A subset of CD8+ T cells acquiring selective suppressive properties may play a role in GvHD management

Transplant Immunology 28 (2013) 57–61

Contents lists available at SciVerse ScienceDirect

Transplant Immunology journal homepage: www.elsevier.com/locate/trim

A subset of CD8 + T cells acquiring selective suppressive properties may play a role in GvHD management Irit Avivi a, b, Dina Stroopinsky b, Jacob M. Rowe a, b, Tamar Katz a, b,⁎ a b

Department of Hematology & Bone Marrow Transplantation, Rambam Health Care Campus, Haifa, Israel Bruce Rappaport Faculty of Medicine, Technion, Israel Institute of Technology, Haifa, Israel

a r t i c l e

i n f o

Article history: Received 10 September 2012 Received in revised form 14 November 2012 Accepted 19 November 2012 Keywords: Regulatory CD8+ T cells Graft-versus-host disease Selective suppression

a b s t r a c t Difficulty in segregating graft-versus-tumor effect (GvT) from graft-versus-host disease (GvHD) remains a major limitation of allogeneic stem cell transplantation (Allo SCT). Naturally occurring regulatory T cells have been suggested to suppress alloreactive T cells involved in GvHD; however, their non-selective suppressive effect raises concern regarding probable attenuation of the GvT effect. Recent studies suggested inducible CD8 (iCD8) cells to be useful in suppressing autoimmune reactions, although their function in the Allo SCT setting has not been fully explored. The current study assessed in-vitro the properties of iCD8 T cells, generated in response to allogeneic dendritic cells (DCs), imitating the Allo SCT conditions. CD25− peripheral blood mononuclear cells (PBMCs) were stimulated with allogeneic DCs in mixed lymphocyte culture (MLC). The resultant iCD8+CD25+ population was isolated and assessed for phenotypic markers, cytokine expression profile, cell proliferation, inhibitory capacity and anti-viral response. The generated CD8+CD25+FOXP3+ T cells selectively inhibited the primary allogeneic response, without attenuating T cell response against other stimuli, such as mitogens or a cytomegalovirus (CMV) recall antigen. In conclusion, iCD8+CD25+ cells suppress allogeneic stimulation, while maintaining the capacity to respond to infectious pathogens. These cells could be potentially efficient in the Allo SCT setting, where GvHD prevention is required. © 2012 Elsevier B.V. All rights reserved.

1. Introduction The anti-tumor effect obtained with Allo SCT is often offset by a high mortality rate, caused by a simultaneous GvHD. Generation of an immunosuppressive cell population that selectively inhibits GvHD without adversely affecting post-transplant immune reconstitution and anti-tumor immunity remains a major challenge for improving patient outcome. Naturally occurring CD4 regulatory T cells (nTregs) have been suggested to oppose GvHD [1,2]. However, their non-selective inhibitory effect [3] raises the question whether their administration would also inhibit GvT activity, mediated through the induction of response targeted against antigens shared by tumor and healthy cells, tumor-specific antigens, or over-expressed self-antigens (proteinase-3, WT-1) [4].

Abbreviations: Allo SCT, allogeneic stem cell transplantation; CMV, cytomegalovirus; DCs, dendritic cells; EAE, experimental autoimmune encephalomyelitis; GvHD, graft-versus-host disease; GvT, graft-versus-tumor; iCD8, inducible CD8; IDDM, insulin dependent diabetes mellitus; IRB, Institutional Review Board; mAb, monoclonal antibodies; MLC, mixed lymphocyte culture; nTregs, naturally occurring CD4 regulatory T cells; PBMCs, peripheral blood mononuclear cells. ⁎ Corresponding author at: Department of Hematology and Bone Marrow Transplantation, Rambam Medical Center, P.O. Box 9602, Haifa 31096, Israel. Tel.: + 972 4 854 2541; fax: + 972 4 854 2343. E-mail address: [email protected] (T. Katz). 0966-3274/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.trim.2012.11.006

Regulatory CD8 + T cells, both expanded and inducible, possessing suppressive function, were mainly explored in the cancerous [5] and chronic-infection settings [6]. The abundance of these cells in involved tissues proved to be associated with a worse prognosis, given their ability to promote tumor progression and infection spread by suppressing an effective immune response [5,6]. Conversely, the employment of suppressive capacities of regulatory CD8 cells for treating T cell-dependent autoimmune disorders such as multiple sclerosis (or its non-humanized version; experimental autoimmune encephalomyelitis/EAE) and insulin dependent diabetes mellitus (IDDM) appears promising, leading to inhibition of T cell-related autoimmune inflammatory responses [7,8]. Characteristics and potential applicability of CD8 regulatory T cells for preventing GvHD, a life-threatening complication mediated through donor's T cell activation by host's DCs, have not been fully elucidated [9–11].

2. Objective The purpose of the current study was to explore in-vitro the profile and significance of induced CD8 + T cells, generated in response to allogeneic DCs, aiming to mimic the allogeneic stem cell transplant setting, where selective inhibition of GvHD without attenuating the GVT response is required.

I. Avivi et al. / Transplant Immunology 28 (2013) 57–61

100 101 102 103 104

CTLA-4

100

FOXP3

101

102

100 101 102 103 104

100

50%

100

103

104

30%

101

102

103

104

81%

101

102

CD4

103

104

100

100

101

102

100

100

103

104

68%

101

102

103

104

77%

101

102

103

104

60%

101

102

103

104

100 101 102 103 104

100 101 102 103 104

CD25 104

104

100

101

102 FL4-H

I 100 101 102 103 104

103

103

0%

100

100

100

103

104

4%

101

102 FL4-H

100 101 102 103 104

102

CD8

102

95%

Induced CD8+CD25-

100 101 102 103 104

101

GITR

100

100 101 102 103 104

102

103 104

101

100 101 102 103 104

103

100

90%

100 101 102 103 104

102

CD8

Induced CD8+CD25+

Expanded nTregs

100 101 102 103 104

101

100

100

100

20%

101

102

CD25

103

0%

101

CD25

104

B

104

A

100 101 102 103 104

58

103

104

5%

101

102 FL4-H

103

104

3%

101

102

103

104

CD8

Fig. 1. Induced CD8+CD25+ T cells express regulatory markers. CD25 depleted PBMCs were stimulated with irradiated DCs for 6 days. (A) Induction of CD8+CD25+ T cells was assessed with flow cytometry using PE-anti-CD25 and FITC-anti-CD8 monoclonal antibodies. The dot plot represents one of the 7 experiments prior to (left panel) and following (right panel) induction. (B) ICD8+CD25+ and CD8+CD25− T cells were isolated and stained with FITC-anti-GITR, FITC-anti-CTLA-4 or intracellular FITC-anti-FOXP3 as shown in the representative FACS dot plot (n = 5). Expanded nTregs were induced and analyzed in the same experiment.

3. Material and methods 3.1. Human samples Peripheral blood mononuclear cells (PBMCs) obtained from healthy donors (IRB approval number 2419) were isolated by centrifugation over Ficoll–Hypaque gradients (Sigma-Aldrich, St Louis, MO, USA). All experiments were performed using PBMCs obtained from different unrelated donors.

3.2. Flow cytometery and intracellular staining Cells were stained with the following monoclonal antibodies (mAb): CD25, CD4, CD8, CD3, and CD45 (BD Biosciences, San Jose, CA); GITR and CTLA-4 (R&D systems, Minneapolis, Minnesota, USA). The antibodies were conjugated to FITC, PE, PerCP or APC. For intracellular staining, cells were activated with 40 ng/ml PMA and 1 μg/ml ionomycin (Sigma-Aldrich) and treated with 2 μM/ml GolgiStop (BD Bioscience). Upon staining for surface markers, cells were permeabilized (cytofix-cytoperm kit, BD Bioscience) and incubated with FITC anti-FoxP3 (clone PCH101, eBioscience) and PE conjugated anti-cytokine mAb: IL-2, IFN-γ, IL-10 (BD Pharmigen) or TGF-β1 (clone TB21, IQ products Groningen, The Netherlands). Necrotic cells were detected by the addition of 0.1 mg/ml propidium iodide (PI) (Sigma-Aldrich). Apoptotic cells were detected by Annexin V (APC) apoptosis detection kit (eBioscience).

Flow cytometry analysis was done using the CellQeust software on a fluorescence-activated cell sorting (FACS) — Calibur instrument (BD Bioscience). 3.3. Dendritic cell generation Monocyte-derived DCs were generated from adherent PBMCs cultured with 1% AB plasma (Tel Hashomer Blood Bank, Tel-Aviv, Israel), 1000 U/ml granulocyte-macrophage colony stimulating factor (GM-CSF) and 500 U/ml interleukin-4 (IL-4) (R&D systems, Minneapolis, Minnesota, USA) for 5 days. Maturation was induced by addition of 1000 U/ml tumor necrosis factor α (TNF-α), 300 U/ml IL-1β, 1000 U/ml IL-6 (R&D systems), and 1 μg/ml prostaglandin E2 (PGE-2) (Sigma-Aldrich) for 48 h. 3.4. Allo-stimulation and purification of cell populations CD25 − PBMCs generated by negative selection using anti-CD25 magnetic beads (Miltenenyi Biotech) were stimulated with mature irradiated DCs (7500 cGy, Gammacell 300) in the mixed lymphocyte culture (MLC) containing 10% AB plasma for 5 days. After stimulation, CD8 + T cells were purified by negative selection using the untouched CD8 + T cell isolation kit (Miltenyi Biotech). CD8+CD25+ cells were isolated by positive selection, using anti-CD25 magnetic beads. Naturally occurring CD4+CD25+ T (nTregs) cells were expanded in parallel with DCs. For functional analysis, expanded nTregs

I. Avivi et al. / Transplant Immunology 28 (2013) 57–61

*

* PBMCs

B

1.4 1.2

*

1 0.8

*

0.6 0.4 0.2 0

PBMCs iCD8+C25- iCD8+CD25- nTregs

PBMCs non stimulated

+Induced iCD8

D

PBMCs + DCs

1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

PBMCs non stimulated PBMCs + pp65-CMV

PBMC s 1.6%

3.6%

PBMCs + DCs 3%

2.5%

+Expanded - nTregs

PBMCs + DCs + iCD8+CD25+ 1%

1%

* 10.7%

2.6%

2.7%

PI

Proliferation (O.D. 450 nm)

C

- DCs + DCs

1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

Proliferation (O.D. 450 nm)

Proliferation (O.D. 450 nm)

A

59

Annexin V

1:5

1:1

11:1

+Induced +Expanded iCD8+CD25+ nTregs Fig. 2. Induced CD8+CD25+ T cells selectively suppress response against original stimuli. CD25 depleted PBMCs were stimulated with irradiated DCs for 6 days. ICD8+CD25+ T cells were isolated and functional tests were performed: (A) ICD8+CD25+ and iCD8+CD25− T-cells underwent secondary MLC with irradiated DCs from the original stimulator. Proliferation rate was determined with the use of tetrazolium salt. Data (mean ± SD) show mean O.D. values derived from 5 independent experiments. Proliferation rate of iCD8+CD25+ and iCD8+CD25− T cell fractions was significantly different (P b .05) as indicated by asterisks. (B) Irradiated iCD8+CD25+ T cells were added at 1:5 or 1:1 ratio to responder PBMCs, stimulated by original allo-DCs for 5 days. Expanded nTregs (after sorting) were added at 1:1 ratio as a control. Data (mean ± SD) show means O.D. values derived from 4 independent experiments. (C) Irradiated iCD8+CD25+ T cells were added at 1:1 ratio to responder PBMCs, stimulated by pp65-CMV recombinant protein for 4 days. Expanded nTregs (after sorting) were added as a control. Data (mean ± SD) show mean O.D. values derived from 4 independent experiments. Suppression rate following addition of iCD8+CD25+ cells versus nTregs was significantly different (P b .05), as indicated by asterisks. (D) Un-stimulated PBMCs, co-cultures of PBMCs and original allo-DCs with and without addition of iCD8+CD25+ cells, were subjected to necrosis and apoptosis analysis as demonstrated by PI (PE) and Annexin V (APC) staining. Representative dot plots are shown (n = 3).

were sorted to obtain CD25high subpopulation (FACSAria™ Cell Sorter, BD, Heidelberg, Germany). All generated cell fractions demonstrated >90% purity, as confirmed by flow cytometry.

affecting their function. Upon 4 days of incubation in the plasma-free medium, cells were pulsed for the last 3 h with tetrazolium salt (EZ4U kit, Biomedica, Austria). Color absorbance, proportional to proliferation rate, was measured using a microplate-reader set at 450 nm, with 620 nm as a reference.

3.5. Response to CMV antigens iCD8 +CD25 + cells generated from CMV-positive donors were incubated with 20 μl/ml of pp65-CMV recombinant protein in presence of 2 μM/ml of GolgiStop (BD) for 6 h. CMV specific cells were measured by intracellular staining with PE-anti-IFN-γ (BD Pharmigen).

3.8. Statistical analysis Data of the tested groups were compared using a Student's t-test. P-values lower than 0.05 were considered significant. 4. Results

3.6. Granzyme B production

4.1. Induction of CD8+CD25+ FoxP3+ T cells following DC stimulation +

−

+

+

Induced CD8 CD25 and iCD8 CD25 T cells were incubated for 48 h with PHA-P (Sigma-Aldrich, St Louis, MO), and then co-stained with surface markers, fixed, permeabilized, and assessed for intracellular Granzyme B expression, using PE-anti-Granzyme B (BD Pharmigen). 3.7. T-cell proliferation and suppression To measure their proliferative capacity, generated T-cells were co-cultured with irradiated immature DCs (1:1 and 5:1 ratios) derived from the original stimulator. Suppressive capacity was evaluated by addition of irradiated tested populations to original MLC (CD25− PBMCs+ DCs). Irradiation was performed in order to annul a potential, though minimal, proliferation of the tested populations, without

The current study investigated the nature of iCD8+CD25+ T cells generated upon allogeneic stimulation with highly potent mature DCs, aiming to induce a marked suppressive T cell population. Stimulation of CD25− T cells with allogeneic mature DCs for 3 days resulted in the induction of CD8+CD25high T cells, accounting for approximately 5% of CD8+ cell population. A longer stimulation, approaching 6 days, yielded a greater population of CD8+CD25high T cells, comprising about 20% of CD8+ cell population (Fig. 1A; n= 7). Further characterization revealed these cells to concurrently express additional regulatory markers: 65% for GITR, 70% for CTLA-4 and 58% for FOXP3 (n = 5) (Fig. 1B). iCD8+CD25+ T cells exhibited the regulatory phenotype, similar to that observed in expanded nTregs, with mean levels of FOXP3, GITR and CTLA-4 of 63%, 70% and 78%, respectively (n=5). In contrast, the expanded CD8+CD25− cells (representing 80% of the allo-stimulated CD8+ population) did not express regulatory markers. Fig. 1B demonstrates the mean levels of regulatory markers, exhibited by iCD8+CD25+ T cells, their CD8+CD25− allo-induced counterparts and by expanded nTregs.

60

I. Avivi et al. / Transplant Immunology 28 (2013) 57–61

Fig. 3. Induced CD8+CD25+ T cells exhibit inflammatory properties. CD25 depleted PBMCs were stimulated with allogeneic DCs. (A) ICD8+CD25+ were isolated on day 6 of culture. ICD8+CD25+ T cells and expanded nTregs were incubated with PMA, inomycin and GolgiStop for 5 h and stained with PE anti-IFN-γ, anti-IL-2 and FITC-anti-FOXP3. A summary of 3 independent experiments showing the mean ± SD co-expression of inflammatory cytokines and FOXP3 is presented. Expression of inflammatory cytokines by expanded nTregs significantly differed from that of iCD8+CD25+ T cells (P b .05). (B) A representative FACS plot showing co-expression of FOXP3 and inflammatory cytokines (IFNγ and IL-2) on iCD8+CD25+ cells (n = 3). (C) On day 5 of culture, isolated iCD8+CD25+ T cells and expanded nTregs were incubated with pp65-CMV protein and GolgiStop for 6 h and stained with PE anti-IFN-γ and surface CD8 or CD4 as shown in the representative FACS dot plot (n = 3). (D) Isolated iCD8+CD25− and iCD8+CD25+ were incubated with PHA for 48 h and stained with PE-anti granzyme B and surface markers. Representative experiment is shown (n = 3).

Of note, all experiments were performed using pairs of stimulator and responder cells of unrelated donors, thereby, reducing the likelihood of shared HLA allo types between the involved cell populations.

4.2. iCD8+CD25+ T cells selectively suppress response against original stimuli We examined the potential proliferative and suppressive capacities, associated with the acquired regulatory phenotype exhibited by iCD8+CD25+ cells. In contrast to induced CD8+CD25− T cells, iCD8+CD25+ cells presented low proliferative capacity when stimulated by the original allogeneic DCs (Fig. 2A; n = 5). iCD8+CD25+ T cells when added to MLC, markedly suppressed T cell proliferation in response to the original DC stimulus. The suppression level was found to be dose dependent, with a mean suppression level reaching 60% at 1:1 ratio (Fig. 2B; n = 5). To investigate whether the suppression effect is selectively directed against T-cell responsiveness to the original stimulus, we looked at the capacity of CD8+CD25+ T-cells to attenuate the response to other stimuli, such as CMV. Remarkably, opposite to nTregs exhibiting non-selective inhibitory capacity against both original DCs and CMV antigen, iCD8+CD25+ cells, generated from CMV seropositive subjects, selectively suppressed T cell reaction against the original stimulus, without attenuating T cell responsiveness against CMV (Fig. 2C). The employment of IL-2 stimulated PBMCs did not alter the suppression capacity of iCD8 cells (data not shown). Notably, staining for both PI and Annexin V was negative, indicating that addition of iCD8+CD25+ T cells could lead to a true reduction in T cell proliferation, rather than to T cell death (Fig. 2D). 4.3. iCD8+CD25+ T cells exhibit inflammatory properties iCD8+CD25+ T cells, stimulated with PMA and ionomycin, failed to express suppressive factors such as IL-10 and TGF-β, with levels approaching 0.2% and 0.5%, respectively (data not shown). However, iCD8+CD25+ cells exhibited inflammatory cytokines, including IL-2 and IFN-γ, with mean levels of 21% and 50%, respectively (Fig. 3A; n=3). Of note, a large number of iCD8+CD25+ cells highly expressing inflammatory cytokines, have co-expressed FOXP3+ (Fig. 3B; n=3); suggesting that they represent a unique cell population,

characterized by dual inhibitory and effector properties. In contrast, nTregs exposed to the same stimulation conditions failed to express IL-2 and IFN-γ. In accordance with this “pro-inflammatory” profile, iCD8+CD25+ T cells obtained from a CMV sero-positive donor, responded to stimulation with pp65 CMV recombinant protein, as reflected by increased IFN-γ excretion, opposite to nTegs, which failed to show such activity (Fig. 3C; n = 3). Further evaluation of the effector cytotoxic capacities of the iCD8 cells revealed their ability to produce granzyme B in response to stimulation with PHA for 48 h. The mean level of granzyme B production by the iCD8+CD25+ T cells approached 80%, as opposed to a 3-fold lower production, measured for iCD8+CD25− cells (mean = 25% n = 3) (Fig. 3D).

5. Discussion Acute GvHD, involving activation of host DCs by damaged tissues [12], followed by stimulation of donor CD4 and CD8 T cells, results in a local tissue damage, mainly involving skin, gut and liver [13]. While the excess of regulatory iCD8 cells is judged to be harmful in cancerous and infectious settings [5,6], it appears to be advantageous in the autoimmune scenario (e.g., IDDM, EAE, rheumatoid arthritis), where both expanded and induced CD8+ regulatory T cells were shown to potently inhibit T cell responses involved in the autoimmune process [7,8]. Regulatory iCD8 cells may be a potential target for immunological therapeutic intervention in the allogeneic GvHD setting. In the current study, stimulation of human CD8+CD25− T cells with mature allogeneic DCs, simulating the allogeneic GvHD setting, led to induction of a prominent CD8 +CD25+ cell population. Mature DCs were chosen over other allogeneic stimulators (i.e., PBMCs or immature DCs) due to their superior capacity to generate T cells activation, which in turn could lead to induction of inhibitory CD25+ T cells [14]. This unique cell population, induced without addition of exogenous cytokines (e.g., IL-2, TGF-β), was generated due to an exceptionally

I. Avivi et al. / Transplant Immunology 28 (2013) 57–61

potent stimulation of the T cell receptor by allogeneic mature DCs. Previous studies exploring iCD8+ cells generated in both mice and humans by Teplizumab, a humanized anti-CD3 antibody [15,16] or IL6 [17], showed iCD8+ cells to express regulatory phenotypic markers including FOXP3. Consistent with their regulatory “nTreg-like” phenotype, demonstrated in our study, iCD8 +CD25 + cells failed to proliferate in response to the original allogeneic DCs (though described to restore this capacity in the presence of specific experimental conditions) [18], and markedly suppressed T cell proliferation evoked by original stimuli, confirming their inhibitory function. Exogenous administration of IL-2, previously reported to slightly enhance suppressive activity of regulatory cells [19,20], did not alter iCD8 inhibitory effect, reflecting differences in iCD8 populations generated with different stimuli. A recent publication demonstrated in mice that the induction of highly suppressive gut iCD8+ T cells in response to gut-specific antigens maintains the intestinal homeostasis by down-modulating effector T cell function [21]. These findings support our hypothesis regarding the potential induction of allo-stimulated iCD8 cells, which could be used to inhibit effector T cells involved in GvHD related damage. Remarkably, in contrast to CD4 + nTregs exhibiting a non-specific T cell inhibition capacity, iCD8 were found to suppress T cell activity against the original stimuli only, preserving T cell responsiveness against recall antigens such as CMV. This novel finding of a unique suppressive capacity provides a potential approach to selective inhibition of undesirable GvHD, whilst preserving T cell activity against life threatening infectious pathogens and residual tumor cells. Previous studies, inducing iCD8+CD25+ by cytokines or anti-T cell MoAbs, demonstrated non-antigen specific inhibition [15,17], making these cells less suitable for GvHD prevention or treatment. Similar to other inducible suppressive cell populations (either CD4 or CD8 T cells) [15,17], allo-iCD8 cells obtained in our study, showed a high expression of INF-γ in response to mitogens. Interestingly, the inhibitory iCD8 population described in the current study, showed a high FOXP3 expression along with a marked production of INF-γ and IL-2 in response to mitogens, similarly to induced CD4+CD25+ cells [14]. Furthermore, stimulation with CMV peptide induced an INF-γ excretion, reflecting a true residual effector capacity, retained from their paternal CD8+CD25− cells. In conclusion, iCD8 cells, a prominent population generated by allo-DC stimulation, possess unique properties of selective allosuppressive capacity, while retaining anti-infections responsiveness. Such characteristics make these cells potentially valuable for the management of GvHD. However, further in-vivo studies are warranted to prove their employment in this setting. Author contribution statement Irit Avivi — designed research and wrote the paper. Dina Stroopinsky — designed and performed research, and wrote the paper. Jacob M. Rowe — wrote the paper. Tamar Katz — designed and performed research, and wrote the paper.

61

Conflict of interest statement There are no conflicts to declare. Acknowledgment The authors wish to acknowledge with thanks the assistance of Sonia Kamenetsky in the preparation of the manuscript. References [1] Sakaguchi S, Sakaguchi N, Shimizu J, Yamazaki S, Sakihama T, Itoh M, et al. Immunologic tolerance maintained by CD25+ CD4+ regulatory T cells: their common role in controlling autoimmunity, tumor immunity, and transplantation tolerance. Immunol Rev 2001;182:18–32. [2] Shevach EM. CD4+ CD25+ suppressor T cells: more questions than answers. Nat Rev Immunol 2002;2:389–400. [3] Thornton AM, Shevach EM. Suppressor effector function of CD4+CD25+ immunoregulatory T cells is antigen nonspecific. J Immunol 2000;164:183–90. [4] Brusic A, Wu CJ. Enhancing graft-versus-leukemia after transplant: the rise of anti-cancer vaccines. Front Biosci 2012;17:635–55. [5] Yang ZQ, Yang ZY, Zhang LD, Ping B, Wang SG, Ma KS, et al. Increased liver-infiltrating CD8+FoxP3+ regulatory T cells are associated with tumor stage in hepatocellular carcinoma patients. Hum Immunol 2010;71:1180–6. [6] Billerbeck E, Thimme R. CD8+ regulatory T cells in persistent human viral infections. Hum Immunol 2008;69:771–5. [7] Zozulya AL, Wiendl H. The role of CD8 suppressors versus destructors in autoimmune central nervous system inflammation. Hum Immunol 2008;69:797–804. [8] Chen ML, Yan BS, Kozoriz D, Weiner HL. Novel CD8+ Treg suppress EAE by TGF-beta- and IFN-gamma-dependent mechanisms. Eur J Immunol 2009;39: 3423–35. [9] Sawamukai N, Satake A, Schmidt AM, Lamborn IT, Ojha P, Tanaka Y, et al. Cell autonomous role of TGFbeta and IL-2 receptor in CD4+ and CD8+ inducible regulatory T cell generation during graft-versus-host disease. Blood 2012;119:5575–83. [10] Beres AJ, Haribhai D, Chadwick AC, Gonyo PJ, Williams CB, Drobyski WR. CD8+ Foxp3+ regulatory T cells are induced during graft-versus-host disease and mitigate disease severity. J Immunol 2012;189:464–74. [11] Funatake CJ, Marshall NB, Kerkvliet NI. 2,3,7,8-Tetrachlorodibenzo-p-dioxin alters the differentiation of alloreactive CD8+ T cells toward a regulatory T cell phenotype by a mechanism that is dependent on aryl hydrocarbon receptor in CD4+ T cells. J Immunotoxicol 2008;5:81–91. [12] Duffner UA, Maeda Y, Cooke KR, Reddy P, Ordemann R, Liu C, et al. Host dendritic cells alone are sufficient to initiate acute graft-versus-host disease. J Immunol 2004;172:7393–8. [13] Choi SW, Levine JE, Ferrara JL. Pathogenesis and management of graft-versus-host disease. Immunol Allergy Clin North Am 2010;30:75–101. [14] Stroopinsky D, Avivi I, Rowe JM, Avigan D, Katz T. Allogeneic induced human FOXP3+IFN-gamma+ T cells exhibit selective suppressive capacity. Eur J Immunol 2009;39:2703–15. [15] Ablamunits V, Bisikirska B, Herold KC. Acquisition of regulatory function by human CD8+ T cells treated with anti-CD3 antibody requires TNF. Eur J Immunol 2010;40: 2891–901. [16] Bisikirska B, Colgan J, Luban J, Bluestone JA, Herold KC. TCR stimulation with modified anti-CD3 mAb expands CD8+ T cell population and induces CD8+CD25+ Tregs. J Clin Invest 2005;115:2904–13. [17] Nakagawa T, Tsuruoka M, Ogura H, Okuyama Y, Arima Y, Hirano T, et al. IL-6 positively regulates Foxp3+CD8+ T cells in vivo. Int Immunol 2010;22:129–39. [18] Robb RJ, Lineburg KE, Kuns RD, Wilson YA, Raffelt NC, Olver SD, et al. Identification and expansion of highly suppressive CD8+FoxP3+ regulatory T cells after experimental allogeneic bone marrow transplantation. Blood 2012;119:5898–908. [19] Popmihajlov Z, Smith KA. Negative feedback regulation of T cells via interleukin-2 and FOXP3 reciprocity. PLoS One 2008;3:e1581. [20] Ablamunits V, Bisikirska BC, Herold KC. Human regulatory CD8 T cells. Ann N Y Acad Sci 2008;1150:234–8. [21] Fleissner D, Hansen W, Geffers R, Buer J, Westendorf AM. Local induction of immunosuppressive CD8+ T cells in the gut-associated lymphoid tissues. PLoS One 2010;5:e15373.