Recipient natural killer cells alter the course of rejection of allogeneic heart grafts in rats

Rejection of solid organ grafts is regarded to be dependent on T cell responses. Nonetheless, numerous studies have focused on the contribution of NK cells in this process, resulting in contradictory theories. While some conclude that there is no participation of NK cells, others found an inflammatory or regulative role of NK cells. However, the experimental settings are rarely comparable with regard to challenged species, strain combinations or the nature of the graft. Thus, clear definition of NK cell contribution might be impeded by these circumstances. In this study we performed heterotopic heart transplantation (HTx) in rats, choosing one donor-recipient-combination leading to a fast and a second leading to a prolonged course of graft rejection. We intervened in the rejection process, by depletion of recipient NK cells on the one hand and by injection of activated NK cells syngeneic to the recipients on the other. The fast course of rejection could not be influenced by any of the NK cell manipulative treatments. However, the more prolonged course of rejection was highly susceptible to depletion of NK cells, resulting in significant acceleration of rejection, while injection of NK cells induced acceptance of the grafts. We suggest that, depending on the specific setting, NK cells can attenuate the first trigger of immune response, which allows establishing the regulatory activity leading to tolerance of the graft.


Introduction
The participation of NK cells in rejection of solid organ grafts was first described in the early 1980s, with evidence of influx of NK cells, then also described as large granular lymphocytes (LGL) into rat renal allografts [1]. Shortly thereafter the missing self-theory postulated by Kaerre et al. gave a reasonable explanation, why NK cells should respond to foreign grafts: Cells that lack expression of self MHC-I molecules are targets for NK cell mediated killing [2]. However, a study by Shelton et al. reported of non-rejection of allogeneic skin-grafts in SCID mice, whereas rejection was re-established by transfer of-predominantly CD4+-T cells a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 The donor was intravenously injected with 500 I.U. heparin, the thorax was subsequently opened and the ascending aorta and the pulmonary trunk were severed. The inferior and superior vena cava, as well as the pulmonary veins, were ligated with a single 5-0 ligature. The explanted heart was perfused with 30 ml isotonic saline solution supplemented with 1000 I.U. heparin. For implantation the abdominal aorta and the inferior vena cava were temporarily clamped using a Cooley vascular clamp and the vessels were incised in longitudinal direction. The donor ascending aorta was connected to the abdominal aorta and the donor pulmonary trunk to the inferior vena cava via 8-0 monofilament suture. For perioperative analgesia, rats were treated with a subcutaneous depot of 5 mg Carprofen per kilogram bodyweight. Graft survival was controlled by daily palpation through the abdominal wall.

Treatment of rats prior and after transplantation
In vivo depletion of NK cells using mAb HT30. Monoclonal antibody (mAb) HT30 detects the allomorphic protein NKR-P1A LEW . It has been developed in our lab, using the rat strain pair LEW.TO-NKC2 (donor, carries NKR-P1A LEW ) and LEW.TO-NKC (recipient, carries NKR-P1A TO ). Both NKC-congenic strains share the same genetic background and allelic sequences of the natural killer gene complex (NKC) downstream of NKR-P1A. LEW.TO-NKC animals were immunized with donor spleen cells. Supernatants of established hybridoma cells were screened using 293 cells transfected with NKR-P1A LEW .
NK cells were depleted in vivo by a singular subcutaneous application of 500 μg of mAb HT30 one day prior to HTx.
Injection of NK cells. NK cells were positively isolated using biotinylated mAb 3.2.3 and Streptavidin-labelled microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany). The isolated NK cells were incubated for 7-10 days in culture medium supplemented with rat IL-2. 1-2x10 6 NK cells (with a purity � 90%) were then injected intravenously as a single shot directly after transplantation.
Treatment with Ciclosporin (CsA). Selected recipients were injected daily with a subtherapeutic dose of 1.25 mg/kg body-weight subcutaneously. This treatment led to 60% graft survival after the first 21 days of observation.

Subcutaneous placement of heart cells in the ear
Perfused explanted hearts from donors were chopped up into 3x3 mm blocks and digested with 0,5 mg/ml collagenase I for 30 min at 37˚C. The tissue was mashed through a large-pore sieve, resulting in vital muscle cell congeries, (mostly dead) single heart cells and remaining blood cells. By passage through a 40 μm cell strainer, the congeries were separated and 1x10 4 were injected subcutaneously in the ear of specified recipients.

Histology and scores for infiltration
Cryostat sections of Tissue-Tek (Sakura, Alphen aan den Rijn, Netherlands) embedded grafts were air dried, acetone/methanol fixed, and incubated with mAb to TCRα/β (clone R73), CD4 (W3/25), CD68 (ED1, AbD Serotec, Düsseldorf, Germany), CD161 (3.2.3) and NKR-P1A (HT30). Antibodies were purified in our lab, except where noted. Stained cells were detected with bridge antibodies (rabbit anti-mouse Ig) and alkaline phosphatase anti-alkaline phosphatase (APaAP) (both Dako, Hamburg, Germany). Nuclear staining of sections was performed with hematoxylin (Merck, Darmstadt, Germany). Infiltration of lymphocytes was assessed in double-blind evaluation by light. Of note, our quantification followed a fixed classification starting with 0.5 = singular distributed positive stained cells marginally occurring in the tissue section; 1.0 = singular distributed positive stained cells occurring in every field of view; 1.5 = numerous positive stained cells uniformly distributed over the whole tissue section; 2.0 = strong distribution and 2.5 = very strong distribution of positive stained cells.

Mixed lymphocyte reaction
2x10 5 responder cells were either stimulated and re-stimulated applying a specific stimulus with equal numbers of lethally irradiated allogeneic splenocytes or using plate-bound CD3 and soluble CD28. After 5 days incubation in 96 well-round bottom plates, lymphocytes were pulsed with 0.5-1 μCi [ 3 H]thymidine/well for 16 hours and [ 3 H]thymidine incorporation was assessed after scintillation using a β-counter (LKB Wallac, Turku, Finland). In certain mixed cultures, NK cells were depleted from bulk splenocytes using mAb 3.2.3 and MACS beads. IFN-γ production was measured in the supernatants of the culture by enzyme-linked immunoabsorbant assay (ELISA).

Assessment of mRNA expression
Draining cervical lymph node (LN) cells from Lew wt and Lew.1a rats grafted with allogeneic heart cells in the ear were lysed and mRNA was isolated using NucleoSpin RNA II kit (Macherey-Nagel, Düren, Germany) according to manufacturer's instructions. mRNA was subsequently translated into complementary DNA by supplementing 1 μg RNA with Oligo (dT) primer and RevertAid Transcriptase (both Fermentas, St. Leon-Rot, Germany) and incubation for 1 h at 42˚C. Reaction was stopped at 72˚C for 10 min. cDNA was amplified by PCR using primer for TGF-β, IL-10, FoxP3, IL-1β and IFN-γ ( Table 1). Expression of GAPDH served as housekeeping gene. PCR included 30 cycles with 30 sec 94˚C, 30 sec 58˚C and 1 min 30 sec 72˚C.

Statistical analysis
Results are presented in general as the mean ± standard deviation, depicting at least five individuals per group. In vitro experiments were performed independently usually five to six times (minimum of three times). Statistical analysis was generally performed with the unpaired twotailed Student's t-test using prism software (Graph Pad Software, San Diego, CA, USA).

Results
The comparison of two different, fully MHC disparate rat strain combinations discloses divergent effects on NK cells To assess the possible regulative function of NK cells in allograft rejection in vivo, we performed heterotopic HTx for two reasons: 1. Graft survival can be assessed directly and noninvasively until final cessation of heartbeat, since graft rejection does not impede vital functions of the recipient. 2. We were able to refer to a comprehensive study by Klempnauer et al. who followed heart-graft survival in numerous MHC-disparate rat strains that were combined in a vice-versa manner [25].
We included a fast rejecting donor-recipient combination and a combination of strains resulting in a prolonged rejection, to analyze the regulatory capacity of NK cells, depending on the severity of the immune response. Hence, we chose the strain combination Lew.1a ! Lew wt with a reported mean graft rejection of 8.4 days and the strain combination Lew.1u7B ! Lew.1a, with a mean rejection time of 17.5 days [25].
Allogeneic grafts from Lew.1a donors were rejected by Lew wt recipients in a mean time of 7.3 days, in line with results obtained by Klempnauer et al. [25], whereas syngeneic grafts showed vital function over the complete observation period (Fig 1A).
In Lew.1a recipients of allogeneic grafts, derived from Lew.1u7B donors, we observed a mean graft survival of 36.4 days, which was significantly longer than reported in the aforementioned study (Fig 1A). Of note, three of the seven Lew.1a recipients had vital graft function over the observation period of 60 days (Fig 1A and 1C).
Analysis of NK cell populations derived from lymphatic tissue (lymph node and spleen) and the blood of both recipient strains revealed that fast rejecting Lew wt recipients reacted with an increase of NK cell ratios upon allogeneic engraftment, when compared to naïve rats, accentuated in the spleen and blood. Lew.1a recipients on the other hand, displayed a general decrease of NK cell ratios compared to naïve rats ( Fig 1B). A detailed comparison of rejecting and non-rejecting Lew.1a recipients demonstrated significantly lower NK cell ratios in the analyzed compartments in rejecting recipients ( Fig 1D, S1A Fig).
Assessment of recipient T cell ratios in the two strain combinations also revealed distinct alterations: Whereas Lew wt rats engrafted with allogeneic hearts responded with a significant decrease of T cells in spleen and blood compared to the naïve rats or recipients of syngeneic grafts, Lew.1a recipients showed only moderate changes in T cell ratios upon allogeneic engraftment (S1B Fig). The specific alterations of NK cell ratios between the recipients of the two strains (Lew wt and Lew.1a) and the distinct results derived from rejecting and non-rejecting recipients in the latter strain, prompted us to investigate general differences in NK cell biology between the specific rat strains.

NK cells show varying potential to influence the response of T cells depending on the specific rat strain
Bulk splenocytes and those depleted of NK cells from Lew wt and Lew.1a rats were stimulated in vitro with irradiated splenocytes derived from the involved allogeneic rat strains of this study (cells derived from Lew.1u7B rats were only used for stimulation). The stimulation of untreated bulk splenocytes disclosed a highly distinct reaction to the respective trigger. Lew wt An illustration displaying a rat with a highlighted heart indicates that the presented data resulted from the model of HTx in distinction from data obtained from the second transplantation model, in which allogeneic heart muscle cells were injected into the ear of the respective recipients (see also Fig 3). (A) Perfused hearts derived from Lew.1a (RT1 a ) rats were heterotopically transplanted into the abdomen of Lew wt (RT1 l ) recipients (fast rejection) and Lew.1u7B (RT1 u ) hearts were engrafted in Lew.1a (RT1 a ) recipients (prolonged rejection). Syngeneic HTx served as a control group (n = 5). (B) Flow cytometric assessment of frequencies of NK cells in Lew wt and Lew.1a recipients after syngeneic and allogeneic HTx compared to naïve rats, respectively. The data implies divergent reactions of the two strains to the grafts: Upon allogeneic engraftment Lew wt recipients displayed increased NK cell ratios, which was observed especially in spleen and blood of the recipients, whereas Lew.1a recipients reacted with a general decrease of NK cell ratios compared to naïve rats or recipients of syngeneic grafts. (C) Three of 7 Lew.1a recipients receiving allogeneic grafts developed prolonged graft survival with no differences in viability, when compared to syngeneic controls over the complete observation period of 60 days resulting in the prolonged mean survival of allogeneic grafts in Lew.1a recipients. (D) NK cell frequencies of rejecting and non-rejecting Lew.1a recipients in cervical (cerv) and mesenterial (mes) LN, spleen and blood. cells responded to Lew.1a splenocytes with higher proliferation (mean of 21 000 cpm) than to the Lew.1u7B strain (mean 8700 cpm). Lew.1a responder cells revealed a strong proliferation of about 47 000 cpm provoked by irradiated Lew wt cells, whereas the stimulation with Lew.1u7B derived stimulators resulted in a proliferation of 19 000 cpm (Fig 2A, upper panel). These strain dependent differences became more pronounced after depletion of NK cells from the responding splenocytes: Responder cells from Lew wt rats showed enhanced proliferation (> 2 times higher after stimulation with Lew.1a cells compared to bulk splenocytes; > 7 times higher in case of stimulation with Lew.1u7B cells). In responder cells of Lew.1a, the absence of NK cells resulted in almost no divergent proliferation after exposure to Lew wt stimulators (bulk: 46 800 cpm vs. NK depleted: 49 900 cpm), but in a > 2 times increase after stimulation with Lew.1u7B splenocytes, reflecting the strain combination used in vivo.
This potential influence on T cell proliferation by NK cells could also be observed by evaluation of IFN-γ release after NK cell depletion (Fig 2A, lower panel), which was increased in both responding strains.
Contrary to the allogeneic induction of proliferation, the assessment of spontaneous activation of T cells by incubation of splenocytes solely in medium induced a pronounced proliferation of CD4 + cells derived from Lew wt and especially from Lew.1a rats. This tendency was accentuated in the absence of NK cells (Fig 2B).
Rejecting rats showed a significant decrease of NK cells, when compared to non-rejecting (and untreated) rats. Survival of grafts is shown as a Kaplan-Meier presentation (A) and in a diagram, respectively (C). The ratios of NK cells are presented as mean ± SEM of three (in D) and at least five (in B) recipients. Significance is indicated with � for p-values � 0.05, �� for p-values � 0.005 and ��� for p-values � 0.0005. If no asterisk is shown in the respective graph the results were not significantly different. https://doi.org/10.1371/journal.pone.0220546.g001

Fig 2. Distinct capacity of NK cells to control T cell activation, depending on the rat strain. (A) Proliferation (upper panel)
and IFN-γ release (lower panel) of 2x10 5 naïve splenocytes derived from Lew wt and Lew.1a rats after in vitro stimulation with 2x10 5 irradiated splenocytes from the respective MHC disparate strains for 6 days. Responder splenocytes were either bulk or depleted of NK cells. The proliferation of T cells was highly controlled by NK cells in Lew wt rats, while Lew.1a rats showed a less consistent outcome in the absence of NK cells. The amount of secreted cytokine was highly related to the degree of proliferation, which indicates T cells as source of the detected IFN-γ. (B) Spontaneous proliferation of CMFDA-labelled Lew wt and Lew.1a splenocytes in the presence or absence of NK cells, respectively, incubated in medium alone for 6 days, was elevated in the absence of NK cells in both rat strains. Results are presented as exemplary dot blots or as mean ± SEM of at 4 least independent in vitro experiments. Significance is indicated with � for p-values � 0.05, �� for p-values � 0.005 and ��� for p-values � 0.0005. If no asterisk is shown in the respective graph the results were not significantly different. Analyzing the process of initiation of immune responses in Lew wt and Lew.1a recipients towards allogeneic grafts in vivo is challenging, especially since the explicit identification of graft draining LN is difficult in heterotopic HTx. As a consequence, the crucial local environment to follow the commencement of T cell activation after engraftment cannot be assessed.
Before manipulating NK cells in recipients of allogeneic heart grafts, we therefore performed additional in vivo experiments by challenging the two recipient strains by placing clusters of allogeneic heart muscle cells subcutaneously in the ear of the respective rat strain. By applying this method it is more feasible to track immunological processes in cervical lymph nodes (LN) draining the pinna.

Subcutaneous placement of heart muscle cells induces an attenuated immune response in Lew.1a rats
After subcutaneous placement of heart muscle cells in the pinna, cells harvested from (draining) cervical and mesenteric LN (serving as intrinsic control) were re-stimulated specifically with splenocytes from the respective donor strain. In Lew wt rats, re-stimulation led to an increase in proliferation, however, there was almost no detectable distinction in proliferation of lymphocytes derived from draining and mesenteric LN. In contrast, Lew.1a draining LN cells revealed strongly suppressed proliferation in comparison to mesenteric LN control cells and to draining LN from Lew wt (Fig 3A, S2 Fig). This suppression is mainly based on a significant reduction of activated T cells and not on T cell division in general (S2 Fig). Interestingly, unspecific activation of lymphocytes using CD3/28 antibodies induced pronounced proliferation in all tested LN of both strains, revealing a general proliferative capacity of T cells, even in Lew.1a draining (cervical) LN (Fig 3B, S2 Fig).
Parallel to these observations, relative NK cell numbers from draining LN of Lew.1a rats were significantly increased, when compared to intrinsic (mesenteric) control LN and draining LN from Lew wt (Fig 3C). Of note, the draining LN of Lew wt rats also revealed a significant increase of the NK cell ratio, when compared to the mesenteric LN, although this was much less accentuated than in Lew.1a recipients (p-value LN drain vs LN mes in Lew.1a: p = � 0.0001 and Lew wt: p = 0.0145).
We took aliquots from the draining LN cells used for analysis of lymphocyte proliferation and NK cell ratios, prepared the cell mRNA and subsequently the corresponding cDNA. Cells derived from Lew wt rats had higher mRNA expression for IL-10 and FoxP3, but revealed a complete lack of IL-1 and IFN-γ expression ( Fig 3D). In contrast, LN cells derived from Lew.1a rats displayed lower FoxP3 mRNA levels and no IL-10 but low levels of IL-1 and IFNγ. Only the detected amounts of TGF-β were almost identical in both rat strains. Thus, regardless of the significantly higher proliferative activity of Lew wt T cells harvested from draining LN, a classical suppressive cytokine milieu was observed in the draining LN. Cells derived from Lew.1a rats displayed a more inflammatory profile according to the IL-1 and IFN-γ expression, although specific re-stimulation revealed reduced proliferative capacity.
The increased expression of suppressive cytokine mRNA could also indicate an enhanced immune response in Lew wt rats. For that reason, we also analyzed the infiltration of immune cells at the site of heart muscle cell placement in the pinna by immunohistochemistry, as an indication of the existence of activated immune cells. Indeed, Lew wt rats revealed a well detectable infiltration of immune cells, predominantly T cells and monocytes. In Lew.1a rats the influx of immune cells into the ear tissue was only marginal (Fig 3E). Thus, Lew.1a recipients challenged with Lew.1u7B cells displayed an attenuated immune response against the allogeneic stimulus, accompanied by an increase of NK cell ratios when compared to the response of Lew wt recipients to Lew.1a heart cells. The obtained cytokine profiles showed a simultaneous increase of IL-1 and IFN-γ, possibly accounting for an increase in activation of NK cells. However, the secretion of regulatory cytokines such as IL-10 and TGF-β in this specific setting could not be observed. Considering the initial in vivo as well as the above mentioned in vitro results, NK cells seem to exert specific immunoregulatory function in Lew.1a rats upon allogeneic challenge to Lew.1u7B cell or organ grafts.

Successful interference on the course of allogeneic graft rejection by manipulation of NK cells in Lew.1a but not in Lew wt recipients
To further characterize NK cell functions in vivo we performed HTx in the above mentioned strain combinations and manipulated NK cells in the recipients either by depletion of recipient NK cells before HTx using our lab-made mAb HT30 (which is specific for NKR-P1A and does not cross-react with NKR-P1B) or by injection of activated syngeneic NK cells subsequently after successful surgery. The depletion of NK cells was highly efficient in both strains (S3 Fig). The injection of NK cells led to a decrease of NK cell ratios in Lew wt rats, whereas Lew.1a recipients did not reveal detectable alterations of NK cells ratios in the analyzed compartments (S3 Fig). Fig 4A shows the outcome of this respective treatment in Lew wt rats engrafted with Lew.1a hearts. Neither depletion nor injection of syngeneic NK cells could influence the course of rejection. Of note, daily administration of subtherapeutic doses of CsA (1.25 mg/kg body weight) led to a decline of CD4 + and CD8 + T cells and significantly prolonged the rejection of the grafts, indicating that the rejection is based on T cell activity (S4 Fig). The graft infiltrating cells were mainly T cells, macrophages and monocytes, whereas the latter dominated the infiltration of the allogeneic grafts (Fig 4B, S5A Fig).
The evaluation of the histological sections was based on the respective infiltration of syngeneic heart grafts (S5A and S5B Fig). The allografts in Lew wt recipients were invaded significantly above this baseline. The pattern of infiltration in untreated recipients was comprehensive and not restricted to certain local areas of the grafts. The same pattern could be observed in NK cell depleted and injected rats (data not shown). However, the treatment with CsA led to clusters of immune cells in the proximity of the ventricle and the edge of the muscle tissue ( S5A Fig). Furthermore, we detected a significant increase of NK cell influx and a decrease of macrophages in the composition of infiltrating cells, whereas T cells revealed a slight increase compared to untreated rats (S5B and S5C Fig). Treatment of Lew wt recipients with NK cell depleting antibodies had only marginal effects on the graft infiltration, except for a significant decrease of NK cells among the infiltrating cells (Fig 4B). Injection of NK cells on the other hand had a surprising effect: While NKR-P1A/B + (monocytes/NK and NKT cells) cells increased in the heart tissue, influx of NKR-P1A (NK and NKT cells) single positive cells was significantly reduced, when compared to syngeneic controls, which reflects a broad increase of monocytes into the graft tissue in the presence of additional syngeneic NK cells (Fig 4B).
In general, broad graft infiltration was already observed three days post HTx, especially with a significant influx of CD68 + macrophages and NKR-P1A/B + monocytes when compared to the time of graft rejection (S5D Fig). In contrast to the observations made in Lew wt recipients, the manipulation of NK cells revealed a key function of these cells in the regulation of rejection of Lew.1u7B heart grafts by Lew.1a recipients (Fig 4C). Depletion of recipient NK cells significantly reduced allograft survival to a mean survival of 15 days, whereas injection of syngeneic NK cells completely abrogated allograft rejection. Of note, the recipients received no other treatment for immunosuppression or foregoing immunization; solely the absence or enforced presence of NK cells induced the accelerated rejection and long-term graft survival, respectively. Histological analyses of the grafts revealed increased infiltration of syngeneic grafts when compared to syngeneic grafts of Lew wt recipients. Although an increase of cell influx in allogeneic grafts of untreated Lew.1a recipients was evident at the time of rejection, we observed a significant decrease of immune cells upon NK cell depletion, especially of the NKR-P1A/B + cell population, despite an accelerated rejection (S5E Fig). However, since additional infiltration of allogeneic grafts was only slightly above the baseline of syngeneic grafts, analysis by immunohistochemistry was insufficient in providing detailed changes depending on the performed intervention. We therefore decided to prepare heart tissue for flow-cytometry for further analysis of distinct lymphocyte subsets, despite the risk of selectively losing specific cell populations during tissue preparation.

The injection of NK cells in Lew.1a recipients reduced the number of infiltrating monocytes and activated T cells drastically
Flow cytometric analysis of heart infiltrating cells in Lew.1a rats exposed that the majority of infiltrating T cells were CD25 + . The depletion of recipient NK cells did not significantly alter the population of CD4/CD25 + and CD8/25 + T cells despite an accelerated rejection. A substantial reduction of these cells however was observed after injection of NK cells, which diminished all infiltrating cell populations in comparison to untreated recipients (Fig 5A). Of note, the decrease of CD25 + T cells could not be solely attributed to either activated T cells or T reg since flow cytometric analysis of FoxP3 expression revealed a distribution of T reg among all CD25 expressing T cells in both recipient strains (S6 Fig). Therefore, the decrease of CD25 + T cells could account for both populations.
A reduced graft infiltration was also observed for the most abundantly detected cell population in the heart tissue: CD172 + /CD11b/c + monocytes and macrophages. Of note, a comparison of dot blots of blood circulating and heart infiltrating immune cells revealed that all populations detected in the blood were also present in the graft (Fig 5B). The CD172 + /CD11b/ c + cells could be divided into two populations with respect to their predominant localization in blood and grafts. While the majority of CD172 + /CD11b/c + cells in the blood displayed a phenotype with a high mean fluorescence intensity of CD11 (termed here CD172 + /CD11b/ c high ), there was a considerable CD172 + /CD11b/c + population among the graft infiltrating cells with a lower mean fluorescent intensity (MFI) of CD11, when compared to the fraction of CD172 + /CD11b/c high cells (termed here CD172 + /CD11b/c low ). Interestingly, there was a correlation of accelerated rejection and detected cell populations: NK cell depletion caused increasing percentages of CD172 + /CD11b/c high cells with simultaneous decrease of CD172 + /CD11b/ c low cells. As assumed for the rejection process in Lew wt recipients, cells of the monocyte/ macrophage fraction seemed to be critically associated with the rejection of the allogeneic graft in Lew.1a recipients. This is also reflected by the observations made after injection of syngeneic NK cells in Lew.1a recipients: all CD172 and CD11b/c positive cells (CD11b/c low and high ) were considerably reduced (Fig 5C).

Discussion
Although more commonly known for their pivotal role in cytotoxic and inflammatory processes, NK cells can also exhibit a strong regulative capacity towards T cells. This characteristic feature is not restricted to certain infections or immunological triggers. For instance, in mice it was reported for models of viral infections [27,28] and after allogeneic skin transplantation [29,30]. Previously, our workgroup has demonstrated the regulatory effects of rat NK cells on the onset of naïve T cells [16]. To evaluate the immunological significance of our in vitro results, we performed transplantation of allogeneic heart grafts in two strain combinations, characterized by a fast and a more prolonged course of rejection, respectively [25]. In grafts of untreated Lew.1a recipients we observed increased fractions of CD4 + , CD8 + T cells and especially of activated T cells (CD25 + ). These percentages did not correlate with the analysis of TCR + cells, however, activated T cells often internalise their TCR, which might be an explanation for the discrepancy [26]. NK cells invaded the graft tissue only moderately. Depletion of NK cells led to a significant reduction of CD4 + / CD25and CD8 + /CD25cells, whereas the influx of activated T cells was unaffected. The injection of syngeneic NK cells on the other hand, almost completely abrogated the infiltration of activated T cells into the graft. (B) Dot blots of lymphocytes and granulocytes in the blood and allografts defined by TCR vs. NKR-P1A/B and CD172 vs. CD11b/c expression of untreated recipients. Monocytes were among the major graft infiltrating cells and could be further discriminated according to the MFI of the CD11b/c expression. (C) Summary of the flow cytometric analysis of CD172/CD11b/c expressing cells in the respective group. Graft rejection, in Despite providing further evidence for the distinct characteristic of NK cells to regulate T cell responses in vitro, we were surprised to learn that the manipulation of recipient NK cells either by depletion or injection, did not influence the course of fast rejection in vivo. This is most probably due to the severe activation of the adaptive and innate immune system upon allogeneic engraftment.
It is worth noting that, the fast and severe infiltration of Lew.1a grafts in Lew wt recipients by macrophages and monocytes (as precursors of tissue macrophages) raised doubts, that the rejection in this specific strain combination is predominately T cell mediated. However, the depletion of monocytes/macrophages by administration of Clodronate did not influence rejection (data not shown), whereas subtherapeutic administration of CsA significantly prolonged the course of rejection.
In conclusion, NK cells did not exhibit a central role in this rigid experimental setting of fast rejection of allogeneic heart grafts. Accordingly, previous studies have implicated that allogeneic grafts, provoking a strong immune response, are rejected independently of the presence of NK cells [4]. In contrast, tolerance mediated by NK cells was shown to be induced in a moderate course of rejection, if T cell activity is hindered by co-stimulatory blockage [7]. In line with these reports, the presence or absence of NK cells in the second strain combination, characterized by a more prolonged rejection, significantly influenced the survival of the graft: Depletion of recipient NK cells led to a significantly accelerated loss of graft function, whereas injection of syngeneic NK cells completely prevented graft rejection, without simultaneous immunosuppression.
Of note, the applied strain combination showed occasional indefinite graft survival for untreated recipients, indicating an exceptional immunological setting. Specific strain combinations with similar observations of prolonged graft survival, despite fully disparate MHC, were identified in the past [31]. However the underlying mechanisms explaining these results were not further clarified.
In past experiments immunization of Lew.1a recipients with Lew.1u7B derived blood [32] or skin [33] led to complete tolerance of Lew.1u7B heart grafts without further treatment, which was demonstrated to be infectious for naïve recipients. This is especially noteworthy, since others postulated that donor-specific transfusion (DST)-mediated tolerance can only be induced by depletion of recipient CD4 + T cells [34] or mild gamma-irradiation of recipients [35]. As an explanation for these observations, the authors concluded that by attenuating the initial trigger for an allogeneic immune response, the induction of a repertoire of CD4 + regulatory T-cells is facilitated, which results in a highly efficient mechanism to maintain tolerance, even after several "infectious transfers". Interestingly, TGF-β seems to be crucially involved in this procedure [33], which would be in line with our previous data, revealing that suppression of T cells by NK cells is mediated directly by secretion of TGF-β [16]. Therefore, we assessed ex vivo mRNA expression of selected cytokines in draining LN cells from Lew.1a and Lew wt rats after exposure to donor (Lew.1u7B and Lew.1a respectively) derived allogeneic heart cells. We found similar TGF-β expression in both strain combinations, but a more inflammatory cytokine profile in Lew.1a recipients with enhanced expression of IFN-γ, whereas Lew wt recipients tended to an expression of a more suppressive profile, with expression of IL-10 and FoxP3. A possible explanation could be, that a strong immune response provokes a higher particular after NK cell depletion, is accompanied by increased percentages of CD172 + /CD11b/c high cells. Injection of syngeneic NK cells impeded the invasion of macrophages/monocytes into the heart tissue. Results are presented as exemplary dot blots or as mean ± SEM of at least 5 individuals per group. Significance is indicated with � for p-values � 0.05, �� for p-values � 0.005 and ��� for p-values � 0.0005. If no asterisk is shown in the respective graph the results were not significantly different.
https://doi.org/10.1371/journal.pone.0220546.g005 induction of suppressive cytokines in general, since mitigating a highly inflammatory immune response is of high priority for the host, as has been demonstrated by Lee et al. in mouse model of CMV [36].
While it is a common notion, that regulatory T cells play a central role in inducing and maintaining graft acceptance, our flow cytometric analyses to define the role of regulatory T cell populations (CD4 + CD25 + FoxP3 as defined by Nessi et al. in rats [37]) in selected experiments, did not reveal significant alterations of this cell population, leading to the assumption that these are not major contributors to the process of graft acceptance in our setting (preliminary data not shown). Accordingly, Lair et al. reported on tolerance transfer in the same transplantation model and rat strain combination as used in our experiments (Lew.1w ≙ Lew.1u ! Lew.1a), providing proof for blood non-T cells as effective mediators for induction of indefinite graft survival upon adoptive transfer, as opposed to a transfer of blood T cells, which failed to induce tolerance in this special setting [38].
As stated above, monocytes and macrophages were the predominant graft infiltrating cell population and were detected as early as 3 days post HTx, which is in line with previous reports [39,40,41]. The simultaneous expression of CD4 and CD68 points towards a CD4 + population of macrophages that invades the grafts, as was reported by Wallgren et al. in the past [41]. The infiltrating monocytes and macrophages were shown to express inflammatory molecules, like inducible NO synthase (iNOS) and TNF-α, which can damage the function of the graft, eventually resulting in rejection [40]. The participation of graft infiltrating macrophages/monocytes could also be tracked in the Lew.1a recipients engrafted with hearts of Lew.1u7B, whereby the presence of CD172 + /CD11b/c + cells was abundant in untreated recipients rejecting the grafts. For dendritic cells (DC) it has been shown that the CD172 + /CD11b high expressing cells enhance inflammation in LN and directly at the site of inflammation [42]. Upon depletion of NK cells and concomitant accelerated course of rejection, we observed a significant increase of CD172 + /CD11b/c high expressing cells in the graft, whereas injection of NK cells abrogated the occurrence of these cells. We assume that these observations are caused by an attenuated activation of T cells mediated by NK cells, leading to a reduction of signals attracting inflammatory cells into the graft tissue. This is supported by the almost complete disappearance of CD25 + (activated) T cells in the heart grafts after injection of syngeneic NK cells.
Furthermore, we believe that the attenuation of T cell activation occurs in the draining LN, as has also been demonstrated by Laffont et al. in a model of mouse skin transplantation [30]. This is indicated by the significant increase of NK cells in draining LN of Lew.1a, but not of Lew wt rats upon subcutaneous placement of allogeneic heart cells into the pinna. Lew.1a rats reacted with a significant increase of NK cells in draining LN in comparison to naïve rats and suppressed the T cell response upon allogen-specific re-stimulation significantly, which could not be observed in Lew wt rats. We assume that this attenuated response in combination with the pronounced characteristic of NK cell homing to draining LN in Lew.1a is the key of tolerance induction in these rats. This is additionally suggested by the significant decrease of circulating NK cells in untreated Lew.1a recipients, rejecting the allogeneic heart grafts.
The mechanism of this suppressed T cell activity can only be assumed, since cytokine profiles did not give clear indications of an immunosuppressive environment. Aside from suppressive cytokine production, direct lysis of activated T cells (possibly via perforin-dependent mechanisms) could be hypothesized, as was reported by several authors [29,30], however previous analyses for NK cell cytotoxicity in the two implemented strains in our laboratory did not reveal significant differences between the two strains and flow cytometric analysis for CD107a expression in experiments did not indicate NK-cell mediated lysis of target cells (data not shown).
In this study, we provide evidence that NK cells are regulative agents, able to control T cell activity during rejection of completely MHC-disparate grafts. This observation however, was dependent on the chosen strain combination. The diversity of animal models used to evaluate the role of NK cells in graft rejection might provide a reasonable explanation, why such diverse and to some extent contrary outcomes are postulated, ranging from no participation [3,4,5], to participation which is outweighed by T cells [8,43] and finally with active participation [7,10,13]. The direct comparison of the two presented strain combinations in this study gives an indication that the degree of allogeneic immune activation is crucial, when evaluating participation of specific cell populations in the course of rejection and leads us to following conclusion: NK cells participate and are able to influence the course of solid organ rejection, as long as the stimulation of T cells is of moderate nature. lymph node cells of Lew wt and Lew.1a rats after subcutaneous placement of allogeneic heart muscle cells (derived from Lew.1a and Lew.1u7B, respectively) and 6 days after in vitro-stimulation with splenocytes from the respective donor strain (specific re-stimulation) or with splenocytes from a third party strain (for Lew wt responder ! Lew.1u7B cells and for Lew1.a responder ! Lew wt cells; unspecific re-stimulation). Whereas Lew wt T cells responded with an enhanced proliferation to the specific re-stimulation in comparison to an unspecific stimulus, T cells derived from Lew.1a rats showed a significant reduced proliferation upon specific re-stimulation. This reduction of proliferation was mainly based on a reduced number of activated T cells in general and not by means of cell divison as indicated by the dot blots. Of note, the ability to regulate the T cell responses by NK cells in Lew.1a rats is promoted by a weak induction of proliferation through Lew.1u7B cells in Lew rat strains in general, and especially in Lew.1a rats. (TIF)

S3 Fig. Relative frequencies of NK cells in Lew wt and Lew.1a recipients upon intervention.
Flow cytometric analysis of the relative frequencies of NK cells among lymphocytes derived from cervical and mesenteric LN (white and black bars) and from spleen and blood (bright and dark grey bars) of the respective group in Lew wt and Lew.1a recipients, after rejection and at the end of the observation period, respectively. While NK cell depletion was successful in both strains and in all analyzed compartments, untreated Lew wt recipients, unlike Lew.1a recipients, showed significantly increased NK cell ratios upon engraftment in spleen and blood.  Fig  4B. The graph highlights the early graft invasion by macrophages in untreated recipients, especially when compared to recipients after NK cell depletion. (E) Infiltration of syngeneic and allogeneic grafts of Lew.1a recipients assessed by nuclear hematoxylin staining. Due to the intense infiltration of syngeneic grafts it was difficult to analyse additional infiltration of allogeneic grafts. Of note, the degree of infiltration in untreated rejecting and non-rejecting Lew.1a recipients (see Fig 1C and 1D) revealed no significant differences, which is why the allogeneic group was presented undivided as one group. (TIF) S6 Fig. Distribution of CD25 + /FoxP3 + T reg in draining lymph nodes of Lew wt and Lew.1a rats. (A) Flow cytometric analyses were performed using cervical lymph node cells of naïve Lew wt and Lew.1a rats with a distinct gate on the lymphocyte population (including lymphoblasts). Detailed assessment of CD25 + CD4 + T cells revealed a homogenous distribution of FoxP3 + cells. Therefore, a consistent discrimination between activated T cells and (FoxP3 expressing) T reg cannot be made, since additional FoxP3 staining was not performed for all samples. Naturally occurring T reg cells in rat tend to express higher levels of CD25, than activated T effector cells, however this is not always a valid tool for discrimination of these cell populations as was pointed out in both recipient strains. (TIF) S1 File. ARRIVE guidelines. (DOCX)