Original Clinical Science

CD27low Natural Killer Cells Prolong Allograft Survival in Mice by Controlling Alloreactive CD8+ T Cells in a T-Bet–Dependent Manner Margareta Lantow, Elke Eggenhofer, Manije Sabet-Baktach, Philipp Renner, Jordi Rovira, Gudrun E. Koehl, Hans J. Schlitt, Edward K. Geissler, and Alexander Kroemer Background. Natural killer (NK) cells play a dichotomous role in alloimmune responses because they are known to promote both allograft survival and rejection. The aim of this study was to investigate the role of functionally distinct NK cell subsets in alloimmunity with the hypothesis that this dichotomy is explained by the functional heterogeneity of distinct NK cell subsets. Methods. Because T-bet controls the maturation of NK cells from CD27high to terminally differentiated CD27low NK cells, we used Rag−/−T-bet−/− mice that lack mature CD27low NK cells to study the distinct roles of CD27low versus CD27high NK cells in a model of T cell–mediated skin transplant rejection under costimulatory blockade conditions. Results. We found that T cell–reconstituted Rag1−/− recipients (possessing CD27low NK cells) show significantly prolonged allograft survival on costimulatory blockade when compared to Rag1−/−T-bet−/− mice (lacking CD27low NK cells), indicating that CD27low but not CD27high NK cells enhance allograft survival. Critically, Rag1−/−T-bet−/− recipients showed strikingly increased alloreactive memory CD8+ T cell responses, as indicated by increased CD8+ T cell proliferation and interferon-γ production. Therefore, we speculated that CD27low NK cells directly regulate alloreactive CD8+ Tcell responses under costimulatory blockade conditions. To test this, we adoptively transferred CD27low NK cells into Rag1−/−T-bet−/− skin transplant recipients and found that the CD27low NK cells restore better allograft survival by inhibiting the proliferation of alloreactive interferonγ+CD8+ T cells. Conclusions. In summary, mature CD27low NK cells promote allograft survival under costimulatory blockade conditions by regulating alloreactive memory CD8+ T-cell responses.

(Transplantation 2015;99: 391–399)

T

he role of natural killer (NK) cells in transplantation is dichotomous and poorly defined as indicated by their involvement in both allograft rejection1-3 and transplant tolerance.4-7 For instance, NK cells have been found in large Received 6 August 2014. Revision requested 2 October 2014. Accepted 22 October 2014. Department of Surgery, University Hospital Regensburg, University of Regensburg, Regensburg, Germany. This work was supported by the EU 7th Framework Programme (The ONE Study, grant agreement 260687; E.K.G, A.K) and the Deutsche Forschungsgemeinschaft (KFO 243/1, KR 3631/2-1, KR3631/2-2, A.K., E.K.G., H.J.S., G.E.K., and E.E.). The authors declare no conflicts of interest. M.L. performed and participated in designing the experiments, analyzed and interpreted data, and wrote the article. M.S-B, E.E., P.R., J.R. and G.E.K performed experiments and wrote the article. A.K., E.K.G, and H.J.S designed the experiments, provided funding, administrative and material support, interpreted data, and wrote the article. Correspondence: Alexander Kroemer, MD, Department of Surgery, University Hospital Regensburg, Franz-Josef-Strauß-Allee 11, 93042 Regensburg, Germany. ([email protected]) Supplemental digital content (SDC) is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal’s Web site (www.transplantjournal.com). Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0041-1337/15/9902-391 DOI: 10.1097/TP.0000000000000585

Transplantation



February 2015



Volume 99



Number 2

numbers in rejecting allografts8 and intensify T cell–dependent alloresponses by promoting the maturation of antigen-presenting cells, thereby indirectly amplifying T-cell activation.2,5 Additionally, NK cells activated by interleukin (IL)-15 develop potent effector functions that render them capable of directly rejecting allografts in the absence of adaptive immunity.5 Conversely, NK cells have also been shown to promote the induction of transplant tolerance under costimulatory blockade conditions.4-7 Importantly, NK cells in a naive state readily reject graftderived allogeneic cells, thus preventing the priming of alloreactive T cells.5,9 Moreover, NK cells have been shown to delay allograft rejection by reducing the homeostatic proliferation of memory CD8+ T cells by competing for IL-15.10 In line with this observation, it has been shown that the depletion of host NK cells in skin allograft recipients enhances the expansion of alloreactive Th2 cells.10,11 Therefore, the dichotomous role of NK cells could in part be explained by differences in their activation status.12 Moreover, we speculate that this dichotomy1-7,12 is caused by the phenotypic and functional heterogeneity of NK cells,13-16 which warrants a detailed investigation of functional NK cell subsets and their involvement in allograft rejection and tolerance.1-7,10-13,17 In mice, NK cells undergo a sequential 4-step maturation process, which is regulated by the T-box transcription factors T-bet and Eomesodermin (Eomes).18 T-bet controls not only the homeostatic maturation of NK cells13-15 but also the www.transplantjournal.com

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

391

392

Transplantation



February 2015



Volume 99



Number 2

expression of the NK cell maturation markers KLRG1 and CD2719 and is required for activation-induced maturation of NK cells during infection.20 T-bet deficiency in mice reduces the overall maturation status of the peripheral NK cell compartment and the expression of KLRG1.20 Based on different expression levels of CD27, CD11b, and KLRG1, NK cells can be phenotypically divided into subsets with distinct effector functions including cytotoxicity and cytokine production.13,15,17 More specifically, CD27highKLRG1lowCD11bhigh NK cells are less mature than their terminally differentiated CD27lowKLRG1highCD11bhigh counterparts and express high levels of Eomes. Moreover, CD27high NK cells are proinflammatory innate effector cells, given that they are potent producers of interferon (IFN)-γ and proliferate strongly in response to cytokines.10,16 CD27lowKLRG1highCD11bhigh NK cells, on the other hand, are terminally differentiated effector cells that are highly regulated by Ly49 surface receptors and critically depend on T-bet, regulating the peripheral homeostasis and differentiation of NK cells.18,21 In this study, we investigated the role of functionally and developmentally distinct NK cell subsets in allograft rejection. We studied Rag1−/− mice that possess both fully functional CD27low and CD27high NK cells5 versus Rag1−/−T-bet−/− mice that lack mature CD27low NK cells as transplant recipients in an adoptive transfer model of T cell–mediated allograft rejection. We found that CD27low NK cells play a crucial regulatory role in alloresponses by limiting the homeostatic proliferation and expansion of IFN-γ–producing memory CD8+ T cell responses under costimulatory blockade conditions in a T-bet–dependent manner.

www.transplantjournal.com

TABLE 1.

Flow cytometric antibodies Staining

Antibody

Clone

Company

Surface

CD4 CD8 B220 CD317 CD11b CD11c IgG IgG IgG2b K NKG2A/C/E NKp46 Ly49C/I/F/H CD27 KLRG1 CD3 CD44 CD62L CD25 IFN-γ IL-17A

RM4-5 53-6.7 RA3-6B2 eBio927 M1/70 HL3 eBio299Arm golden Syrian hamster eB149/10H5 20d5 29A1.4 14B11 LG.7 F9 2 F1 145-2C11 IM7 MEL-14 PC61.5 XMG1.2 eBio17B7

BD Pharmingen eBiosciences eBiosciences eBiosciences eBiosciences eBiosciences eBiosciences eBiosciences eBiosciences eBiosciences eBiosciences eBiosciences eBiosciences eBiosciences BD Pharmingen eBiosciences eBiosciences eBiosciences eBiosciences eBiosciences

Intracellular

Rag1−/−T-bet−/− mice were reconstituted intravenously with 2–3  106 CD27lowCD11bhigh NK cells one day before STx. Costimulatory Blockade

MATERIALS AND METHODS Mice

C57BL/6 wild-type (wt), BALB/c wt, Foxp3-EGFP (BALB/ c-Foxp3tm1Flv/J), Rag1−/− (BALB/c-Rag1tm1Mom/J) and T-bet−/−(BALB/c-Tbx21tm1Glm) mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and Charles River Laboratories (Sulzfeld, Germany). BALB/c.Rag1−/−T-bet−/− mice were obtained through crossbreeding of Rag1−/− with T-bet−/− mice. All animal experiments were approved by the local Institutional Animal Care and Use Committee and the regional authorities of Upper Palatinate, Germany. Skin Transplant

Segments of full-thickness tail skin (C57BL/6 to BALB/c background) were transplanted onto recipient mice at day 0 as described before.16,22 The survival of grafts was monitored by daily visual inspection, and graft rejection was defined as the complete loss of viable graft tissue.16,22 CD3+ Cell Purification and Adoptive Transfer

CD3+ T cells were MACS-sorted (MiltenyiBiotec, Germany) from the spleens and lymph nodes of Foxp3-EGFP mice according to the manufacturer's protocol. On day 8 after engraftment, skin transplant (STx) recipients were reconstituted intravenously with 2–3  106 CD3+ MACS-sorted T cells. NK Cell Purification and Transfer

CD27lowCD11bhigh NK cells were flow-sorted (LSR II, BD Biosciences, Germany) from Rag1−/− mice.

Anti-CTLA4Ig monoclonal antibody (mAb) (0.50 mg) and MR1 mAb (0.50 mg) (BioXCell, Lebanon, NH) were administered intraperitoneally on the day of adoptive cell transfer (day 8 after STx), and 0.25 mg of both antibodies were given intraperitoneally every 2 days until day 16 after STx. Flow Cytometry

At the time of graft rejection draining lymph nodes and spleen were harvested and processed for further analyses. For surface and intracellular staining, cell preparations were processed and stained with mAbs shown in supplement Table 1 as previously described.16,22 All antibodies and isotype-specific controls were purchased from eBiosciences or BD Pharmingen, Germany. Carboxyfluorescein Diacetate Succinimidyl Ester Staining and Proliferation Assay

The wt BALB/c splenocytes or wt C57BL/6 splenocytes were labeled with 1 μM carboxyfluorescein diacetate succinimidyl ester (CFSE, Invitrogen, Germany) and were MACS-sorted for CD3+ T cells before intravenous injection into Rag1−/− and Rag1−/−T-bet−/− recipients. T-cell proliferation was analyzed in lymph nodes and spleens by flow cytometry for CFSE after 3, 4, and 6 days. Enzyme-Linked Immunosorbent Assay

At the time of rejection serum levels of IFN-γ were determined using a commercially available enzyme-linked immunosorbent assay kit according to the manufacturer's instructions (BD Biosciences, Germany).

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Lantow et al

© 2015 Wolters Kluwer

Statistical Analysis

Graft survival was calculated via the Kaplan-Maier method and the log-rank test was used for graft survival comparison. All other statistical analyses were performed using a 2-tailed nonparametric Student t test (Mann-Whitney), or analysis of variance with Bonferroni multiple comparison tests when multiple groups were compared (GraphPad Prism5 Software San Diego, CA). Graphs show means ± SEMs and statistical significance. RESULTS Both Rag1 / and Rag1 / T-bet / Recipients Promptly Reject Skin Allografts on Adoptive T-Cell Reconstitution

We hypothesized that CD27high NK cells are proinflammatory effector cells that can accelerate T-cell–mediated allograft rejection, whereas terminally differentiated T-bet–dependent CD27low NK cells play a regulatory role in prolonging allograft survival. To test this idea, we used T-bet−/− mice on a Rag1−/− background that possess a large fraction of CD27high NK cells, but lack terminally differentiated CD27low NK cells (Figure 1A, Figure S1A, B, SDC, http://links.lww.com/TP/B136), in comparison to Rag1−/− mice that feature both fully functional CD27high and CD27low NK cell subsets; both strains were recipients of fully-mismatched skin allografts. We found that both Rag1−/− and Rag1−/−T-bet−/− mice show indefinite allograft survival, suggesting that T-bet-deficient CD27high NK cells per se do not cause rejection of skin allografts in the absence of adaptive immunity (Figure 1B). According to our theory, we anticipated that CD27high and T-bet–dependent CD27low NK cells indirectly regulate allograft rejection by controlling alloreactive T-cell responses. To address this, we reconstituted Rag1−/− and Rag1−/−T-bet−/− STx recipients with MACS-purified CD3+ T cells and determined allograft survival. Our results showed that both strains promptly reject skin allografts with similar kinetics (Figure 1B; rejection after about 3 weeks). Further analyses revealed a high propensity for IFN-γ–producing CD4+ and CD8+ T cells and reduced levels of IL-17–producing CD4+ T cells (Figure 1C), suggesting that IFN-γ–producing CD4+ and CD8+ T cells are the major effector cells in our model. Thus, our results indicate that a deficiency for T-bet in the NK cell compartment does not alter allograft rejection in our stringent STx model in the absence of pharmacological immunosuppression. Rag1 / T-bet / STx Recipients Show Accelerated Allograft Rejection Under Costimulatory Blockade Conditions

Rag1−/−T-bet−/− NK cells play a dual role in alloimmune responses under costimulatory blockade conditions because they have been shown to promote both allograft rejection and tolerance induction in different costimulatory blockade models.1,2,23,24 Based on this, we hypothesized that this dichotomy can be explained by the functional heterogeneity of distinct NK cell subsets. Therefore, we next investigated the regulation of T-cell–mediated alloresponses by CD27low versus CD27high NK cells in our adoptive transfer STx model under costimulatory blockade with MR1 and CTLA4Ig.

393

Importantly, we found that Rag1−/− recipients possessing T-bet–competent CD27low NK cells show significantly prolonged allograft survival (1 week) when compared to Rag1−/−T-bet−/− recipients that lack CD27low NK cells (Figure 2A), suggesting that the deficiency for CD27low NK cells in Rag1−/−T-bet−/− mice precipitates acceleration of allograft rejection. Moreover, we did not find any differences in the absolute numbers of NKp46+, CD27low (Figure S2A and B, SDC, http://links.lww.com/TP/B136) or T-bet–expressing CD27low NK cells (Figure S3, SDC, http://links.lww.com/TP/B136) between costimulatory blockade-treated and untreated recipients, suggesting that these differences in allograft survival are not influenced by costimulatory blockade-induced alterations in NK cell subsets. Next, we studied the mechanistic effects of CD27high versus CD27low NK cells on alloreactive CD4+ and CD8+ T cell responses in our transplant model in vivo. We found that proinflammatory IL-17A+CD4+ T cell responses that have recently been associated with costimulatory blockade-resistant allograft rejection25 were sufficiently suppressed in both mouse strains by double costimulatory blockade (Figure 2B). Critically, however, there was a significantly higher frequency of IFN-γ–producing CD8+, but not CD4+, T cells in Rag1−/−T-bet−/− versus the Rag1−/− mice (Figure 2B and C). Moreover, we found significantly decreased levels of IFN-γ in the serum of treated STx recipients when compared to untreated controls in both strains (Figure S4, SDC, http://links.lww.com/TP/B136). These results indicate that Rag1−/−T-bet−/− mice lacking CD27low NK cells have stronger alloreactive CD8+ T-cell responses in these mice.

Rag1 / T-bet / STx Recipients Show Enhanced Memory CD8+ T Cell Responses Due to Higher Homeostatic Proliferation

Based on the results above, we hypothesized that CD27low NK cells are important regulators of alloreactive CD8+ T cell differentiation and homeostasis. To address this, we first analyzed the distribution and absolute numbers of CD8+ and CD4+ T cells in Rag1−/−T-bet−/− and in Rag1−/− STx recipients in the presence or absence of costimulatory blockade. We found that CD4+ T cell numbers are similarly reduced on costimulatory blockade without strain-specific differences (Figure 3A and B). However, CD8+ T-cell numbers were notably higher in Rag1−/−T-bet−/− mice than in Rag1−/− mice under costimulatory blockade conditions (Figure 3A and B), suggesting that alloreactive CD8+ T cells in Rag1−/−T-bet−/− mice might undergo more homeostatic proliferation than in Rag1−/− mice. To verify this, we analyzed the proliferation response of CD8+ and CD4+ T cells in Rag1−/−T-bet−/− versus Rag1−/− mice via a CFSE proliferation assay. Indeed, we observed a strongly enhanced proliferation response in Rag1−/−T-bet−/− mice, which showed significantly greater proliferation of CD8+ T cells (Figure 3C, right) in Rag1−/−T-bet−/− than that in Rag1−/− mice. CD4+ T-cell proliferation (Figure 3C, left) did not show a significant difference between the 2 strains. Further analyses revealed that proliferating alloreactive CD8+ T cells in Rag1−/− and Rag1−/−T-bet−/− mice feature a memory phenotype (CD44+CD62L+ effector memory and CD44+CD62L− central memory cells), which is consistent with previous studies10 (Figure 3D). Thus, our results suggest

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

394

Transplantation



February 2015



Volume 99



Number 2

www.transplantjournal.com

FIGURE 1. Both Rag1−/− and Rag1−/−T-bet−/− recipients promptly reject skin allografts on adoptive T-cell reconstitution. A, Representative flow cytometry analysis (left) of NK cell subset distribution (CD27, CD11b and KLRG1 expression) and absolute numbers of CD27−CD11b+ or CD27−KLRG1+ cells (right) from spleens of Rag1−/− and Rag1−/−T-bet−/− mice (n =3), gated on viable NKp46+CD3− cells. B, Graft survival after adoptive transfer of purified CD3+ T cells (2.5  106 cells) 8 days after STx into Rag1−/−T-bet−/− and Rag1−/− recipients; Rag1−/−T-bet−/− and Rag1−/− controls received no T-cell transfer. C, Representative flow cytometry analysis of IL-17A and IFN-γ expression (left) in splenic CD4+ and CD8+ T lymphocytes (gated on viable CD3+CD4+ and CD8+CD3+ T cells) from STx recipients after adoptive transfer of CD3+ T cells is shown; the absolute number of CD4+IFN-γ+ and CD8+IFN-γ+ cells in spleens after ex vivo restimulation (50 ng/mL Phorbol 12-myristate 13-acetate and 550 ng/mL ionomycin for 5.5 hours) is also shown (right) (n =3-6).

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Lantow et al

© 2015 Wolters Kluwer

that T-bet–competent CD27low NK cells prolong allograft survival by controlling the homeostatic proliferation of alloreactive CD8+ T cells. Adoptively Transferred CD27low NK Cells Promote Allograft Survival in Rag1 / Mice by Controlling the Expansion of IFN-g-Producing CD8+ T cells

To confirm that CD27low NK cells can promote allograft survival by controlling alloreactive CD8+ T cells, we adoptively transferred flow-sorted wt CD27lowCD11bhigh NK cells into Rag1−/−T-bet−/− STx recipients and analyzed allograft survival under costimulatory blockade conditions. Our results show that reconstituted Rag1−/−T-bet−/− recipients (Figure 4A) show significantly prolonged allograft survival when compared to Rag1−/−T-bet−/− controls (Figure 4B), supporting our hypothesis that CD27low NK cells prolong allograft survival under costimulatory blockade conditions. Moreover, we found a strikingly reduced expansion of alloreactive CD8+ T cells in NK cell-reconstituted Rag1−/−T-bet−/− recipients in comparison to unreconstituted controls (Figure 4C and D). In addition, we found significantly reduced numbers of IFN-γ–producing CD8+ T cells in NK cell–reconstituted Rag1−/−T-bet−/− mice

395

(Figure 4E and F), confirming that T-bet–competent CD27low NK cells suppress the homeostatic expansion of alloreactive IFN-γ+CD8+ T cells. In summary, our results indicate that terminally differentiated CD27low NK cells, but not CD27high NK cells, are critical for prolonging allograft survival by limiting the expansion of alloreactive CD8+ T-cell responses. DISCUSSION The controversial role of NK cells in transplantation1–7 is likely due to the existence of specific NK cell subsets. Here, we analyzed the role of CD27low versus CD27high NK cells in an adoptive transfer model of T-cell–mediated skin allograft rejection in T-bet-deficient mice and found that CD27low, but not CD27high, NK cells are critically dependent on T-bet and prolong allograft survival by limiting the homeostatic proliferation and expansion of IFN-γ– producing memory CD8+ T cells under costimulatory blockade conditions. Because T-bet has been shown to be a key transcription factor regulating the maturation and differentiation of NK

FIGURE 2. Rag1−/−T-bet−/− STx recipients show accelerated allograft rejection under co-stimulatory blockade conditions. A, Graft survival in Rag1−/−T-bet−/− and Rag1−/− STx recipients after CTLA4Ig and MR1 costimulatory blockade and adoptive transfer of purified CD3+ T cells (2.5 x 106 cells) 8 days after STx; Rag1−/−T-bet−/− and Rag1−/− controls received no T-cell transfer. B, Representative flow cytometry analysis of IL-17A and IFN-γ expression in ex vivo restimulated (50 ng/mL PMA and 550 ng/mL ionomycin for 5.5 hours) splenic CD4+ and CD8+ T lymphocytes (gated on viable CD3+CD4+ and CD3+CD8+ cells) from STx recipients (n =6) is shown. C, Absolute numbers of splenic CD8+IFN-γ expressing T cells from Rag1−/−T-bet−/− and Rag1−/− STx recipients (n =6) after CTLA4Ig and MR1 costimulatory blockade (0.50 mg on day 8 and 0.25 mg on days 10, 12, 14, and 16 after STx).

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

396

Transplantation



February 2015



Volume 99



Number 2

www.transplantjournal.com

FIGURE 3. Rag1−/−T-bet−/− STx recipients show enhanced memory CD8+ T-cell responses due to higher homeostatic proliferation. A, Representative flow cytometry analysis of adoptively transferred CD3+ Tcells (2.5  106 cells) 8 days after STx and the distribution of CD4+ and CD8+ Tcells (gated on viable CD3+ cells) in Rag1−/− and Rag1−/−T-bet−/− mice with or without costimulatory blockade. B, Absolute number of splenic CD4+ T cells (left) and CD8+ T cells (right) with or without costimulatory blockade. C, Absolute numbers of CFSE+CD4+ (left) and CFSE+CD8+ T cells (right) from lymph nodes of Rag1−/−and Rag1−/−T-bet−/− mice on day 3 after T-cell transfer (n =3-4). D, CD44 expression in CD8+ T cells from lymph nodes of Rag1−/−and Rag1−/−T-bet−/− recipient mice with or without costimulatory blockade. CFSE, carboxyfluorescein diacetate succinimidyl ester.

cells,18,21 T-bet–deficient mice provide a valuable tool to study the role of functional NK cell subsets in alloresponses. However, the use of T-bet−/− mice for this purpose is limited due to the striking effects of the absence of T-bet on adaptive immune responses in those mice. For instance, T-bet–deficient mice fail to generate functional Th1 responses, show evidence of pronounced Th2 and Th17 responses, and experience accelerated allograft rejection under costimulatory blockade conditions.9,26,27 More specifically, T-bet−/− recipients of fully mismatched heart transplants show a CD8+ T17 cell– mediated alloimmune response that mediates acute, costimulatory blockade–resistant, allograft rejection, which can be overcome by targeting the costimulatory Tim-1/Tim-4 pathway.26,27

Importantly, CD8−/−T-bet−/−, but not CD4−/−T-bet−/−, recipients accepted their allografts long term,26 suggesting that memory CD8+ T cells are key alloreactive effector cells in this model. To specifically address the role of T-bet in regulating NK cells in alloimmune responses in the absence of adaptive immunity, we used Rag1−/−T-bet−/− mice that were reconstituted with T-bet–competent T cells as transplant recipients. Our findings in this model add a new perspective to the current literature by showing that CD27low NK cells control the expansion and proliferation of alloreactive memory CD8+ T cells in a T-bet–dependent manner. Moreover, our results lead us to speculate that the absence of CD27low NK cells in regular T-bet–deficient mice partially accounts for the proinflammatory

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Lantow et al

© 2015 Wolters Kluwer

397

FIGURE 4. Adoptively transferred CD27low NK cells promote allograft survival in Rag1−/−T-bet−/− mice by controlling the expansion of IFN-γ-producing CD8+ T cells. A, Representative flow cytometry analysis of CD27 and CD11b expression (gated on viable CD3−NKp46+ cells) and NK cell subset distribution from Rag1−/−T-bet−/− and CD27low NK cell-reconstituted Rag1−/−T-bet−/− STx recipients. B, Graft survival after CTLA4Ig and MR1 costimulatory blockade and transfer of flow cytometry–sorted 2.5  106 CD27low NK cells (1 day before STx) and 2.5  106 CD3+ T cells (8 days post STx) into Rag1−/−T-bet−/− recipients; Rag1−/−T-bet−/− and Rag1−/− controls received no Tcell transfer. C, Representative flow cytometry analysis of CD4+ and CD8+ Tcell distribution (gated on viable CD3+CD4+ and CD8+CD3+ cells) in Rag1−/−T-bet−/− and CD27low NK cell–reconstituted Rag1−/−T-bet−/− STx recipients. D, Absolute numbers of splenic CD8+ T cells in Rag1−/−T-bet−/− and CD27low NK cell–reconstituted Rag1−/−T-bet−/− STx recipients (n =5-6). E, Representative flow analysis of IL-17A and IFN-γ expression in ex vivo restimulated splenic CD8+ cells (gated on viable CD8+CD3+ cells) from Rag1−/−T-bet−/− and CD27low NK cell–reconstituted Rag1−/−T-bet−/− STx recipients. F, Absolute cell numbers of splenic CD8+IFN-γ–producing T cells from Rag1−/−T-bet−/ − and CD27low NK cell–reconstituted Rag1−/−T-bet−/− STx recipients (n = 5-6).

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

398



February 2015



Volume 99



Number 2

www.transplantjournal.com

CD8+ T17 cell–mediated alloresponse in those mice. Thus, our study closes a knowledge gap concerning the role of T-bet in regulating native immunity in alloresponses and provides detailed insights into how CD27low versus CD27high NK cell subsets control CD8+ T cell alloresponses and allograft rejection. Besides T-bet, Eomes was shown to play an essential role in determining NK cell differentiation18 and NK cell responses against major histocompatibility complex-mismatched targets.28 Although T-bet expression is more prominent in CD27low NK cells, Eomes expression is higher in CD27high NK cells.16,28 In this context, we recently showed that NK cells are highly responsive to IL-15, which promotes their expansion, differentiation, and effector maturation into highly cytotoxic KLRG1+ NK cells.12,28 Moreover, treatment with IL-15 could particularly restore the overall differentiation, maturation, and extralymphoid-homing capabilities of T-bet–deficient Eomes+ NK cells, indicating that Eomes can compensate for the lack of T-bet in the presence of abundant IL-15.28 Consistent with this literature, we found that IL-15–stimulated NK cells can mediate prompt rejection of skin allografts in the complete absence of adaptive immunity, suggesting that the role of NK cells in alloimmune responses is contingent on their cytokinedependent activation status.12 Therefore, we speculate that Eomes+CD27high NK cells, which are highly responsive to IL-15, represent a proinflammatory subset of NK cells that may trigger accelerated allograft rejection in Rag1−/−T-bet−/− recipients under costimulatory blockade conditions. From a mechanistic point of view, Eomes has been shown to regulate IFN-γ expression in CD8+ T cells and NK cells,26,27 which is in line with our results showing that Eomes+CD27high NK cells produce higher levels of IFN-γ than their T-bet–competent CD27low counterparts. Additional studies are clearly warranted using novel mouse models including conditional KO mice in which Eomes is selectively deficient in NK cells. Tools like these will help to precisely delineate the direct and indirect roles of Eomes-expressing versus T-bet–expressing NK cells in alloimmune responses. Zecher et al10 showed that long-term allograft survival in a homeostatic proliferation model is critically reliant on the presence of NK cells, which downregulate the homeostatic proliferation of CD8+ T cells in lymphopenic environments by competing for IL-15. However, the identity of NK cells involved and the underlying mechanisms regulating NK cell– mediated control of CD8+ T-cell responses still remained unclear. Homeostatic proliferation plays a significant role in our adoptive transfer model, which allowed us to mimic the effects of depletive induction therapy in transplant recipients, a clinically highly relevant scenario in transplantation because homeostatically generated memory T cells in the setting of organ transplantation precipitate allograft rejection and resist tolerance induction.10 Therefore, our results provide important clues to understand the mechanisms that regulate homeostatic T-cell proliferation of CD8+ T cells in an NK cell subset– dependent manner. Our findings identify CD27low NK cells as key effector cells promoting allograft survival by inhibiting the homeostatic proliferation of IFN-γ–producing memory CD8+ T cells under costimulatory blockade conditions in a T-bet– dependent manner. Intriguingly, we have noted higher IL-15 cytokine levels in the serum of Rag1−/−T-bet−/−versus Rag1−/− mice, an effect that is even more pronounced on skin transplantation (data not shown). One possible explanation for

these results could be that the large fraction of IL-15receptor-β-expressing CD27low NK cells in Rag1−/− mice may consume IL-15, thus leading to relatively lower IL-15 levels in Rag1−/− mice. Clearly, more experiments are required to elucidate respective underlying mechanisms. In summary, distinct NK cell subsets have different roles in alloimmune responses contingent on their regulation by the hallmark T-box transcription factors T-bet and Eomes. Eomes-expressing CD27high NK cells may be proinflammatory effector cells that accelerate allograft rejection, whereas T-bet–dependent CD27low NK cells prolong allograft survival by controlling the homeostatic proliferation of alloreactive memory CD8+ T cells. This may be of high relevance for the development of novel treatment strategies against costimulation blockade-resistant allograft rejection in the clinic.

Transplantation

ACKNOWLEDGMENTS The authors thank Lydia Schneider, Kathleen Burghardt, Azucena Martin-Santos, and Michael Rieger for excellent technical help and support. REFERENCES 1. Maier S, et al. Inhibition of natural killer cells results in acceptance of cardiac allografts in CD28−/− mice. Nat Med 2001;7(5):557. 2. McNerney ME, et al. Role of natural killer cell subsets in cardiac allograft rejection. Am J Transplant 2006;6(3):505. 3. Uehara S, et al. NK cells can trigger allograft vasculopathy: the role of hybrid resistance in solid organ allografts. J Immunol 2005;175(5):3424. 4. Yu G, et al. NK cells promote transplant tolerance by killing donor antigenpresenting cells. J Exp Med 2006;203:1851. 5. Kroemer A, et al. The innate NK cells, allograft rejection, and a key role for IL-15. J Immunol 2008;180(12):7818. 6. Beilke JN, et al. NK cells promote islet allograft tolerance via a perforindependent mechanism. Nat Med 2005;11(10):1059. 7. van der Touw W, et al. NK cells are required for costimulatory blockade induced tolerance to vascularized allografts. Transplantation 2012; 94(6):575. 8. Habiro K, et al. Effect of inflammation on costimulation blockade-resistant allograft rejection. Am J Transplant 2005;5(4 Pt 1):702. 9. Yuan X, et al. A novel role of CD4 Th17 cells in mediating cardiac allograft rejection and vasculopathy. J Exp Med 2008;205(13):3133. 10. Zecher D, et al. NK cells delay allograft rejection in lymphopenic hosts by downregulating the homeostatic proliferation of CD8+ T cells. J Immunol 2010;184(12):6649. 11. Laffont S, et al. Natural killer cells recruited into lymph nodes inhibit alloreactive T-cell activation through perforin-mediated killing of donor allogeneic dendritic cells. Blood 2008;112(3):661. 12. Kroemer A, Edtinger K, Li XC. The innate natural killer cells in transplant rejection and tolerance induction. Curr Opin Organ Transplant 2008; 13(4):339. 13. Hayakawa Y, et al. Functional subsets of mouse natural killer cells. Immunol Rev 2006;214:47. 14. Hayakawa Y, Smyth MJ. CD27 dissects mature NK cells into two subsets with distinct responsiveness and migratory capacity. J Immunol 2006; 176(3):1517. 15. Huntington ND, Vosshenrich CA, Di Santo JP. Developmental pathways that generate natural-killer-cell diversity in mice and humans. Nat Rev Immunol 2007;7(9):703. 16. Sabet-Baktach M, et al. Double deficiency for RORgammat and T-bet drives Th2-mediated allograft rejection in mice. J Immunol 2013; 191(8):4440. 17. Vivier E, et al. Functions of natural killer cells. Nat Immunol 2008;9(5):503. 18. Daussy C, et al. T-bet and Eomes instruct the development of two distinct natural killer cell lineages in the liver and in the bone marrow. J Exp Med 2014;211:563. 19. Gordon SM, et al. The transcription factors T-bet and Eomes control key checkpoints of natural killer cell maturation. Immunity 2012;36(1):55.

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Lantow et al

© 2015 Wolters Kluwer

20. Robbins SH, et al. Direct effects of T-bet and MHC class I expression, but not STAT1, on peripheral NK cell maturation. Eur J Immunol 2005; 35(3):757. 21. Townsend MJ, et al. T-bet regulates the terminal maturation and homeostasis of NK and Valpha14i NKT cells. Immunity 2004;20(4):477. 22. Rovira J, et al. A color-coded reporter model to study the effect of immunosuppressants on CD8+ T-cell memory in antitumor and alloimmune responses. Transplantation 2013;95(1):54. 23. Kim J, et al. The activating immunoreceptor NKG2D and its ligands are involved in allograft transplant rejection. J Immunol 2007; 179(10):6416.

399

24. Moreland L, Bate G, Kirkpatrick P. Abatacept. Nat Rev Drug Discov 2006;5(3):185. 25. Ford ML, Adams AB, Pearson TC. Targeting co-stimulatory pathways: transplantation and autoimmunity. Nat Rev Nephrol 2014;10(1):14. 26. Yuan X, et al. Targeting Tim-1 to overcome resistance to transplantation tolerance mediated by CD8 T17 cells. Proc Natl Acad Sci U S A 2009; 106(26):10734. 27. Burrell BE, et al. CD8+ Th17 mediate costimulation blockade-resistant allograft rejection in T-bet-deficient mice. J Immunol 2008;181(6):3906. 28. Malaise M, et al. KLRG1+ NK cells protect T-bet–deficient mice from pulmonary metastatic colorectal carcinoma. J Immunol 2014;192(4):1954.

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Three-year outcomes following 1420 ABO-incompatible living-donor kidney transplants performed after ABO antibody reduction: results from 101 centers.

Reports from experienced centers suggest that recipients of an ABO-incompatible living-donor kidney transplant after reduction of ABO antibodies exper...
8MB Sizes 0 Downloads 3 Views