Carboxyfluorescein succinimidyl ester
Killer Ig-like receptor
The encounter of NK cells with dendritic cells (DC) undergoing maturation may result in the induction of NK cell proliferation. Whether such proliferation involves most NK cells or just a subset has yet to be determined. In the present study we analyzed the nature of such proliferating NK cells by combining carboxyfluorescein succinimidyl ester staining and double-fluorescence cytofluorimetric analysis. Freshly isolated peripheral blood NK cells cultured with LPS and immature DC underwent proliferation; however, proliferating cells were confined to a minor NK cell subset. This subset is characterized by the CD56brightCD16–NKG2A+KIR– surface phenotype (KIR, killer Ig-like receptor). This was further confirmed by the fact that, after cell sorting, only the CD56bright NK cells were able to proliferate in response to the DC stimulus, whereas the CD56dull were not. We also provide evidence that the CD56bright subset is the main source of IFN-γ-producing NK cells, upon interaction with DC. The CD56brightCD16– NK cells express a panel of surface molecules including CD62L, CCR7 and CXCR3 that may allow their homing either to secondary lymphoid compartments or to inflamed tissues. This implies that, in vivo, the interactions between DC undergoing maturation and CD56bright NK cells may occur in different tissues and have different functional implications.
The main priority of the innate immune response is to respond to pathogenic threats with maximal speed and potency, thus providing a first line of defense against infections. In the meanwhile,cells of the innate immunity may cooperate in a coordinated fashion to provide the adaptive immune system with suitable information to start antigen-specific responses. In this context, DC can be considered as true sentinels that play a strategic defensive role in both innate and adaptive immune responses 1–5.
On the other hand, NK cells appear to play a major role in the direct control of early viral infections by killing infected cells and by producing IFN-γ 6–9. In addition, they have been shown to cooperate with DC during the early phases of innate immune responses. Indeed a complex cross-talk may occur between these two cell types 10–13. In the course of these interactions, NK cells appear to exert a quality control by shaping both numbers and functional properties of DC before the DC migrate to secondary lymphoid organs. DC are capable of activating NK cells that acquire the ability to kill otherwise NK-resistant tumor target cells 14–17. In addition, DC-stimulated NK cells can kill immature DC (iDC) while sparing mature DC (mDC) 14, 16, 17. Such DC-induced NK cytotoxicity requires physical contact between the two cell types 14. It is conceivable that the induction of NK cytotoxicity may be a consequence of the effect of IL-12 released by Ag-activated DC 17.
It has been proposed that the process of NK-cell-mediated DC selection represents a checkpoint that provides a mechanism of quality control for DC undergoing maturation 11. According to this view, NK cells equipped with suitable chemokine receptors are recruited together with iDC at sites of inflammation. Recruited NK cells acquire the ability to kill DC that do not express adequate amounts of HLA class I (NKp30 has been identified as the main triggering NK receptor involved in DC killing) 14. Thus, under these conditions, there will be a "positive selection" only of those DC that, undergoing proper maturation, will be capable of efficient T cell priming. In this context, it has been shown that iDC, after their differentiation from monocytes, display a down-regulation of the expression of the HLA-E surface molecules 18. This renders iDC potentially susceptible to killing mediated by NK cells equipped with the HLA-class I-specific inhibitory receptor CD94–NKG2A 18. However, the process of NK-mediated regulation of DC maturation is based not only on the process of quality control of DC but also on the ability of NK cells to favor DC maturation via a mechanism which is likely dependent on the TNF-α that is released by NK cells themselves 15, 16.
Another event resulting from the close contact between NK and DC is represented by the induction of NK cell proliferation. Whereas the process allowing NK cells to kill iDC requires only few hours of NK/DC co-culture, the induction of NK cell proliferation is detectable after 6–8 days. DC pulsing with LPS, BCG or Escherichia coli significantly enhanced the proliferative responses of co-cultured NK cells 17.
Although the studies above provided relevant information on the outcome of DC/NK interactions, the nature of NK cells undergoing proliferation has not been investigated so far. In the present study we analyzed such proliferating NK cells both for their surface phenotype and for their functional capabilities. We show that NK cells undergoing DC-dependent proliferation are restricted to a minor subset of peripheral blood NK cells characterized by the CD56brightCD16–NKG2A+KIR– surface phenotype (KIR, killer Ig-like receptor). This subset also expresses CCR7 and CD62L, thus suggesting its ability to migrate to secondary lymphoid organs 19–21. Finally, these cells in the presence of LPS-pulsed DC produce high levels of IFN-γ that may play an additional regulatory role on the quality of the subsequent adaptive immune response.
2.1 Identification of the NK cell subset undergoing proliferation upon interaction with DC
In these experiments, peripheral blood NK cells that had been purified from healthy donors were cultured together with monocyte-derived DC. Since previous studies indicated that, during DC/NK interactions, efficient NK cell activation and proliferation requires the presence of DC undergoing maturation 15, 16, the NK cell response to DC was analyzed in the presence of 1 μg/ml LPS (i.e. an inducer of DC maturation). Controls included NK cells cultured in the absence of DC with either LPS or IL-2.
In agreement with previous data, NK cells cultured together with maturing DC expressed activation markers 15, 16. As shown in Fig. 1, after 2 days most NK cells cultured with DC+LPS expressed surface CD69. NK cells cultured in IL-2 also expressed CD69 but those cultured with LPS alone did not. Thus, the majority of NK cells became activated upon interaction with DC+LPS. We further analyzed whether NK cells were homogeneously capable of proliferating under these culture conditions. To this end, NK cells purified from PBL were labeled with the green-fluorescent dye carboxyfluorescein succinimidyl ester (CFSE) before culture and were analyzed at various time intervals by cytofluorimetric analysis. On the basis of this analysis, proliferating NK cells were detectable, in cultures containing DC+LPS, after 6–7 days (maximally after 9–11 days).
Remarkably, although virtually all NK cells were activated upon interaction with DC, only a small fraction elicited proliferative responses. This proliferating NK cell subset was assessed for the expression of informative cell surface markers by double-fluorescence cytofluorimetric analysis. These experiments revealed that proliferating cells were characterized by the surface phenotype CD3–CD56+CD16–CD62L+. Importantly this selective proliferation only occurred when NK cells interacted with DC pulsed with LPS. In contrast, in the presence of exogenous IL-2 a non-selective, homogeneous proliferation of both CD16+ and CD16– NK cells occurred (Fig. 2A) (in this case NK cells displayed a partial down-regulation of CD62L expression).
DC-induced, proliferating NK cells were homogeneously stained by anti-NKG2A antibodies whereas they were KIR– as assessed by using a mixture of antibodies specific for various KIR (including KIR2DL1, L2, L3, S1, S2, S4 and KIR3DL1, L2, S1) 22. Regarding the expression of triggering NK receptors 23, DC-induced NK cells expressed medium/high levels of NKp46, NKp30, NKp44 (Fig. 2B) and NKG2D (not shown). It is of note that, differently from CD69, and consistent with previous data, NKp44 24 was expressed at later stages of NK cell activation, primarily by NK cells displaying a shift in CFSE staining (i.e. cells that had undergone proliferation).
Similar results were obtained in a large panel (>15) of normal donors both in autologous and KIR-mismatched NK/DC culture combinations. These data suggest that the HLA class I haplotype of DC does not play a significant role in the process of selection of the proliferating NK cells.
Altogether the surface phenotype displayed by NK cells that were exhibiting proliferative responses upon interaction with LPS-pulsed DC is reminiscent of that of a small subset of CD56brightCD16– NK cells that is detectable in peripheral blood samples from most normal donors 12, 13, 25. This subset accounts for only 3–10% of the total NK cell pool in normal individuals. Thus, it appears that NK cells undergoing proliferation upon interaction with DC originate from a minor peripheral blood cell subset.
2.2 Analysis of isolated CD56++CD16– and CD56+CD16+ cell subsets
DC-induced proliferating NK cells could originate from the CD56++CD16– (i.e. CD56bright) peripheral blood cell subset. Alternatively, they could (at least in part) derive from CD56+NKG2A+CD16+ cells upon modulation of surface CD16. To analyze these possibilities, we cultured purified CD56+CD16+ (i.e. CD56dim ) or CD56bright populations in the presence of iDC and LPS. As shown in Fig. 3, the CD56bright population that was analyzed was characterized by a homogeneous NKG2Abright surface phenotype and lacked both KIR molecules and LIR-1. They expressed high levels of NKp46 whereas the expression of NKp30 and NKG2D was comparable to that of CD56dim cells. Consistent with previous studies 21, CD56bright cells also expressed high levels of CD62L. Importantly, the chemokine-receptors CCR7 and CXCR3 were virtually confined to CD56bright cells whereas CXCR1 and CX3CR1 were mostly expressed by CD56dim cells.
As shown in Fig. 4, when purified CD56bright cells were cultured in the presence of iDC+LPS they displayed strong proliferative responses as assessed by CFSE labeling. On the contrary, the CD56dim population did not proliferate under the same culture conditions. Remarkably, both NK cell subsets underwent proliferation when cultured in IL-2 (not shown).
2.3 CD56bright NK cells produce high levels of IFN-γ in the presence of DC
It has previously been shown that, following interaction with Ag-pulsed DC, peripheral blood NK cells produced IFN-γ 14, 15. We analyzed the levels of IFN-γ in supernatants collected from purified NK cells that had been cultured for 2 days with DC+LPS. Under these culture conditions, large amounts of IFN-γ could be also detected in the supernatants (Fig. 5A). IFN-γ release could not be observed when purified NK cells were cultured with either LPS alone or iDC alone. IFN-γ production appears to be dependent on the presence of IL-12 secreted by LPS-induced DC since IFN-γ production was strongly inhibited by anti-IL-12 antibodies. On the other hand, no effect could be detected upon mAb-mediated masking of NKp30 (Fig. 5A), a triggering NK-receptor that plays a crucial role in the NK-mediated killing of iDC 14. Thus, at variance with previous data on activated NK cells 14, IFN-γ production by freshly derived peripheral blood NK cells does not require NK cell triggering via NKp30, but is induced by IL-12. Notably, purified NK cells, when cultured with supernatants derived from DC that had been pulsed for 2 days with 1 μg/ml LPS, did not release significant amounts of IFN-γ (Fig. 5A). Taken together these data suggest that both IL-12 and DC/NK interactions are needed to induce IFN-γ production by NK cells.
The observation that DC induce a preferential proliferation of CD56bright NK cells prompted us to assess whether IFN-γ production also preferentially resulted from this NK cell subset. In this context, previous studies already suggested that CD56bright cells may represent an NK cell subset capable of prominent cytokine production 25. As shown in Fig. 5B, after depletion of the small CD56bright cell fraction, the remaining NK cells displayed a strongly reduced ability to secrete IFN-γ when stimulated by DC+LPS. These data strongly suggest that, as a result of the DC/NK interaction, CD56bright cells not only undergo strong proliferation, but also represent the major NK cell population secreting IFN-γ.
In the present study, we provide clear evidence that the small population of peripheral blood NK cells that is characterized by the CD56brightCD16– surface phenotype displayssignificant functional responsiveness to pathogen-pulsed DC. Indeed, only CD56brightCD16– cells underwent proliferation and released large amounts of IFN-γ upon culture with DC and LPS.
Although, under the culture conditions used, most peripheral blood NK cells became rapidly activated and expressed CD69, only the minor subset expressing the CD56brightCD16– phenotype and characterized by the expression of CD94–NKG2A and by the lack of KIR was capable of proliferation. Remarkably, the same NK cell subset also represents the main source of IFN-γ production by NK cells cultured with DC and LPS.
In line with previous studies 14–16, the ability of NK cells to respond to DC appears to require cell-to-cell contact. Indeed, in no instance could supernatants, collected at various time intervals from DC/NK cultures, induce either NK cell proliferation or IFN-γ production. Although we cannot rule out that NK cell proliferation might be sustained by one or another lymphokine 26, the DC-driven NK cell proliferation was not inhibited by the addition of several combinations of different neutralizing anti-lymphokine antibodies(including anti-IL-12+anti-IL-15, anti-IL-12+anti-IL-2 or anti-IL-15+anti-IL-2) (not shown). In contrast, the production of IFN-γ by the same NK cell subset required the release of IL-12 by DC since this activity could be abrogated by anti-IL-12 antibodies.
The cross-talk between NK cells and DC appears to be a complex phenomenon that involves bi-directional cell-to-cell interactions 10–13. Thus, it is possiblethat distinct NK cell subsets may be involved in different types of NK/DC interactions occurring in different tissue compartments. For example, we showed that the ability to kill iDC is not a common feature of the whole NK cell population, but it is rather confined to a well-defined NK cell subset characterized by the KIR–NKG2A+CD16+ surface phenotype 18. This NK cell subset is likely to operate a selection process on DC that are present at inflammatory sites by a mechanism of "quality control". This control is essentially based on the ability of NK cells to sense the level of HLA-E expression by DC that are undergoing maturation and to eliminate those with insufficient expression 18.
In the present study we identified an additional NK cell subset with the unique capability of proliferating in response to DC undergoing maturation. The proliferative response of NK cells cultured together with DC was recently documented by Ferlazzo et al. 14. Since the experimental approach was based on [3H]thymidine incorporation, no information could be provided on the nature of proliferating cells. Now, by the use of CFSE staining combined with double-fluorescence analysis, we demonstrate that these NK cells are included in the CD56brightCD16– NK cell population. Differently from the CD56dimCD16+ NK cells, the CD56brightCD16– NK cells express high levels of L-selectin (CD62L) and CCR7 (Fig. 3); these are two surface molecules that allow lymphocyte migration to LN through high endothelial venules 27–29. On the other hand, the CD56bright NK cells lack the expression of chemokine receptors, such as CXCR1 and CX3CR1, that drive CD56dimCD16+ NK cells into inflamed tissues. Thus CD56bright NK cells may prevalently populate secondary lymphoid compartments rather than peripheral tissues.
Since, following Ag capture or pathogen-mediated signaling, DC migrate from peripheral tissues to LN 30, it is conceivable that the interaction between DC and CD56bright NK cells may prevalently take place in these compartments. In this context, recent reports 31, 32 indicated that the NK cells that are present in the T cellareas of LN and tonsils express a surface phenotype (CD56brightCD16–KIR–NKG2A+) identical to that of the NK subset characterized in the present study that proliferate in response to DC+LPS (Fig. 2B). On the other hand the analysis of the chemokine receptor repertoire revealed that CD56bright NK cells also express CXCR3 (Fig. 3), a receptor specific for chemokines (CXCL9, CXCL10 and CXCL11) that are released in the course of acute inflammatory events 27, 29. This may suggest that CD56bright NK cells may also interact with pathogen-stimulated DC in peripheral tissues.
Our studies also suggest that monocyte-derived DC, in the presence of LPS, can selectively induce CD56bright NK cells to secrete large amounts of IFN-γ. This is in line with the current hypothesis that CD56brightCD16– NK cells may represent an NK cell subset specialized in cytokine production 25. As noted above, the NK/DC interaction leading to IFN-γ production could take place in LN where T cells are primed to initiate adaptive immune responses. As a consequence, the release of IFN-γ by NK cells may influence the type of downstream adaptive immune response by polarizing the T cell functional phenotype towards Th1. In this context, the recent evidence that CD56bright NK cells are present in the T cell areas ofLN (see above) may suggest that functional interactions between DC, NK and T cells may occur primarily in these compartments.
Remarkably, in inflamed peripheral tissues, CD56bright NK cells that undergo proliferation may also play a role in DC maturation by enhancing the inflammatory responses through the secretion of IFN-γ. Although CD56bright NK cells have been reported to be constitutively poorly cytolytic 25, our experiments revealed that upon interaction with maturing DC they up-regulate the expression of perforin, thus suggesting that they may acquire cytolytic properties (not shown). Finally, what could be the final tissue target for CD56bright NK cells having undergone proliferation in inflamed peripheral tissues in the presence of maturing DC? They still express CD62L but, differently from the original peripheral blood population, after interaction with DC they have lost the expression of the secondary lymphoid compartment (SLC)-tropic CCR7 receptor (not shown). Thus, although their initial surface phenotype suggests a possible migration to SLC, it is likely that during their entire lifespan these cells may actually be mostly confined within peripheral tissues, at least when they interact with DC in these tissues.
In conclusion, our present study demonstrates that, upon LPS stimulation, DC can induce the effector functions of a well-defined circulating NK cell subset. These NK cells are characterized bythe CD56brightCD16– phenotype and express a pattern of chemokine receptors and adhesion molecules that enable their homing to different body compartments. Through the CXCR3 receptor, circulating CD56bright NK cells can be recruited into inflamed peripheral tissues where, following contact with DC undergoing maturation, may be induced to proliferate, produce IFN-γ and even enhance their cytolytic potential. In this case the DC/NK interaction will strengthen the innate response either by promoting the inflammatory reaction or by enhancing the clearing rate of NK-susceptible pathogens. On the other hand, through CCR7 and CD62L, circulating CD56bright NK cells may also reach secondary lymphoid compartments such as LN 12, 13. At these sites the DC/NK interaction may modulate adaptive immune responses by skewing the T cell response towards a Th1 polarization. Thus, the DC-mediated stimulation of CD56brightCD16– NK cells may acquire different functional significance depending on the site (inflamed peripheral tissues vs. LN) where these interactions actually occur.
4 Materials and methods
The following mAb, produced in our laboratory, were used in this study: JT3A (IgG2a, anti-CD3), c227 (IgG1, anti-CD69), AZ20 and F252 (IgG1 and IgM, respectively, anti-NKp30), BAB281 and KL247(IgG1 and IgM, respectively, anti-NKp46), ECM217 (IgG2b, anti-NKG2D), c127 (IgG1, anti-CD16), c218 (IgG1, anti-CD56), EB6b (IgG1, anti-KIR2DL1 and KIR2DS1), GL183 (IgG1, anti-KIR2DL2, KIR2DL3 and KIR2DS2), FES172 (IgG2a, anti-KIR2DS4), Z27 (IgG1, anti-KIR3DL1), XA185 (IgG1, anti-CD94), Z199 (IgG2b, anti-NKG2A), and F278 (IgG1, anti-LIR-1).
Anti-CXCR3 (IgG1) (R and D Systems, Minneapolis, MN, USA), anti-CX3CR1 (IgG2B–FITC; MBL), anti-CXCR1 (IgG1; Santa Cruz Biotechnologies, Santa Cruz, CA, USA), anti-CCR7 (IgM; Pharmingen, Becton-Dickinson), PC5-conjugated anti-CD62L (IgG1), PE-conjugated anti-CD1a (IgG1), anti-CD14 (IgG2a), anti-CD83 (IgG2b) PE-conjugated anti-CD86 (IgG2b), a mixture of PC5-conjugated anti-CD56 mAb and FITC-conjugated anti-CD3 mAb (Immunotech, Marseille, France), allophycocyanin-conjugated anti-CD56 (IgG1; Miltenyi Biotec, Bergisch Gladbach, Germany) and anti-cytokine (rabbit antisera; Peprotech, London, GB) were commercially available.
D1.12 (IgG2a, anti-HLA-DR) mAb was provided by Dr R. S. Accolla (Pavia, Italy). HP2.6 (IgG2a, anti-CD4) mAb was provided by Dr P. Sanchez-Madrid (Madrid, Spain).
4.2 Generation of polyclonal NK cell populations and CFSE labeling
NK cells were derived from healthy donors as previously described 18. Briefly, PBMC were isolated on ficoll-hypaque gradients and depleted of plastic-adherent cells. Enriched NK cells were isolated by incubating PBL with anti-CD3 (JT3A), anti-CD4 (HP2.6) and anti-HLA-DR (D1.12) mAb (30 min at 4°C) followed by goat anti mouse-coated Dynabeads (Dynal, Oslo, Norway) (30 min at 4°C) and immunomagnetic depletion. In selected experiments, the CD56bright and CD56dull NK cell subsets were separated (according to CD56 expression) by FACS cell sorting (DIVA; BD Biosciences). Reanalysis of the isolated subsets using a FACSCaliburTM (BD Biosciences) showed that cell purity was at least 98%.
To obtain NK cells labeled with CFSE, PBL were washed twice in serum free RPMI-1640, and incubated at a final concentration of 107 cells/ml in 2 μM CFSE (Molecular Probes) for 7 min at 37°C 33. Labeled cells were washed twice in RPMI-1640 containing 10% FCS and were then subjected to the purification procedures described above.
4.3 Generation of DC
DC were generated as previously described 18. Briefly, PBMC were derived from healthy donors, and plastic-adherent cells were cultured in RPMI-1640 containing 10% FCS, in the presence of IL-4 and GM-CSF (Peprotech) at final concentrations of 20 ng/ml and 50 ng/ml, respectively. After 6 days of culture, cells were characterized by the CD14–CD1a+CD83– phenotype corresponding to iDC. To generate CD14–CD1a+CD83+CD86+ mDC, the iDC were stimulated for 2 days with LPS (Sigma-Aldrich, St. Louis, MO, USA) at afinal concentration of 1 μg/ml.
4.4 NK/DC cultures and proliferation assay
NK cells were cultured in RPMI-1640 containing 10% FCS, in 96-well round-bottomed microtiter plates (105 cells/well) in the presence of either iDC + 1 μg/ml LPS (NK:DC ratio 5:1) or 1 μg/ml LPS alone. As positive control, NK cells were also cultured in the presence of 600 IU/ml rhIL-2 (Proleukin, Chiron Corp., Emeryville, CA, USA). After 48 h, NK cells were stained with the anti-CD69 mAb and assessed by cytofluorimetric analysis.
To assess DC-induced NK cell proliferation, the various NK cells populations (including total NK cells, the CD56bright NK cell subset and the CD56dull NK cell subset), labeled with the green-fluorescent dye CFSE, were cultured in the above culture conditions. After 9–11 days, cells were harvested, stained with the mAb indicated in the text and assessed by double-fluorescence cytofluorimetric analysis.
4.5 Flow cytofluorimetric analysis
For one-, two- or three-color cytofluorimetric analysis (FACSCalibur; Becton Dickinson & Co, Mountain View, CA, USA), cells were stained with the appropriate labeled or unlabeled mAb. In the case of unlabeled mAb, staining was followed by PE- or FITC-conjugated isotype-specific goat-anti-mouse second reagent (Southern Biotechnology Associated, Birmingham, AL, USA) as described previously18.
4.6 IFN-γ secretion assays
NK cells (derived either from the total circulating NK cell population or from the CD56dull NK cell subset) were cultured in 96-well round-bottomed microtiter plates (105 cells/well) under the experimental conditions indicated in the text. The supernatants of the cultures were collected after 48 h and assayed by ELISA (Biosource International).
4.7 mAb-mediated blocking experiments
In the blocking experiments, saturating amounts of one or another of the following reagents were added at the onset of the cell cultures: anti-IL-12, anti-IL-15, anti-IL-2 neutralizing antisera, anti-NKG2D (BAT221; IgG1), anti-NKp30 (F252) and anti-NKp46 (KL247). The anti-cytokine antisera were also added in different combinations (anti-IL-12 + anti-IL-15; anti IL-12 + anti-IL-2; anti IL-15 + anti-IL-2).
This work was supported by grants awarded by Associazione Italiana per la Ricerca sul Cancro (A.I.R.C.), Istituto Superiore di Sanità (I.S.S.), Ministero della Salute, Ministero dell'Istruzione dell'Università e della Ricerca (M.I.U.R.), Ministero dell'Università e della Ricerca Scientifica e Tecnologica (M.U.R.S.T.). The financial support of Fondazione Compagnia di San Paolo, Torino, Italy, is also gratefully acknowledged.