum, F. von Weizsacker, and R. Thimme. 2005. T cells with a CD4CD25 regulatory phenotype suppress in vitro proliferation of virusspecic CD8 T cells during chronic hepatitis C virus infection. J. Virol. 79: 7860 7867.

cell-cell contact. We postulate that this suppression is mediated by similar mechanisms that were described previously in other nonHIV regulatory T cell populations. Suppression may occur through direct binding of inhibitory cell surface molecules such as CTLA-4 to costimulatory molecules (including CD80 and CD86) on effector T cells (43, 44). These HIV-specic suppressor CD8 T cells may also recognize MHC peptide complexes on the cell surface of APCs, triggering the up-regulation of the inhibitory receptors Iglike transcripts (ILT)3 and ILT4 and the down-regulation of costimulatory molecules rendering the APC tolerogenic (33, 45, 46). In addition, programmed death-1 ligation has been suggested as a suppressor mechanism of regulatory T cells resulting in decreased T cell proliferation and IL-2 production (47, 48) and may also be involved. Our current ndings do not rule out a concomitant suppressive effect of IL-10 on cytolysis and IL-2 production. We have previously observed that rIL-10 suppressed in vitro HIV-specic CD107a/b expression (20). IL-10 has been shown to inhibit IL-2 production by CD4 T cells (49) and to induce the up-regulation of ILT4 expression on monocytes and dendritic cells from HIVinfected individuals (50). Suppressor CD8 T cells have been shown to decrease the expression of MHC class I on target cells through IL-10 secretion (3). Suppression by the HIV-specic IL10-positive CD8 T cells may involve the down-regulation of HLA expression on APC and reduce the effectiveness of Ag presentation. It is likely that these IL-10-positive CD8 T cells exert their suppressive function in vivo by multiple mechanisms that may be complementary or independent. Gag-specic IL-10-positive CD8 T cells may produce other cytokines that contribute to the observed suppressive effect. Of particular interest is the role of TGF- that was previously shown to trigger the up-regulation of CTLA-4 expression (5153) and inhibit T cell proliferation and IL-2 production (54). In this study, we identify a novel population of HIV-specic IL-10-positive CD8 T cells that appear to suppress multiple HIVand non-HIV-specic T cell functions and likely play an important role in HIV-associated immune dysfunction. The presence of these suppressor IL-10-positive CD8 T cells may inuence the maintenance of T cell responses and the effectiveness of T cell-based vaccines in HIV-infected individuals with advanced disease. This suppressor IL-10-positive CD8 T cell population is likely distinct from the previously described effector Tc2 cells (reviewed in Ref. 55). Tc2 cells were shown to be cytotoxic via the perforin and Fas pathways and produce mainly IL-4, IL-5, IL-10, and TGF- (56). Understanding the precise mechanisms by which these HIV-specic IL-10-positive CD8 T cells exert their suppressive functions will elucidate their contribution to the impaired immune system in HIV infection.

Disclosures

The authors have no nancial conict of interest.
1. Cortesini, R., J. LeMaoult, R. Ciubotariu, and N. S. Cortesini. 2001. CD8CD28 T suppressor cells and the induction of antigen-specic, antigenpresenting cell-mediated suppression of Th reactivity. Immunol. Rev. 182: 201206. 2. Filaci, G., M. Fravega, D. Fenoglio, M. Rizzi, S. Negrini, R. Viggiani, and F. Indiveri. 2004. Non-antigen specic CD8 T suppressor lymphocytes. Clin. Exp. Med. 4: 86 92. 3. Filaci, G., M. Fravega, S. Negrini, F. Procopio, D. Fenoglio, M. Rizzi, S. Brenci, P. Contini, D. Olive, M. Ghio, et al. 2004. Nonantigen specic CD8 T suppressor lymphocytes originate from CD8CD28 T cells and inhibit both T-cell proliferation and CTL function. Hum. Immunol. 65: 142156. 4. Jarvis, L. B., M. K. Matyszak, R. C. Duggleby, J. C. Goodall, F. C. Hall, and J. S. Gaston. 2005. Autoreactive human peripheral blood CD8 T cells with a regulatory phenotype and function. Eur. J. Immunol. 35: 2896 2908.
30. Thornton, A. M., and E. M. Shevach. 1998. CD4CD25 immunoregulatory T cells suppress polyclonal T cell activation in vitro by inhibiting interleukin 2 production. J. Exp. Med. 188: 287296. 31. Balashov, K. E., S. J. Khoury, D. A. Haer, and H. L. Weiner. 1995. Inhibition of T cell responses by activated human CD8 T cells is mediated by interferon- and is defective in chronic progressive multiple sclerosis. J. Clin. Invest. 95: 27112719. 32. Filaci, G., S. Bacilieri, M. Fravega, M. Monetti, P. Contini, M. Ghio, M. Setti, F. Puppo, and F. Indiveri. 2001. Impairment of CD8 T suppressor cell function in patients with active systemic lupus erythematosus. J. Immunol. 166: 6452 6457. 33. Manavalan, J. S., S. Kim-Schulze, L. Scotto, A. J. Naiyer, G. Vlad, P. C. Colombo, C. Marboe, D. Mancini, R. Cortesini, and N. Suciu-Foca. 2004. Alloantigen specic CD8CD28 FOXP3 T suppressor cells induce ILT3 ILT4 tolerogenic endothelial cells, inhibiting alloreactivity. Int. Immunol. 16: 10551068. 34. Scotto, L., A. J. Naiyer, S. Galluzzo, P. Rossi, J. S. Manavalan, S. Kim-Schulze, J. Fang, R. D. Favera, R. Cortesini, and N. Suciu-Foca. 2004. Overlap between molecular markers expressed by naturally occurring CD4CD25 regulatory T cells and antigen specic CD4CD25 and CD8CD28 T suppressor cells. Hum. Immunol. 65: 12971306. 35. Bisikirska, B., J. Colgan, J. Luban, J. A. Bluestone, and K. C. Herold. 2005. TCR stimulation with modied anti-CD3 mAb expands CD8 T cell population and induces CD8CD25 Tregs. J. Clin. Invest. 115: 2904 2913. 36. Cosmi, L., F. Liotta, E. Lazzeri, M. Francalanci, R. Angeli, B. Mazzinghi, V. Santarlasci, R. Manetti, V. Vanini, P. Romagnani, et al. 2003. Human CD8CD25 thymocytes share phenotypic and functional features with CD4CD25 regulatory thymocytes. Blood 102: 4107 4114. 37. Nakamura, K., A. Kitani, I. Fuss, A. Pedersen, N. Harada, H. Nawata, and W. Strober. 2004. TGF- 1 plays an important role in the mechanism of CD4CD25 regulatory T cell activity in both humans and mice. J. Immunol. 172: 834 842. 38. Rao, P. E., A. L. Petrone, and P. D. Ponath. 2005. Differentiation and expansion of T cells with regulatory function from human peripheral lymphocytes by stimulation in the presence of TGF-. J. Immunol. 174: 1446 1455. 39. Allez, M., J. Brimnes, I. Dotan, and L. Mayer. 2002. Expansion of CD8 T cells with regulatory function after interaction with intestinal epithelial cells. Gastroenterology 123: 1516 1526. 40. Uss, E., A. T. Rowshani, B. Hooibrink, N. M. Lardy, R. A. van Lier, and I. J. Ten Berge. 2006. CD103 is a marker for alloantigen-induced regulatory CD8 T cells. J. Immunol. 177: 27752783. 41. Rouse, B. T., P. P. Sarangi, and S. Suvas. 2006. Regulatory T cells in virus infections. Immunol. Rev. 212: 272286. 42. Thornton, A. M., and E. M. Shevach. 2000. Suppressor effector function of CD4CD25 immunoregulatory T cells is antigen nonspecic. J. Immunol. 164: 183190.

43. Manzotti, C. N., H. Tipping, L. C. Perry, K. I. Mead, P. J. Blair, Y. Zheng, and D. M. Sansom. 2002. Inhibition of human T cell proliferation by CTLA-4 utilizes CD80 and requires CD25 regulatory T cells. Eur. J. Immunol. 32: 2888 2896. 44. Zheng, Y., C. N. Manzotti, M. Liu, F. Burke, K. I. Mead, and D. M. Sansom. 2004. CD86 and CD80 differentially modulate the suppressive function of human regulatory T cells. J. Immunol. 172: 2778 2784. 45. Manavalan, J. S., P. C. Rossi, G. Vlad, F. Piazza, A. Yarilina, R. Cortesini, D. Mancini, and N. Suciu-Foca. 2003. High expression of ILT3 and ILT4 is a general feature of tolerogenic dendritic cells. Transplant Immunol. 11: 245258. 46. Chang, C. C., R. Ciubotariu, J. S. Manavalan, J. Yuan, A. I. Colovai, F. Piazza, S. Lederman, M. Colonna, R. Cortesini, R. Dalla-Favera, and N. Suciu-Foca. 2002. Tolerization of dendritic cells by T(S) cells: the crucial role of inhibitory receptors ILT3 and ILT4. Nat. Immunol. 3: 237243. 47. Nishimura, H., M. Nose, H. Hiai, N. Minato, and T. Honjo. 1999. Development of lupus-like autoimmune diseases by disruption of the PD-1 gene encoding an ITIM motif-carrying immunoreceptor. Immunity 11: 141151. 48. Carter, L., L. A. Fouser, J. Jussif, L. Fitz, B. Deng, C. R. Wood, M. Collins, T. Honjo, G. J. Freeman, and B. M. Carreno. 2002. PD-1:PD-L inhibitory pathway affects both CD4 and CD8 T cells and is overcome by IL-2. Eur. J. Immunol. 32: 634 643. 49. Moore, K. W., R. de Waal Malefyt, R. L. Coffman, and A. OGarra. 2001. Interleukin-10 and the interleukin-10 receptor. Annu. Rev. Immunol. 19: 683765. 50. Vlad, G., F. Piazza, A. Colovai, R. Cortesini, F. Della Pietra, N. Suciu-Foca, and J. S. Manavalan. 2003. Interleukin-10 induces the upregulation of the inhibitory receptor ILT4 in monocytes from HIV positive individuals. Hum. Immunol. 64: 483 489. 51. Park, H. B., D. J. Paik, E. Jang, S. Hong, and J. Youn. 2004. Acquisition of anergic and suppressive activities in transforming growth factor--costimulated CD4CD25 T cells. Int. Immunol. 16: 12031213. 52. Zheng, S. G., J. D. Gray, K. Ohtsuka, S. Yamagiwa, and D. A. Horwitz. 2002. Generation ex vivo of TGF--producing regulatory T cells from CD4CD25 precursors. J. Immunol. 169: 4183 4189. 53. Chen, W., W. Jin, N. Hardegen, K. J. Lei, L. Li, N. Marinos, G. McGrady, and S. M. Wahl. 2003. Conversion of peripheral CD4CD25 naive T cells to CD4CD25 regulatory T cells by TGF- induction of transcription factor Foxp3. J. Exp. Med. 198: 18751886. 54. Sung, J. L., J. T. Lin, and J. D. Gorham. 2003. CD28 co-stimulation regulates the effect of transforming growth factor-1 on the proliferation of naive CD4 T cells. Int. Immunopharmacol. 3: 233245. 55. Mosmann, T. R., L. Li, and S. Sa. 1997. Functions of CD8 T-cell subsets secreting different cytokine patterns. Semin. Immunol. 9: 8792. 56. Vukmanovic-Stejic, M., B. Vyas, P. Gorak-Stolinska, A. Noble, and D. M. Kemeny. 2000. Human Tc1 and Tc2/Tc0 CD8 T-cell clones display distinct cell surface and functional phenotypes. Blood 95: 231230.