Innate immunity is the ancient defence system against infection that is conserved in most of organisms including prokaryotes and invertebrates throughout evolution. Among the cells that participate in the early innate immune response, natural killer (NK) cells occupies special positions, not only owing to its unique cytotoxic ability but its role between the innate and adaptive immunity put it to the high level of interest. The NK cells were first discovered in 1975 by two groups, R.Kiessling and R.Herberman when studying specific cytotoxic effect of lymphocytes against target tumour cells. The observation of the NK cells brought many controversial issues because of its reactivity level was observed from not only limited to the animals showing the active immune response, but also to those non-immunised one. After several studies suggested the capacity to kill the tumour cell without prior sensitisation, NK cells has been accepted as a sub-population of the lymphocytes, which conduct highly regulated killing mechanism.
The current understandings of NK cells originate from the models of missing-self and induced-self recognition. These models introduce that NK cells unlike other T or B lymphocytes do not recognise foreign antigens but are initiated by detecting changes in the surface molecules of the cells. Such phenomenon describes that certain cells expressing less major histocompatiblity complex (MHC) class I are the major target of the NK cells. This MHC class I dependent recognition is responsible for the killing of virus-infected or malignant tumour cells with deficient expression of MHC class I are attacked by NK cells, whereas normal healthy cells are not. However, missing-self model does not provide sufficient explanation on how the NK cells recognise their target cells, when those autologous cells that do not express MHC class I molecules, such as erythrocytes are counted. The 'self-induced' hypothesis complements the 'missing-self' model by proposing that NK cell detects cell stress-induced self-ligands that trigger activation of NK cells. Taken together, it is now well established that activation of NK cells depends on a delicate balance between activating and inhibitory signals. Normal autologous cells express inhibitory signals and no activating ligands, whereas malignant or stressed cell actively expresses activating NK ligands with downregulated MHC class I molecules.
Upon the activation, NK cell functions can be categorised into three;
I. Cytotoxicity: NK cells perform their cytotoxicity by conducting two mechanisms. First, they exocytose cytotoxic granules containing perforin and various granzymes, resulting in the apoptotic death induced by permeated granzymes. Second, they express tumour necrosis factor (TNF) receptor superfamily (TNFRSF) members such as FAS that engage with corresponding ligands to induce apoptosis. Third, NK cells hire TNF-related apoptosis-inducing ligand (TRAIL) as a cytotoxic effector molecule.
II. Regulation of immune response through various cytokine production: NK cells produce many cytokines such as IFN-g, IL-3, GM-CSF, TNF-a and chemokines such as MIP-1, RANTES and IL-8
III. Direct contact co-stimulation: Via costimulatory ligands, such as CD40 and OX40L, NK cells induce T cells and B cells stimulation.
Recently, Tbox21 transcription factor (T-bet) is emerging as a crucial factor in NK cell development and maturation. Several studies have demonstrated the defected functions of T-bet KO NK cells in the aspect of cytokine production and NK cell mediated killing. How T-bet transcription factor participates in the NK cell development will be discussed in detail in the introduction.
NK cell development
The interactions between the precursor NK cells and bone marrow stromal cells are pivotal in the development of NK cells. The NK precursor cells in mice showing high expression of CD122, one of the four subunits of IL-2/IL-15 receptor are constantly influenced by IL-15 in the early stages of NK-cell development, especially during the stage of cellular expansion. These immature NK cells that express NK1.1, acquire CD94-NKG2 heterodimer on their surface as well as Ly49 receptor prior to the proliferation phase. After proliferation, terminal maturation is accompanied by the expression of high levels of CD11b and CD43 along with the perfected cytolytic function and IFN-g production. These interaction with various cytokines are not only confined to the mice NK cells, but immature human NK cells are observed to be perform their cytotoxicity in the manner of TNF-related apoptosis-inducing ligand and release type 2 cytokines in their early development.
To elucidate the NK cell development, transcription factors are undeniably key elements as several studies already revealed several transcription factors including interferon regulatory factor(IRF)-1, IRF-2, ETS-1 and T-bet regulates the ability and development of NK cells. For instance, those mice missing GATA-3 transcription factor that are important in Th2 cytokine production, produce NK cells with immature phenotype and also influence the production of IFN-g. Additionally, Ets-family member myeloid elves like factor (MEF) also participates in the determination of NK cell function that its deficiency leads to the failure of expressing perforin and have cytotoxicity. Furthermore, the T box 21(T-bet) transcription factor knockout mice are found to carry significantly less number of NK cells expressing reduced levels of CD43 and CD11b and show receded IFN-g production. T-bet transcription factor will be discussed in more depth later.
Once fully matured NK cells are emerged, it is interferons and cytokines that affect expansion and modulate the function of NK cells. As mentioned briefly, IL-15 and IL-15 receptor (CD122) plays pivotal role in the early development of NK cells. Having said that, IL-15 and its receptor also contribute to the survival and maintenance of mature NK cells. However, it seems CD122, the IL-15 receptor is not solely functioned by the interacting with IL-15 as NK cells from the IL-15R KO mice develop as normal when transferred to a host where the non-hematopoietic cells express IL-15R. Rather, the concentration of IL-15 on stromal cells is decisive in the maturation of NK cells.
Having said that IL-15 contributes to the early development and maturation of the NK cells, IL-21 rather enhances the cytotoxic ability and production of IFN-g. Yet, the clear role of IL-21 is still controversial, since the earlier results claiming the importance of IL-21 for the differentiation of NK cells conflict with the latter showing the normal development of NK cells regardless neither the existence nor the concentration of IL-21. Nevertheless, some studies suggested the antagonistic influence of IL-21 on the cellular expansion of NK cells.
Transcription Factors in NK cell development
It is now a well-known fact that, as with all cells of the hematopoietic system, NK cells are derived from bone-marrow hematopoietic stem cells. From the previous studies, it has been generally accepted that NK cells development is defined into several stages including precursor NK cell (pNK), immature NK cell (iNK), and mature NK cell (mNK). According to the recent publish by Vosshenrich et al, six stages are suggested (Fig1), taking into account up-to-date studies. The first stage A is the precursor NK cells as defined by Rosmaraki et al that express high CD122 (IL-2Rb) but lack other known NK cell marker including NK1.1, DX5 and Ly49R. The two subsequent stages, B and C are immature NK cells (iNK). At stage B, the cells are yet inadequately cytotoxic, but start to possess exclusive NK markers, NK1.1 and NKG2D (but not DX5 nor Ly49R). Ly49 receptor begins to show in the following stage. Vosshenrich defined the mature NK cells by observing high expression of DX5 which appears from the stage D. Two additional markers that are initially appearing from the stage C are upregulated as a functional markers; CD11b expression is elevated first (stage E), followed by CD 43(stage F). The fully matured NK cells attain the cytotoxic ability and also can pr NK cells are fully functional and also can produce IFN-g after the stimulation. These mNK cells will then migrate to the periphery via blood.
Each developmental stage of NK cells is sophisticatedly controlled by diverse transcription factors. These factors can be grouped into three, according to the stages they correlate to (Table 1); the first group includes Ikaros, Id2, Ets-1 and PU.1 and are known to control the population size and maintenance of early precursor NK cells. Additionally, these transcription factors are essential to establish the early expression of CD122 (IL-15Rb), that the previous studies by two different groups, Ikawa et al and Boggs et al both observed the impaired expression of CD122 from Id2 KO and Ikaros KO mice respectively and therefore, reduced population of NK cells along with severely defected cytotoxic ability and cytokine production are observed. The second group of the transcription factors includes Gata-3, IRF-2 and T-bet. These appear to be primarily involved in the maturation of the NK cells specifically the induction of cytotoxic ability as well as migration to the peripheral site. Each deficiency results in the decreased population of NK cells in the liver, spleen and all the peripheral regions respectively, yet the normal NK cell population in the bone marrow is found. Additionally, the peripheral NK cells missing those transcription factors are found to be immature as they adapt low expression of CD11b and CD43. These functionally defected phenotypes are well demonstrated that severely decreased production of IFN-g is resulted from all IRF-2, Gata-3 and T-bet deficient NK cells. However, the defected NK cells from Gata-3 and T-bet KO mice performs essentially normal cytotoxicity while IRF-2 deficient NK cells exhibit severely reduced cytotoxic ability. This observation support the previous reports that the maturation marker, CD11b and CD43 may not have high correlation with cytotoxic ability, rather DX5 has higher association.
The third group of the transcription factors includes myeloid elf factor (MEF), microphthalmia-associated transcription factor(MITF) and CCAAT/enhancer binding protein-g (CEBP-g). These TFs are reported as a regulator of functional development of NK cells as their deficiency results in the significantly reduced cytotoxic ability and to produce cytokines after stimulation.
Secretroy lysosome in NK cells and lytic synapses
NK cells produce cytolytic granules during their development and maturation. These granules that possess acidic lumens as well as several lysosomal hydrolases and marker proteins hire their own special transport systems that allow highly regulated secretion and fusion with the plasma membrane. Defects in the formation or trafficking of the lysosome result in severe immunodeficiency. Upon the target cell recognition, these granules are trafficking to the site of NK cell-target contact site and fuse with the plasma membrane. Subsequently, several soluble effector molecules and lysosomal hydrolases are secreted into the lytic synapses.
Structurally, lytic granules formed in the NK cells are defined as secretory lysosomes that are distinguished from a conventional lysosome due to the specialised transport systems. Nevertheless, these granules have structural similarities to the conventional lysosomes in the manner that both lysosomes are accessible via the endocytic pathway. Conventional lysosomes usually have multi-vesicular structure, whereas secretory lysosomes have dense cores or multi-laminar appearances. Both conventional and secretory lysosomes possess specialised functional proteins such as lysosomal acid hydrolase, receptor proteins and heat shock 70 family proteins. Each is responsible for degradation of 'old protein', recognition and delivery respectively. There are also other resident membrane proteins such as Lamp1 and CD63 which are largely used as activation markers of cytolytic cells including CTL and NK cells. Lamp1, also known as CD107a will be discussed in detail later in this chapter. In addition to the common components in both conventional and secretory lysosomes, there are exclusive lytic molecules in the granules secreted by NK cells. These include Granzyme A & B, Granulysin, and perforin. The detailed function of these molecules will not be discussed here.
Exocytosis of lytic granules in the NK cell involves several stages beginning with target cell recognition. Most importantly, NK cell-lytic synapses are constructed in where NK cells directly contact to the target cells. The formation of lytic synapses is defined into a series of distinct steps. The first step is the initiation stages that include establishing a close-cell to cell contacts. NK cells either accidentally or possibly by chemotactic attraction, gets in contact with target cells. This step involves the interaction between NK receptors and various ligands on the surface of the target cells such as NKG2DL. The initial contact could subsequently lead to the adherence of NK cell to the target cell, followed by activation of NK cells. The integrin family of adhesion molecules expressed on NK cells including lymphocyte function-associated antigen 1 (LFA1) and macrophage receptor 1 (MAC1) build the firm adhesion to the target cells that facilitate the maturation of the lytic synapses. Additionally, integrins may also participate in signalling that can fully activate some NK cells. Yet, it may not involve the full achievement of NK cytotoxic abilities. Importantly, these initial steps in the formation of the NK cell lytic synapse probably occur before molecular patterning or polarisation are evident and completed quickly. Whether the NK cell progresses to molecular reorganisation at the synapse seems to depend on the level of signals received through inihibitory receptors such as killer cell immunoglobulin like receptors, which can establish the so-called inhibitory synapses. Such regulation ensures that NK cells effectively carry out their surveillance function by leaving most cells undisturbed while being ready to destroy cells that are diseased. The inhibitory synapses are especially elegant in that it directly interferes with the ability of the lytic synapse to progress past the initiation stage.
The subsequent stage is the effector stage. Here, the cruicial steps of the granule exocytosis take place; 1. Formation of an immunological cleft where the granules are secreted. 2. Trafficking of granules towards the synapses, 3.fusion of the granule membrane with plasma membrane for the release of the lytic molecules. Under the condition that synapse initiation has occurred in the absence of the inhibitory signals, reorganisation of the immunological synapse is initiated. This begins from the actin reorganisation that involves the formation of F-actin networks from the cellular pool of monomeric globular actin (G-actin) in association with Wiskott-Aldrich syndrome protein (WASP). The absence of WASP is reported to have decrease in the F-actin accumulation at the synapse and NK cell cytotoxicity. As actin reorganisation is processed, following events occurs; receptor clustering, lipid-raft aggregation, further activation signalling, and lytic granule redistribution. Although the detail of such events is not clarified, it seems that these are essential for both cell adhesion and triggering of cytotoxicity. Especially, it has been only shown that although receptor clustering is crucial in the signalling within the NK cells, so far, microclusters has been indentified at supramolecular inhibitory cluster. During the effector stage of lytic synapse, polarisation of lytic granules also takes place. The granule moves along the microtubules towards the microtubules organising centre (MTOC) in association with motor proteins such as Kinesin family proteins; Kinesin protein families have a range of motor proteins that are largely involved in the intracellular trafficking of vesicles and organelles, yet the exact motor used is not revealed. MTOC, in the mean time, begins to polarise towards the immunological synapses, and consequently, the granules move to the NK-target cell contact site. Prior to the fusion of lytic granules to the plasma membrane, there are few final steps of effector stage, which are docking of lysosomes to the synapse and vesicle priming. The vesicle priming is the essential preparation step for fusion through acquisition of biochemical attributes. After vesicle priming, the lytic granules can fuse with the plasma membrane, releasing the granule contents into the cleft between the cells. the mechanism underlying these steps still needs to be elucidated, but information from CTL may give a clue to this process. In CTLs, docking to the plasma membrane requires members of the RAB family of small GTPases, which are important regulators of vesicle trafficking and compartmentalisation. It undergoes post-translational prenylation to gain hydrophobicity which facilitates its interaction with a target membrane. Rab27a is found in NK cells and shown to interact with Unc13D (also known as Munc13-4) for the vesicle priming. Rab27a-Unc13D complex targets solule N-ethylmaleimide0sensitive0factor accessory-protein receptor (SNARE) family protein acting upon the facilitation the fusion of membranes. SNARE proteins are categorised into two; proteins that can be found on the vesicle membrane is called vSNARE whereas those present on the target membrane is tSNARE. Interaction between tSNARE and vSNARE is required for membrane fusion. So far, Vamp7 is the only identified tSNARE protein that involves in NK cell killing.
The Final stage of lytic synapse is the termination stage after the lytic granules are exocytosed. This stage involves a period of inactivity and downmodulation of the accumulated activating receptors followed by detachment of NK cells from the target cells. Once the NK cells exert its cytolytic function, it detaches from the target cell and restore its cytolytic potential by regenerating new lytic granules and re-expresses activating receptors.
Activation signal of NK cells.
Unlike T and B cells, NK cells do not have antigen specific receptors, instead
Dendritic Cell and NK cell interaction
There have been significant advances in the understanding of the activation and function of NK cells both in human and mice. Since the report by Fernandez at al in 1999 described the DC-NK interaction, it is now well-understood that NK cells and DCs activate one another during an immune response.
DC are crucial regulators of both innate and adaptive immune response. As an antigen presenting cell (APC), DC efficiently process antigens in association with MHC molecules. Yet, their functions are followed by the complete maturation of DC triggered by direct encounter to microbial ligands. Once exposed to the pathogen, DC migrate to the T cell area of lymph nodes at where APC-T cell interaction begins.
DC-Mediated NK cell activation: How DC contribute to the activation of the NK cells has been well documented by many different experimental systems using different sources of DC including mouse DC cell lines, mouse bone marrow derived DCs, human monocyte derived DCs and human cord-blood-derived DCs. While these studies generally agree on the key role of DC-derived cytokines and surface molecules in driving NK cell cytotoxicity, the mechanisms involved has been emerged carefully suggested to explain the IFN-g production both in human and mice: 1) Direct cell contact induces the production of IL-12 necessary for NK activation, and other cytokines. 2) DC-derived IL-2 promotes NK cell functions in association with other cytokines.(Fig1). In addition, the maturation state of the DCs might influence their ability to activate NK cells. Several studies have shown that immature DCs require a maturation stimulus to activate NK cells, whereas others have shown that immature and mature DCs are equivalent in their ability to activate NK cells. Nevertheless, it is inevitable to accept that both IL-12 and IL-2 are the key cytokines for the NK cell activation process. IL-12 released after DC-NK direct cell contact appears to be presented to NK cells through the formation of some kind of "synapse" which greatly increases the efficiency of IL-12 even in the low dose. Many studies suggested the role of IL-12 as a key modulator that induces NK cytotoxicity, yet it is still unclear as there has been a suggestion of possible need of other soluble factors. In addition to IL-12, other cytokines such as type 1 IFN and IL-18 may sustain the functions of NK cells in terms of IFN-g production, migration, cytotoxicity and proliferation. IL-18 appears to be playing central role in promoting the migration of NK cells to secondary lymphoid organs where they encounter DC and produce high level of IFN-g once exposed to IL-12, IL-2 and TNF-a. In addition, NK cells showed enhanced cytotoxicity in association with IL-18. IL-2 is now generally accepted as a key cytokine to generate the NK cell activation both in vivo and in vitro, but it had been excluded from the list of those activating NK cells in vivo as it had been understood as an exclusive product of T cells. Soon as it was figured that IL-2 was also produced from activated DC during the first hours of stimulation, Zanoni and his colleagues has established a work based on the theory that IL-2 may have a physiological role in activating NK cells. They discovered that TLR-dependent full maturation stimuli drove DC to elicit NK cell activation via IL-2 interaction whereas TLR-independent stimuli did not induce the release of IL-2, neither the activation of NK cells. Moreover, NK cell activation process in human has also been revealed recently by Newman et al that the capacity of human NK cells to produce IFN-g is dependent upon cell contact-dependent and IL-2 mediated signal derived from myeloid DC, and this interaction may be driven via cell-cell direct contact in association with TNF expressed by DC that binds to RNFR2 on the surface of NK cells.
Having said that NK cells mediate their cytolytic ability upon their activation via various stimuli, it is an interesting question to address how NK cells will react to the different strength of activating signals. Besides, NK cells can also be activated by both immature and mature DC in vivo. Thus, it is worth to assess the effect of both immature DC and mature DC on NK cell activation using various conditions.
The aim of the first part of this study is to investigate the expression level of CD107a on NK cells activated by either different strength of stimuli or different maturation state of DC. This will be done by using mouse splenic NK cells. With these NK cells, it should be able to observe the response of NK cells under different conditions and DC-mediated NK activation.
Lastly, there has been many data published on Tbox21 transcription factor (T-bet) as a essential regulator of NK cell development and maturation. It is well-established that T-bet deficient NK cells have diminished functional activities such as cytotoxicity or cytokine production. However, it is uncertain whether this diminished cytolytic abilitiy is owing to reduced number of lysosome or due to disturbance in the lysosomal trafficking. Thus, it brings an interesting point to have a look at the underlying mechanism of how T-bet is involved in the NK cell mediated killing.
The aim of the second part of this study is to compare the quantity of lysosomal content in both WT and T-bet KO NK cells. Subsequently, RNA level of genes that are likely to be involved in the lysosomal trafficking will be investigated using real-time PCR techniques. This should be able to show the influence of T-bet transcription factor on NK cytolytic abilities, and hopefully identify the genes involved in the trafficking mechanism.
Materials & Methods
Mice and Regents.
BALB/c and C57BL/6 mice were obtained from Harland UK and T-bet KO mice from (Taconics, USA). Cell lines and primary cells were maintained in medium containing RPMI supplemented with Penicillin/Streptomycin, 10% FCS (fetal calf serum), L-glutamine, non-essential amino acid, pyruvate, hepes, and 2-mercaptoethanol.
The mice were sacrificed by cervical dislocation. Mice were placed on their right side and an incision was made on the left side skin with a pair of scissors. The incision was made approximately 2.5 cm long, from between the last rib and the hip joint. Another incision (1-2 cm) was made in the peritoneal wall and the spleen, a red enlongated bean-shape organ, was pulled onto the exterior wall. The spleen was removed by cutting off the mesentery and connective tissues attaching the spleen to the abdominal cavity.
NK Cell Enrichment (DX5 positive enrichment & KIT negative enrichment)
Splenic NK cells obtained from BALB/c, C57BL/6 and T-bet KO mice were enriched using two MACS enrichment protocols; DX5 Positive enrichment and KIT negative enrichment.
- DX5 positive enrichment: splenocytes were obtained extracted using MACs buffer (Sterile PBS with 2nM EDTA and 10% FCS) and centrifuged for 5min at 1200rpm. Using 10ml of MACs buffer, the cells were resuspended, and then counted with trypan blue dye at 50% concentration. Once the cell number was determined, remaining cells were centrifuged for 10min at 1200rpm and resuspended with MACs buffer (amount varies). Subsequently, DX5 microbeads (Miltenyi) was added to the cell according to the manufacturer instruction and incubated in fridge for 10min. The cells were then washed twice with MACs buffer and resuspended with selected amount of MACs buffer. (Amount varies according to the cell count). Using the LS column (Miltenyi), NK-depleted splenocytes were isolated by MACS magnetic cell sorting (Miltenyi). After three times wash, remaining NK cells in LS columns were collected.
- KIT negative enrichment: splenocytes were obtained extracted using MACs buffer (Sterile PBS with 2nM EDTA and 10% FCS) and centrifuged for 5min at 1200rpm. Using 10ml of MACs buffer, the cells were resuspended, and then counted with trypan blue dye at 50% concentration. Once the cell number was determined, remaining cells were centrifuged for 10min at 1200rpm and resuspended with MACs buffer (amount varies). According to the manufacturer instruction, biotin antibody cocktail (Miltenyi) was added to the cell and incubated in fridge for 10min, followed by microbeads coated with anti-biotin(Miltenyi) and another 15min incubation. The cells were then washed twice with MACs buffer and resuspended with selected amount of MACs buffer (Amount varies according to the cell count). Using the LS columns (Miltenyi), flow-through containing NK cells were isolated by MACs magnetic cell sorting (Miltenyi). Additional 3ml of MACs buffer was applied to the LS column to collect all NK cells.
NK cell Sorting using FACS Aria
The splenic NK cells were obtained from WT and T-bet KO mice and enriched using magnetic cell sorting method. The enriched cells were counted with trypan blue dye. Once the cell number was determined, they were stained with either NKp46-Alexa647 or NKP46-biotin followed by Streptavidin-PE, and CD3-APC or CD3-FITC. After the 30min incubation period, enriched NK cells were washed three times with MACs buffer and analysed using FACs Aria II(BD Science) to check the purity. NK cell purity was 60~80% after MACs cell enrichment. Enriched CD3- NKP46+ NK cells were sorted using FACs Aria. Purity after sorting was 97~99%. Sorted cells were directly frozen in Trizol and kept under -80oC.
Plate coating and FACs staining of splenic NK cells
The high-binding well plate coated with a-NK1.1 and aNKG2D was prepared a night before the NK cell stimulation. To prepare the antibody coated plate, the mastermix was prepared at two different concentrations; 50ug/ml and 10ug/ml for a-NK1.1 and a-NKG2D. Each mastermix contained 100ul of carbonate buffer with 5ul or 1ul of antibodies according to the concentration. As a control mastermix, 1ul of isotype antibody was mixed into 100ul of carbonate buffer. The mastermix was dispensed to the well plates and gently swirled for the even spread-out. The plate was incubated in the fridge over night. Next day, splenic NK cells were obtained from C57BL/6 mice. After the cell enrichment, the NK cells were centrifuged and resuspended with the mastermix containing 4ml of medium, 2.6ul golgi stop and 30ul a-CD107a Alexa488 (The amount was determined according to the manufacturer instruction). Prior to the dispense of NK cells onto the well plate, the plate was washed three times with medium. In each well, 200ul of cells were applied and spun for 5min at 800rpm. Subsequently, it was incubated for 4hrs at 37oC. After the incubation period, the cells were transferred to the regular 96 well plate and washed three times with medium.
The second mastermix were prepared according to the manufacturer instudction. It contains ;1) 1.25ml of MACs buffer with 10ul CD3-APC and 20ul NKP46-biotin 2) 1.25ml of MACs buffer with 8ul streptavidin-PE. NK cells were resuspended with the mastermix containing CD3-APC and NKP46-biotin and incubated on ice for 30min. Next, it was washed three times with MACs buffer and stained with the second mastermix with streptavidin-PE. After the incubation, the cells were washed three times and analysed with LSRII (BD science). Propidium Iodide was added just before the analysis in order to recognise the dead cell population.
DC culture & NK stimulation
After removing all muscle tissues with gauze from the femurs and tibias, the bones were placed in a 60-ram dish with 70% alcohol for 1 min, washed twice with PBS and transferred into a fresh dish with RPMI 1640. Both ends of the bones were cut with scissors in the dish, and then the marrow was flushed out using 2 ml of RPMI 1640 with a syringe and gauge needle. The tissue was suspended, passed through nylon mesh to remove small pieces of bone and debris, and red cells were lysed with lysis buffer. After washing, lymphocytes and Ig positive cells were killed with a cocktail of mAbs and rabbit complement for 60 min at 37oC. The mAbs were GK 1.5 anti-CD4, HO 2.2 anti-CDS, B21-2 anti-Ig, and RA3-3A1/6.1 anti-B220/CD45R. 7.5-10 x 10 s cells were placed in 24-well plates in 1 ml of medium supplemented with 500-1,000 U/ml GM-CSF. The cultures were usually fed every 2 d by gently swirling the plates, aspirating 75% of the medium, and adding back fresh medium with GM-CSF. An object of these washes was to remove non-adherent granulocytes without dislodging clusters of developing dendritic cells that were loosely attached to firmly adherent macrophages.
DC maturation was induced by incubating the cell overnight with 0.5ug/m; of LPS. DC was co-cultured with splenic NK cells negatively sorted by NK cell isolation KIT. After 4 hr incubation, NK cells were labelled with CD3-APC, NKP46-Biotin, Streptavidin-PE and analysed by flow cytometry.
4x106 WT NK cells and 2.3x106 T-bet KO NK cells were kept in trizol under -80oC. 0.2ml of chloroform was added to the sample and centrifuged at 12,000g for 15min. The Top layer containing RNA was extracted and mixed with 500ul of isopropanol. Subsequently, the samples were centrifuged at 12,000g for 10min and 1ml of 75% ethanol was added. Samples were centrifuged at 74,000g for 5min and excess ethanol was carefully removed. Samples were allowed open in air to dry any remaining enthanol. 20ul of nuclease free water was added. Nanodrop (spectrometer) was used to analyse the samples. The amount of RNA from WT and T-bet KO NK cells were 71.5 and 38.6ng/uL respectively.
Reverse Transcription PCR.
cDNA was synthesised using the RNA obtained from WT and T-bet KO splenic NK cells. Mastermix was prepared containing Oligo dT(2ul), PCR Mastermix (4ul), Reverse transcriptase (1ul for RT+) and nuclease free water. 4 PCR eppendorf tubes were labeled as WT+ WT- T-bet+ and T-bet- (negative sign indicates RT- tubes for control). The equal amount of RNA was applied to each tube, followed by the mastermix prepared. The samples were placed in the PCR machine and amplified as the program instructed: 90min 42oC, 5min 85oC.
The amount of RNA of interest in splenic NK cells was determined using real-time PCR. The mastermix was prepared containing Taqman mastermix, Beta-actin control, Gene expression assay and cDNA prepared. As instructed, duplex PCR was performed using total volume of 10ul. On the 384 well plate, 5ul of Taqman mastermix, 0.5ul of gene expression assay, 0.5ul of ß-actin control, 3ul of nuclease-free water and 1ul of sample was applied to each well. Once prepared, the 384 well plate was placed in the real-time PCR machine and run as instructed.
CD107a expression is elevated on NK cells following the stimulation.
To investigate whether the activation of NK cells is influenced by intensity of stimulus, the CD107a expression was examined using two different conditions. As previously reported, CD107a is a marker for cytotoxic ability of NK cells and CD8+ cytotoxic T cells as it is expressed at high levels upon both cells being stimulated. Two different concentrations of either a-NK1.1 or a-NKG2D were harboured. Both NK1.1 and NKG2D are reported as receptors of activation of which initiate transmembrane signals that activate cytotoxicity. Freshly isolated splenic NK cells from C57/BL6 mice were incubated on the high-binding 96 well plate coated with a-NK1.1 and a-NKG2D. Each antibody was coated at two different concentrations of 10ug/ml or 50ug/ml. After 4hr incubation, NK cells were stained for CD3 and NKP46. Representative data from one subject is shown in Fig. 1. Surface expression of CD107a was low in unstimulated NK cells (0.32%) (Fig.1A). Unpurified spleen NK cells stimulated by a-NK1.1 at two different concentrations expressed CD107a at much elevated levels (9.35% and 9.31% for 50ug and 10ug concentration respectively) (Fig.1C,D). The unpurified spleen NK cells stimulated with a-NKG2D at 50ug/ml and 10ug/ml both showed increased expression (8.28% and 8.45% respectively), yet it is faintly lower than those stimulated with a-NK1.1. Compared to the unpurified splenocytes, the negatively sorted spleen NK cells by NK cell isolation KIT showed inconsistency that, although the expression of CD107a was all increased on NK cells stimulated with NK1.1 or NKG2D at both concentrations, the degree of elevation was minor except those stimulated with NK1.1 at 50ug/ml concentration (7.32%). Other three showed 4.28%, 3.92%, 3.3% at NK1.1 10ug/ml, NKG2D 50ug/ml, and 10ug/ml respectively. Taken together, these data demonstrate that CD107a is highly upregulated on the surface of NK cells following stimulation, but the intensity of stimuli did not greatly affect on the activation level.
Having said that the CD107a expression was upregulated following the stimulation using a-NK1.1 and a-NKG2D, DC-mediated NK cell activation was subsequently examined, as DC is a source of cytokines required to activate NK cells in vivo. To assess whether the DC-mediated NK cell activation is regardless the maturation of DC, bone-marrow DC from C57BL6 mice (approximately 80% purity) was subdivided into two populations; one incubated overnight with LPS for the stimulation, and the other left untreated as a immature DC. Negatively sorted splenic NK cells by NK cell isolation KIT was co-cultured with DC in the well plate and incubated for 4hrs. Next, NK cells were stained for CD3 and NKP46. As a control, purified NK cells with no culture were hired to compare the CD107a expression. In a control group, hardly detectable expression of CD107a (0.02%) was observed, indicating most of the NK cells were left inactivated as expected (Fig.2.A). The frequency of activated NK cells co-cultured with iDC was appeared to have higher; 3.05% and 2.31% for iDC co-culture at 1:1 and 1:5 ratio (NK:DC) respectively. It clearly demonstrates that iDC can induce the activation of NK cells. However, it should be noted that the LPS-stimulated DC co-culture did not induce the activation of NK cells so efficiently as the immature DC as expected (0.8%, 1.77% and this conflicts with other data published by others that the mature DC elevates the NK functions in all manners. This may be due to the ratio adopted here that NK cells possibly induce their cytotoxic ability upon DC when there is too large number of DC. To assess such an issue, further investigation is required by adopting various environments of DC stimulation such as IFN-a or CD40L. However, although the expression of CD107a on NK cells simulated by LPS-stimulated DC was beyond the expectation, these results demonstrated DC can mediate NK cell activations regardless their maturation.
Taken together, CD107a is upregulated upon the activated NK cells stimulated by either a-NK1.1 and/or a-NKG2D or bone-marrow DC, indicating that NK cells initiates the production of cytolytic molecules. Both results may indicate that the activation of NK cells is regardless to the intensity of stimulation since different concentration of a-NK1.1 or a-NKG2D, or different ratio of DC:NK does not produce noticeable difference in the CD107a expression. However, further investigation is evitable as these data were obtained from only limited number of mice, and repeated procedures shall be cruicial in order to reveal more precise correlation between intensity of stimuli and NK cell activation.
T-bet--/- NK cells has higher RNA level of CD107a than WT NK cells
Following the demonstration that NK cells mediate their cytolytic functions upon the activation, I subsequently assessed to determine whether T-bet transcription factor influence the CD107a expression by looking at the RNA levels in both WT and T-bet. From the previous findings, it is now a known fact that T-bet-/- NK cells mediate less functional activities in both cytotoxicity and cytokine production. However, it is not revealed whether T-bet deficiency results in reduced amount of cytolytic vesicles or disturbance in the trafficking of the lysosomes. In order to address this remaining question, RNA extent of CD107a was measured in both C57/BL6 WT mice and T-bet KO transgenic mice by real-time PCR technique. Splenic NK cells were extracted from WT and T-bet KO transgenic mice and purified by NK isolation KIT. Subsequently, NK cells were stained for CD3 and NKP46 and sorted using FACs AriaII. Purity after sorting on NK cells from both sources were upto 98%. NK cells were kept in the trizol under -80oC. Next, RNA was separated from the cells using chloroform, isopropanol and ethanol and kept frozen under -80oC. The extracted RNA was hired to synthesise complementary DNA strand by reverse transcription PCR technique and readied to perform the real-time PCR. Gene expression assays were purchased from Applied Biosystem according to the genes of interest. These genes are as follow: CD107a, Rab27a, Unc13d, Vamp7, Wipf1, Dnm2. Additionally, gene expression assays for IFN-g and T-bet were also taken into account to examine the corresponding RNA levels. Using the Taqman PCR mastermix, ß-actin, and gene exression assay, the real-time PCR was performed on those eight genes of interest. Subsequently, Delta-delta CT method was hired to calculate the RT-PCR results. As the PCR was repeated several times, the ratio between delta CT values from WT and T-bet-/- NK cells was calculated and averaged. The representative data from one subject is shown in Fig.3.
As a control, T-bet was examined on both WT and T-bet-/-NK cells. As expected, T-bet was not found in the T-bet deficient cells whereas WT NK cells have much higher levels.
By looking at the results, the RNA level of CD107a was observed repeatedly to be higher in T-bet deficient NK cells than WT NK cells (Supplementary data 1). The averaged ratio indicated that CD107a RNA was 1.7 times higher in T-bet-/- cells. This result directly conflicts with the phenotype of T-bet-/- NK cells to have reduced ability to killing. Other possible explanation can be that the capability of cytolytic vesicle production may not be defected, but the trafficking of such vesicles could be diminished. To address the following question, RT-PCR was performed on the genes possibly involved in the lysosomal trafficking. First, Member RAS oncogene family 27a (Rab27a) was tested, which belongs to the small GTPase family, RAS family that are important regulators of vesicle trafficking and compartmentalisation. According to the RT-PCR results, T-bet deficient NK cells produce a reduced amount of Rab27a in terms of its RNA extent that the mean ratio of T-bet KO cells was 85% compared to the WT. Considering that Rab27a participate in the lysosomal trafficking, it is possible that reduced cytotoxic ability is due to the defect in the trafficking, not the amount of lysosomal vesicles. Subsequently, dynamin 2(Dnm2) was examined. This gene is the isoform of the conventional dynamin family of large GTPase that participate in the receptor mediated endocytosis and membrane remodelling. Additionally, it also localises with lytic vesicle and regulates the fusion to the plasma membrane. By looking at the RNA levels, it was elevated in T-bet deficient NK cells unlike Rab27a. Average, it was one half folds higher than WT NK cells', indicating that defect in cytotoxic ability is not owing to the final fusion of the vesicles to the plasma membranes. Next, Wipf1 belonged to Wiskott-Aldrich syndrome (WAS)/WASL interacting family was examined. Wipf1 is known to be involved in the initiation step of NK cell lytic-synapse formation. From the previous findings, the phenotype of defect in this gene shows NK cells with decreased cytolytic ability, nevertheless, overexpression in Wipf1 results in the enhanced ability. Taken into account that T-bet deficiency brings defected ability of cell killing, it was a surprise to see that Wipf1, like CD107a or Dnm2, was more produced in terms of its RNA level in T-bet-/- NK cells. The RT-PCR result revealed that it was almost two fold higher than WT (1:1:97 in WT : T-bet KO). This demonstrates that the formation of lytic synapses is not manipulated by T-bet transcription factor, at least not by Wipf1. Following the Wipf1, the next gene tested was Unc13d (also known as MUNC13-4). This gene belongs to the Unc13 family, which is known to involve in the vesicle maturation and regulation of cytolytic vesicle release. It interacts with Rab27a and primes the lytic vesicles for fusion with plasma membrane at lytic-synapse. Again, T-bet deficient NK cells had higher RNA extent of Unc13D (1:1.56 to T-bet KO). Lastly, Vamp7 was tested. In fact, Vamp7 was the most highly expected gene to test, because previous findings describe its essential role in target cell killing. To be brief, Vamp7 is known as Vesicle Associated Membrane Protein 7 that is found to be colocalise with CD107a molecules and participate in the fusion of transport vesicle to the target membrane. Although any trace of Vamp7 RNA was not found in both WT and T-bet-/- NK cells, it was not surprising fact, because Vamp7 is the soluble N-ethylmaleimide-sensitive-factor accessory-protein receptor (SNARE) among which present on target membranes (t-SNARE). As a matter of fact, this can explain how NK cells can be protected from their own lysosomal vesicles. In the absence of t-SNARE proteins, lytic synapses cannot be constructed, therefore for those NK cells missing Vamp7 is safe from its own vesicles.
Following the investigation on genes of which participate in the lysosomal trafficking, level of IFN-g was also quantitated. Although the RNA level does not directly reflect the IFN-g production, it is worth have a look to assess whether T-bet deficiency influences the RNA transcription. As a matter of fact, T-bet deficient NK cells seem to have a higher IFN-g RNA level than WT that it was approximately 37% elevated. Based upon the previous studies, IFN-g production is diminished in the absence of T-bet, yet its RNA level indicates T-bet is not directly acting on the RNA production but on the translation pathways.
Taken together, how T-bet deficiency affects on the lysosomal trafficking was tracked by looking at the RNA levels of genes supposed to be involved in the formation of lytic-synapses. Except Rab27a, other four genes were shown to be elevated in the absence of T-bet deficiency whereas Vamp7 was not found on both types of cells. This indicates that Rab27a is involved in unknown mechanism of lysosomal trafficking, which results in the functional defects in T-bet-/- NK cells.
The findings presented here have demonstrated the activation of NK cells in different environments and the effect of T-bet deficiency on the genes believed to be involved in lysosomal trafficking. Resting NK cells show minimal functional activities in order to prevent unnecessary cell killing, yet once stimuli is detected, it rapidly induces its lytic functions against the target cells. However, the question I tried to address was whether the different strength of stimuli would make any difference in the NK cell activation. Considering that NK cells mainly target stressed or virus-infected cells, they should be highly sensitive enough to detect any small stimulus, otherwise it may result in the massive infection or even worse, cancer.
This study has shown that the stimulation of resting NK cells boosts up the expression of CD107a, yet the intensity of expression may not be proportional to the strength of stimuli given. CD107a, which is the lysosomal associated protein representing approximately 50% of the proteins embedded on the lysosomal membrane, has been hired as a functional marker for the NK cell activity. Given that CD107a reflects the cytolytic activity of NK cells, it seems that activation of NK cells is irrelevant to the strength of stimuli. By using the different concentration of a-NK1.1 and a-NKG2D, those provide direct activation signals, this study demonstrated that the CD107a was upregulated on NK cells following the stimulation, indicating the production of lysosomes of NK cells. Moreover, the level of its expression was not noticeably varied when different concentration of stimulatory antibodies was used so that shows NK cells can mediate their cytolytic ability even with small stimulus detected-thereby provide rapid response and protection against any danger signal.
Dendritic cells have been shown to be involved in the activation of resting NK cells once they have been matured after the exposure to the antigen. DC produces various cytokines necessary for NK activation, such as IL-2, IL-12, IL-18, IFN-a/b, as a response, NK cells achieve their cytolytic functions as well as cytokine productions. Here, immature DC and LPS-stimulated DC were used at different ration to NK cells to observe change in the expression of CD107a. As expected, LPS-stimulated DC elevated the CD107a expression and this result explains the key role DC during the early events of immune response, that once DC encounter with a pathogen, they release various cytokines that ;IL-2 and IL-15 to induce activation and/or proliferation of NK cells, IL-12 for induction of IFN-g by NK cells, IL-18 to direct NK cells capable of migrating to secondary lymphoid organs where they can interact with DC. However, albeit elevation of CD107a in both ratio used, it was noticeably increased when the DC ratio to NK was higher. Once DC is matured, IL-1b is produced which, along with high mobility group B1 (HMGB1) factor, protect DC from NK cell mediated killing. Thus, DC can keep mediate NK cell activation without themselves being killed. Additionally, sufficient DC in the environment would provide more cytokines, thus resulting in the higher activation rate.
Immature DC, however, have shown the opposite results that the low ratio in NK:iDC mediated more expression of CD107a. Although the underlying mechanisms are elusive so far, iDC has been reported to be capable of inducing NK activation. In fact, NK cells interact with iDC to polarise and secrete IL-18 through synaptic cleft, thus initiate their activation. In turn, activated NK cells produce HMGB1 that induce DC maturation as well as protection against NK-mediated DC killing. Thus it is natural to see the elevated expression of CD107a in the co-culture with iDC. Then why the expression is low in the high NK:iDC ratio? Study by Piccoli et al has proposed the killing of DC by NK cells in infected tissues to amplify the stranger/danger signal derived from the invading pathogens. Although DC can secrete lysosomal hydrolases, this can be defeated by high NK/DC ratio. Thus, high NK/DC ratio rather kills DC rather than maturation and consequently, this results in the reduced number of DC in the NK-DC co-culture, followed by less activation of NK cells. Only those cells received HMGB1 survives and produce cytokines to activate remaining resting NK cells.
Despite the fact that these results all successfully demonstrated the activation of NK cells in various environments, it still leaves few laments for the further investigations. First, due to the limited resources and time, these experimentations were carried out once and thus there was no sufficient data to confirm the analysis in depth. Repetition on these investigations is inevitable to construct more firm and detailed analysis that are more feasible with other findings. Second, what I intended to assess through the project is the functional aspect of NK cells in relation to T-bet transcription factor. Thus, further investigation is required on looking at CD107a expression on both WT and T-bet -/- and this will advance the understandings on NK cells and its function.
It is now well-established that T-bet deficient NK cells have imperfect functions in every aspect, that they show decreased cytotoxicity and cytokine productions. Given that T-bet transcription factor is involved in the early development of NK cells, it is not so surprising fact to have such a consequence. However, the underlying mechanism on diminished cytolytic function is so far elusive and therefore it is still not clear whether T-bet deficient NK cells produce less lytic vesicles or it is due to the disturbance in the lysosomal trafficking. Based upon the preliminary results from Dr.A Martin-Fontecha,which demonstrate that the lysosomal content of T-bet deficient NK cell is reduced compared to the WT NK cells, real-time PCR was performed on those genes that may participate in the lysosomal trafficking to determine whether both results agrees upon each other. By targeting the five genes, CD107a, Rab27a, Unc13D, Vamp7, Dnm2 and Wipf1, I have shown that T-bet transcription factor contributes on all genes except Vamp7 that had no trace on both WT and T-bet deficient NK cells. CD107a encodes the highly glycosylated protein that contributes half of the proteins on the lysosomal membrane. Thus, it has been used as a NK degranulation marker. This study has shown that mRNA content of CD107a was consistently higher in T-bet deficient NK cells than the control. This result implies that there may be defective formation of cytolytic vesicles as the lysosomal content was found to be reduced. The vesicles may simply possess more CD107a protein or possibly their size may vary, yet any of this has not been determined.
CD107a closely interacts with Unc13D protein (also known as Munc 13-4), the member of Munc13 family of proteins, and it is probably responsible for vesicle priming. The targets of Unc13D are members of the SNARE family which can be defined into vSNARE and tSNARE; vSNARE are the SNARE proteins found on the vesicle and tSNARE are those present on target membrane. SNARE proteins interact in a coordinated manner to facilitate the fusion of two distinct membrane. Unc13D has been suggested to interact with SNARE protein and mediate the membrane fusion. Moreover, it has been found to be closely related to CD107a molecules that previous finding has demonstrated CD107a expression is significantly decreased when Unc13D has been mutated, as well as NK cytolytic ability. Often, their close interaction allows using the expression of CD107a as a screen tool of familial hemophagocytic lymphohistiocytosis (FHL), the defective genetic disorder on NK cells by mutated Unc13D. In this study, I have demonstrated that T-bet deficient NK cells possess higher content of Unc13D mRNA, which is consistent to the CD107a mRNA level that has also been found to be upregulated in T-bet deficient NK cells. These two positively regulated genes in the absence of T-bet transcription factor may suggest that lytic vesicles in the cell can possibly be primed and fuse to the target membrane more than the normal level. As a matter of fact, overexpression of Unc13D enhances the degranulation of lysosomes. Thus, at the baseline, this may explain how T-bet deficient NK cells can still induce cytotoxicity even if reduced content of lytic vesicle is synthesised.
In addition to CD107a and Unc13D, Rab27a is a member of Rab family of small GTPases that are important regulators of vesicle trafficking and compartmentalisation. The protein encoded by Rab27a gene undergoes post-translational prenylation to gain hydrophobicity, which facilitate the interaction with target membrane. Consequently, the interaction mediates the docking of lytic granules. In addition to that, Rab27a is a direct partner of Unc13D that both proteins are normally found in not only NK cells but CTL and mast cells where these two colocalise on the lytic vesicles. Unc13D protein binds to the Rab27a, forming a Rab27a/Unc13D complex that plays a essential role as a SNARE complex regulator. Thus, defects in Rab27a would prevent formation of the complex therefore lead to the functional inability to secrete the lysosomal content. I have shown that T-bet deficiency reduced the mRNA level of Rab27a, the only gene showing reduced level among the genes tested. This may suggest the reason why T-bet deficient NK cells has attenuated cytotoxicity, that reduced level of Rab27a leads to downregulated expression of the protein and therefore, Rab27a/Unc13D complex was not fully constructed as in the WT NK cells. Taken together, diminished cytolytic ability in the absence of T-bet transcription factor is not only owing to the decrease in lysosomal content, but the trafficking of lysosomes was also disrupted by less number of SNARE complex. Nevertheless, it should be noted that this defective phenotype can be reversed by culturing NK cells in the presence of IL-2. It may not induce increased production of Rab27a in both RNA or protein level, but would trigger alternative pathway for Rab27a function. Further investigation is needed to determine whether T-bet deficient NK cells recover their functional activity in the presence of IL-2, and this would clarify the role of Rab27a.
As like those genes showing elevated mRNA levels, Wipf1 was also higher in the absence of T-bet deficiency. Considering the function of protein encoded by Wipf1, it may not be the main cause of the reduced ability of NK cells. Wipf1 encoded protein, WIP is known to be important in the NK cell cytotoxicity as previous finding demonstrated WIP knockdown completely deters cytolytic function whereas overexpression enhances the killing. Its initial role is to bind to a region of WAS protein and stabilises filamentous actin and consequently, regulate the cytoskeleton rearrangement. This can be also found in the macrophage for the podosome formation, but in the NK cells, it is crucial for immune synapse dynamics. Moreover, it associates with lytic vesicles in the NK cells and participate in the lysosomal polarisation to the lytic-synapse. Altogether, deficiency in T-bet transcription factor elevates the Wipf1 mRNA level, but this may not be directly responsible for the reduced functional activity of T-bet deficient NK cells, since its elevation would increase the functional activity. At the baseline, it suggests that formation of lytic-synapse is not directly regulated by T-bet transcription factor.
Lastly, this study has shown the complete absence of Vamp7 in both WT and T-bet deficient NK cells. This was an unexpected result as many studies previously demonstrated that Vamp7 deficient NK cells show significant decrease in their cytolytic ability. Despite repeated RT-PCR was performed, Vamp7 was not found in any results. This may be due to the systematic errors during the performance, yet, it is still not easy to understand why Vamp7 did not produce any results or show extremely low level while other genes tested together worked fine. Possible reasons can be either problem with sample NK cells used or gene expression assay. Nevertheless, it is also hard to believe WT and T-bet deficient NK cells both do not produce any Vamp7 RNA. It is very likely it is the problem with Vamp7 gene expression assay, but to clarify, further investigation is required.
Taken together, this study have demonstrated the activation of NK cells in the various environments and RNA levels of genes believed to be involved in the lysosomal trafficking. NK cells, as a pivotal player in the immune response, it induces its cytolytic effects upon its stimulation by various types of stimuli. However, its activation may not totally depend on the intensity of the stimuli. There can be other possible underlying mechanisms in the activation of NK cells. Through this study, immature DC shows its ability to induce the activation of NK cells just like mature DC. However, as iDC express low level of MHC class I molecules, it easily becomes the target of NK cell mediated killing, and as a consequence, high NK:DC ratio may kill iDC rather than maturation. On the other side, mature DC can protect themselves from NK cells, thus stably can mediate activation of NK cells.
Lastly, T-bet deficiency was shown to affect RNA levels of many genes involved in the lysosomal trafficking. Especially, mRNA content of Rab27a was noticeably reduced while other genes of interest increased. This suggests that reduced amount of Rab27a protein leads to the less SNARE complex and therefore, disturbs the overall lysosomal trafficking.