|Home | About | Journals | Submit | Contact Us | Français|
Qin Tang - designed, performed, analyzed experiments and wrote paper
Bartek Grzywacz - designed, performed, analyzed experiments and wrote paper
Hongbo Wang - designed, performed, analyzed experiments
Nandini Kataria - designed, performed, analyzed experiments
Qing Cao - performed statistical analyses
John E. Wagner - provided essential reagents, analyzed experiments and contributed to writing of paper
Bruce R. Blazar - designed experiments and contributed to writing of paper
Jeffrey S. Miller- designed experiments and contributed to writing of paper
Michael R. Verneris - designed and analyzed experiments and wrote paper
The natural cytotoxicity receptors (NCRs), NKp30, NKp44 and NKp46, are thought to be NK lineage restricted. Here we show that IL-15 induces NCR expression on umbilical cord blood (UCB) T cells. NCRs were mainly on CD8+ and CD56+ UCB T cells. Only NKp30 was functional as demonstrated by degranulation, IFN-γ release, redirected killing and apoptosis. Since NCRs require adapter proteins for function, the expression of these adapters were determined. The adapters used by NKp30 and NKp46, FcεR1γ and CD3ζ, were detected in UCB T cells. There was a near absence of DAP12, the adaptor for NKp44, consistent with a hypofunctional state. NKp46 was on significantly fewer UCB T cells, possibly accounting for its lack of function. Adult PB T cells showed minimal NCR acquisition after culture with IL-15. Since UCB contains a high frequency of naïve T cells, purified naïve T cells from adult peripheral blood (PB) were tested. Although NKp30 was expressed on a small fraction of naïve PB T cells, it was non-functional. In contrast to UCB, PB T cells lacked FcεR1γ expression. These results demonstrate differences between UCB and PB T cells regarding NCR expression and function. Such findings challenge the concept that NCRs are NK cell specific.
Umbilical cord blood (UCB) is emerging as a preferred stem cell source for allogeneic transplantation because: 1) it is rich in hematopoietic progenitor cells, 2) it is easily collected and cryopreserved at the time of delivery, with no apparent risks to the donor (infant), 3) it undergoes infectious disease screening and HLA typing at the time of collection, allowing for rapid donor identification and transplantation, 4) it is associated with low rates of both acute and chronic graft vs. host disease (GVHD), despite HLA mismatch and 5) it shows similar rates of leukemia relapse relative to other hematopoietic cell sources, such as bone marrow or peripheral blood (PB) (reviewed in(1)). Thus, it is important to understand both the similarities and differences between various effector cell populations in UCB relative to PB.
Although UCB and PB do not differ with regard to the percentages of T cells, T cell subsets (CD4 and CD8) or the proportions of αβ and γδ T cells (2, 3), functional differences are commonly observed. For instance, the majority of UCB T cells are naïve, while most adult PB T cells are antigen experienced (4). Compared to PB T cells, UCB T cells express less of the transcription factor NFAT2c, which plays an important role in cytokine gene expression following immune activation (5). Accordingly, CD3 stimulated UCB T cells differ from PB, producing less Th1 (IL-2, TNF-α, IFN-γ) and Th2 cytokines (IL-4, IL-10, IL13) (6-10). Likewise, proteins associated with both activation (CD40L, CD25) and cytotoxicity (perforin and FasL) are reduced in UCB T cells relative to PB T cells (7, 11, 12). Although such differences may suggest immaturity, UCB T cells are capable of functional responses. Antigen specific T cells can be found following in utero infections (13) or in the cord blood of HA-1neg infants born to HA-1pos mothers.(14)
A subset of PB T cells can express receptors that are mainly found on NK cells. Included are NK cell inhibitory receptors, such as killer immune globulin-like receptors (KIRs) (15) and CD94/NKG2A (16-18). In addition to these, PB T cells can also express NK cell activating receptors, such as activating KIR (19, 20), CD94/NKG2C (17, 21, 22) and NKG2D (23). In some studies, engagement of these NK cell associated receptors can either negatively or positively modulate TCR triggering (cytotoxicity and cytokine secretion) (24-26). We and others have also shown that triggering of receptors, such as NKG2D, can induce TCR-independent cytotoxicity on IL-2 or IL-15 stimulated T cells (27, 28). Like PB T cells, UCB T cells can also express KIR (CD158a, -b, and -e1) and NKG2A/CD94, but at significantly lower frequencies (29, 30). These results are in line with the supposition that NK associated receptors are mainly expressed on effector T cells from adult PB (28, 31-33). UCB derived T cells that express NK receptors can be expanded after culture with IL-15 (30). Importantly, the chemotherapy commonly used prior to allogeneic hematopoietic cell transplantation results in lymphodeletion and an increase system IL-15 levels. This has been linked to the success of clinical trials using adoptively transferred PB T and NK effector cell populations (34, 35).
Recently, three NK activating receptors (NKp30, NKp44 and NKp46) have been identified. Collectively they have been referred to as the “natural cytotoxicity receptors (NCRs)” because they play a significant role in the killing of malignant targets (reviewed in (36)). The initial reports describing NCRs showed that unlike other NK cell associated receptors, NCRs were restricted to NK cells and not expressed on PB T cells (37-41). Here, we show that a small percentage of freshly isolated UCB T cells co-express NKp30. We further demonstrate that, unlike PB T cells, UCB T cells can acquire NKp30, -44 and -46 following culture with IL-2 or IL-15; however, only NKp30 is functional. Lastly, small amounts of naïve adult PB T cells can acquire NKp30 following IL-15 stimulation, but this receptor is not functional on these cells.
UCB and PB were obtained from healthy donors. Mononuclear cells (MNC) were prepared by density gradient centrifugation using lymphocyte separation medium (Mediatech, Inc). T cells were isolated using CD3 microbeads (Miltenyi Biotec, Auburn, CA). Purity was assessed by flow cytometry and was >97% (not shown). Isolated T cells were cultured at 1 × 106 cells/ml in Ham's F12 + DMEM (1:2 ratio) with 10% male AB-human serum (Seracare Life Sciences, Oceanside, CA), ethanolamine (50 μmol/L), ascorbic acid (20 mg/L), 5μg/L sodium selenite (Na2SeO3), β-mercaptoethanol (24μmol/L), and penicillin (100U/ml)-streptomycin (100U/ml). Recombinant cytokines IL-2 (Chiron, Emeryville, CA), IL-15 (PeproTech Inc., Rocky Hill, NJ), IL-4 (R&D systems, Minneapolis, MN) and IL-7 (gift from National Institute of Health) were added as indicated.
Prior to staining, cells were washed in PBS with 2% fetal bovine serum and 0.2% NaN3 and then stained with monoclonal antibodies (mAb) for 30 minutes at 4°C. Samples were analyzed on a FACScalibur using CellQuest software. Flowjo 5.7.0 software was used for data analysis. For intracellular staining, the cells were fixed and permeabilized with BD Cytofix/Cytoperm™ solution and washed with BD Perm/Wash™ buffer and stained according to manufactures specifications. The following mouse anti-human antibodies were used: CD3-FITC, CD4-FITC, CD8-FITC, CD62L-PE, CD4-APC, CD8-APC, CD56-APC (BD Biosciences, San Jose CA), NKp30-PE, NKp44-PE, NKp46-PE and TCRζ-PE (Beckman Coulter, Inc., Miami, FL), CCR7-APC (R&D systems, Minneapolis, MN), CD45RA-FITC (Ebioscience, San Diego, CA).
Freshly isolated UCB T cells were washed and resuspended in prewarmed PBS at 5×106/ml containing 5μM carboxyfluorescein diacetate, succinimidyl ester (CFSE) (Molecular Probes Inc., Eugene, OR) and incubated for 15 minutes at 37°C. Following this, cells were washed again with prewarmed PBS, then resuspended in prewarmed medium and incubated for another 30 minutes and again washed prior to cell culture.
24-well plates were coated with mAbs against human NKp30 (clone 210847), NKp44 (clone 253415), NKp46 (clone 195314) (R&D systems, Minneapolis, MN), CD3 (OKT3) or isotype control (Sigma, St Louis, MO) all at 5μg/ml in PBS for 4 hours at 37°C. Plates were washed and UCB or PB T cells at day 14 of culture were added at 0.3×106 in 300 μl medium. Anti-CD107a-FITC (BD Biosciences, San Jose, CA) was added to the culture prior to incubation. After 8 hours of incubation at 37°C 5% CO2, cells were harvested, and analyzed for CD107a expression.
96-well plates were coated with antibodies against human NKp30, NKp44, NKp46, CD3 (OKT3) mAbs or isotype control antibody at 5μg/ml in PBS for 4 hours at 37 °C. Plates were washed and cells were added (0.2×106 cells in 200 μl medium). After 16 hours of incubation at 37°C 5% CO2, cells were harvested and washed with cold PBS, resuspended in 1X Binding Buffer (10 mM Hepes/NaOH, pH 7.4, 140 mM NaCl, 2.5 mM CaCl2) and stained with Annexin V-FITC (BD Biosciences, San Jose CA) for 15 minutes at room temperature and and PI (Sigma, St Louis, MO) was added. Cells were analyzed by flow cytometry.
0.2 ×106 cells in 200μl medium were stimulated with plate-bound Ab coated 96-well plates for 24 hours. Supernatants were harvested and assayed for IFN-γ production by ELISA (R&D systems, Minneapolis, MN) according to the protocol by the manufacturer.
The FcγR-positive cell line, P815 (murine mastocytoma) was used as a target for redirected killing assays. P815 cells were labeled with 51Cr (DuPont-NEN, Boston, MA) by incubating 1 × 106 cells in 300 μCi (11.1 MBq) 51Cr for 1 hours at 37°C 5% CO2. Cells were then washed 3 times with PBS, resuspended in culture medium, and added in triplicate to 96-well plates at 104 cells/well. Effector cells were added at the specified ratios and incubated at 37°C 5% CO2. Effector cells were preincubated in PBS in the presence of soluble antibodies (5 μg/1×106 cells in 100μL PBS) for 30 minutes and washed once with PBS prior to coincubation with tumor cell targets. After 4 hours, 100μl of supernatant was counted using a γ counter. The percentage of specific lysis was calculated using the following equation: % specific 51Cr release = 100 × [(test release) - (spontaneous release)]/[(maximal release) - (spontaneous release)].
Total RNA was isolated using the RNeasy Mini Kit (Qiagen, Santa Clarita, CA) and reverse transcribed to cDNA (iScript, Bio-Rad, Hercules, CA). 2ng of cDNA was amplified using recombinant Taq Polymerase (Gibco BRL) for 30 cycles. The following primers were used: DAP12 forward CCGCA AAGAC CTGTA CGCCA, reverse TGGAC TTGGG AGAGG ACTGG; FcεRIγ. forward ATGAT TCCAG CAGTG GTCTT G, reverse GTGCT CAGGC CCGTG TAAA; CD3ζ forward GCACAG TTGCC GATTA CAGA, reverse GGTTC TTCCT TCTCG GCTTT. PCR products of DAP12 (650bp), CD3ζ (273bp) and FcεRIγ (209bp) were resolved on 2% agarose gel and bands were visualized with ethidium bromide.
UCB T cells at day 14 of culture were lysed using freshly prepared lysis buffer (10mMTris-HCl, pH 8, 150 mM NaCl, 1 mM EDTA, 1% NP-40, 0.5% deoxycholate, 0.1% SDS [sodium dodecyl sulfate], protease inhibitor cocktail (complete protease inhibitor; Roche, Indianapolis, IN), PMSF, and 1mM Na3VO4. Samples were separated on a 15% polyacrylamide gel and transferred to a PVDF membrane. Membranes were blocked with 5% nonfat dry milk and probed with anti-actin (SantaCruz Biotechnology, Santa Cruz, CA), anti-FcεRIγ (Upstate USA Inc., Charlottesville, VA) or anti-DAP12 (generous gift from Dr. Paul Leibson, Rochester, MN) followed by a species-specific secondary horseradish peroxidase (HRP)-conjugated antibody (Santa Cruz Biotechnology, Santa Cruz, CA). Blots were developed using chemiluminescence (Pierce, Rockland, IL).
Statistical analyses were performed with Statistical Analysis System statistical software version 9.1 (SAS Institute, Inc., Cary, NC). For non-normally distributed data the Mann-Whitney rank sum test was used in the evaluation of the statistical differences between UCB and PB T cells, CD4+ vs. CD8+ and CD56- vs. CD56+ subpopulations. One-way ANOVA was used when the data was approximately normally distributed with the general linear models procedure (PROC GLM). Adjustments for multiple comparisons were done with the Tukey's method. Groups with values of p ≤ 0.05 were considered to be statistically different.
A small fraction of freshly isolated UCB T cells expressed NKp30 (1.7±0.9%), but not NKp44 (0.4±0.4%) or NKp46 (0.5±0.4%) (figure 1A and C). Such results suggested that UCB T cells may acquire NCRs after activation with cytokines such as IL-15. Purified CD3+ cells form either adult PB or UCB were cultured in IL-15 and analyzed for NCR expression by flow cytometry after 0, 7, 14 or 21 days. As shown in figure 1B, NKp30, NKp44, and NKp46 were markedly increased on UCB T cells with only marginal increases noted on PB T cells (figure 1B). These results were consistent over a series of donors, resulting in a significant increase in NKp30 (36.8±9.6%, p=0.01), NKp44+ (40.9±19.5, p=0.01), NKp46 (12.9±6.9%, p=0.01) on UCB T cells compared to PB at day 14 (figure 1C). Interestingly, at day 14 the percentage of NKp30- and NKp44-expressing UCB T cells was relatively similar; however, in comparison the percentage of NKp46-expressing T cells was significantly lower (figure 1C). Thus, freshly isolated UCB T cells show rare expression of NKp30, but after culture in IL-15 expression of NKp30, -44 and -46 is observed. CD3 positive selection may result in signals that induce NCR expression, however T cells in from unmanipulated UCB lymphocyte cultures (i.e., not positively selected) also showed NCR upregulation following IL-2 or IL-15 exposure (not shown). Collectively, such results show significant differences between UCB and PB T cells in NCR acquisition following culture with IL-15.
As shown above, small numbers of freshly isolated UCB T cells express NKp30; and at day 14 of culture, NKp30 was present on a significant proportion of UCB T cells (figure 1B and C). Such findings suggest NKp30 acquisition after culture, or alternatively the T cells expressing NKp30 could be preferentially expanded in culture. To address this, freshly-isolated UCB T cells were stained with CD3 and NKp30 and the CD3+NKp30- and CD3+NKp30+ fractions were purified by FACS sorting (figure 2A, left). After 14 days of culture with IL-15, cells from both fractions showed similar expansion and both subsets expressed NKp30, indicating that NKp30 is acquired after culture rather than preferential expansion of NKp30+ cells (figure 2A, right). To further investigate whether NCR acquisition was related to cell proliferation, freshly isolated UCB T cells were labeled with the membrane dye CFSE and cultured for 14 days. As shown in figure 2B, expression of NCRs is abundant on cells that underwent multiple rounds of divisions, whereas these receptors are absent from non-proliferating cells. Comparing NKp30 with NKp44 and NKp46, there were subtle, but reproducible changes in NCR acquisition as it related to CFSE dilution. UCB T cells acquired NKp30 after as few as 1 cell division. In contrast, NKp44 or NKp46 were only acquired on those cells that showed significant proliferation and loss of CSFE.
IL-2, IL-4, IL-7 and IL-15 all signal through the common γ chain and act upon T cells. We therefore tested whether these cytokines could induce NKp30, NKp44 and NKp46 expression on UCB T cells. After 14 days of culture with increasing doses of IL-2 and IL-15, there was a dose-dependent induction in the surface expression of all NCRs tested (figure 3A and 3D). In contrast, no NCR expression was observed when cells were cultured in IL-4, despite robust proliferation (not shown) (figure 3B). Interestingly, IL-7 led to a partial induction of NKp30, but not NKp44 or -46 and, unlike IL-2 and IL-15, there was no clear dose dependent induction at the doses tested (figure 3B). Combining IL-4 and IL-15 resulted in a near complete inhibition in NCR expression (not shown). Removing the IL-4 after 7 days of culture and maintaining IL-15 allowed partial acquisition of NCRs (not shown).
As shown in figure 4A, at D+14 after culture with IL-15, both CD4 and CD8 T cell populations could be observed. A considerable proportion of the CD8+ T cells showed NCR expression. About half of CD8+ T cells expressed NKp30 or NKp44 (figure 4A & B). Similar to our observations in bulk cultures, the percentage of NKp46 expressing CD8+ T cells were considerably lower than that of NKp30 and NKp44. In contrast, a significantly lower fraction of CD4+ T cells showed NKp30 expression (p=0.01), whereas NKp44 or NKp46 were absent (figure 4A & B).
Both freshly-isolated and ex vivo-activated T-cells can express CD56 (27, 42). The vast majority of freshly isolated UCB T cells were did not express CD56, but after culture in IL-15, such T cells acquired CD56. Gating on the CD3+CD56- and CD3+CD56+ cells after 14 days of culture showed that NCRs were found at higher density on CD56 expressing T cells relative to CD56- T cells (figure 4C). Comparing the NCR expression of CD3+CD56+ cells to CD3+CD56- across a series of donors (n=4) we found a significantly higher frequency of NKp44 and NKp46 (p=0.05, 0.02 respectively) on CD3+CD56+ cells, but a non-significant trend was observed for NKp30 expression between CD3+CD56- vs. CD3+CD56+ subpopulations (p=0.07) (figure 4D).
Only a minority of CD3+ T cells in culture expressed TCR γδ at D+14 of culture (average=8.4% (range=1.4-22%), n=6). As shown in figure 4E, gating on CD3+γδ+ T cells showed that NKp30, NKp44 and NKp46 were expressed. Comparing the NCR expression on CD3+γδ+ T cells to CD3+γδ- T cells showed no differences between the two cell types for NKp30 (p=0.42), NKp44 (p=0.14) or NKp46 (p=0.25) (figure 4F).
The ligands recognized by NCRs have not been definitively determined. Thus, to evaluate whether NCRs were functional on UCB T cells, assays were performed using agonist mAb. Controls included OKT3 (positive) and murine IgG (negative). To assay for cytotoxic granule release following receptor engagement we used a monoclonal antibody directed against lysosomal-associated membrane protein-1 (LAMP-1, CD107a). CD107a is normally found on the internal membrane of cytotoxic vesicles and is translocated to the cell surface upon degranulation (43). As expected, CD3 ligation (OKT-3) resulted in a significant increase in the percentage of cells undergoing degranulation (p=0.01) (figure 5A). Likewise, ligation of NKp30 on UCB T cells resulted in significant degranulation (p=0.01), while ligation of NKp44 and NKp46 did not (p=1, both) (figure 5A). In further studies IFN-γ production was investigated following receptor engagement (figure 5B). Cells were cultured in the presence of plate-bound antibodies for 16 h, and supernatant was collected for ELISA. Similar to degranulation, cross-linking of either NKp30 or OKT3 resulted in IFN-γ production (p=0.15, 0.03 respectively). In contrast, little or no IFN-γ was detected in supernatants after stimulation with agonist NKp44 and NKp46 antibodies (p=1 for both, figure 5B).
Activation induced cell death (AICD) occurs following prolonged TCR ligation (44), and similar events occur in NK cells after NCR ligation (45). Thus, we tested whether NCR-expressing activated T-cells at D+14 of culture with IL-15 undergo AICD following NCR ligation. As shown in figure 5C, the number of apoptotic cells did not change in the presence of agonist NKp44 or NKp46 mAbs. In contrast, prolonged NKp30 cross-linking, induced Annexin V staining and loss of membrane integrity (PI+) by a significant proportion of cells, consistent with AICD. Similar results were obtained in control (OKT3) treated cells.
Next, we investigated whether NCR ligation on UCB T cells could induce cytotoxicity using a reverse ADCC assay. As shown in figure 5D after 14 days of culture, T cells show potent CD3 redirected lysis (OKT3, positive control). Cross-linking of NKp30 on UCB T cells led to redirected cytolysis. In contrast, a non-specific IgG antibody or agonist antibodies against NKp44 or NKp46 showed no redirected cytolysis. These antibodies triggered redirected cytolysis using NK cells, proving that they can induce reverse ADCC (not shown). We next investigated whether signaling through NKp44, and NKp46 might provide costimulation and increase cytotoxicity relative to NKp30 alone, however no enhancement in killing was observed (figure 5E). Previous studies show other NK cell activating receptors, such as NKG2D can co-stimulate TCR signaling in CD8+ T cells.(24) Thus, we tested whether signaling through both CD3 and NKp30 would augment redirected lysis. Compared to CD3 signaling alone, the addition of NKp30 did not further enhance cytotoxicity (figure 5F). We and others have shown that both freshly isolated and cultured T cells express NKG2D and that signaling through this receptor can trigger TCR independent cytotoxicity.(27, 46, 47) We tested whether NKG2D could cooperate with NKp30 in this setting. Signaling through the combination of NKp30 and NKG2D resulted in higher cytotoxicity relative to either NKp30 or NKG2D alone (figure 5G). Collectively, these results demonstrate that signals through NKp30, but not NKp44 or NKp46, lead to UCB T cell activation that result in granule exocytosis, cytokine secretion, cytotoxicity and apoptosis.
NCRs are unable to directly transmit intracellular signals and rely upon adaptor proteins for signal transduction. NKp30 and NKp46 signal through CD3ζ and FcεRIγ (48) (49), while NKp44 uses DAP12 (50). To investigate whether these adaptors are expressed by UCB T-cells after culture, we performed RT-PCR. Using small quantities of starting mRNA (2 ng), message for all 3 adaptor proteins were readily amplified (figure 6A). To evaluate protein expression, flow cytometry and Western blotting were used. CD3ζ was found in the vast majority of T cells, as well as NK cells (figure 6B). In contrast, DAP12 was abundantly expressed in NK cells, with a near absence of the protein in UCB T-cell lysate (figure 6C). However, overexposure of Western blots did demonstrate small quantities of DAP12 in UCB T cells (not shown). Similarly, FcεRIγ could be detected in UCB T cells; but the quantities were significantly less than in NK cells (figure 6D). Collectively, these data show that the adapter proteins are controlled at the level of translation.
Considering that UCB is composed mainly of naïve T cells, while PB contains a mixture of naïve and memory cells, we hypothesized that the relative differences in naïve T cells may account for the disparity in NCR expression between the two cell sources. T cell subsets from the PB were sorted into naïve (CD3+CD62LhighCCR7+CD45RA+), central memory (CM, CD3+CD62LhighCCR7+CD45RA-), effector memory CD45RA+ (EM RA+, CD3+CD62LlowCCR7-CD45RA+), and effector memory CD45RA- (EM RA-, CD3+CD62LlowCCR7-CD45RA+ ) (reviewed in (51)) and cultured in IL-15 for 14 days. A small fraction of naïve, but not memory PB T cells acquired NKp30 after culture (figure 7A). Acquisition of NKp30, -44 and -46 by purified naïve and memory T cell subsets after culture with IL-15 is shown in Figure 7B. Redirected killing assays with cultured naïve cells from 3 donors (with one donor having >30% NKp30+ cells) showed no cytotoxicity after NKp30 engagement. As a positive control, OKT3 engagement induced killing (figure 7C). To investigate whether differences in the adapter proteins used by NKp30 accounted for these differences, we investigated CD3ζ and FcεRIγ expression in purified PB fractions after culture. CD3ζ was present in all subsets of PB T cells (figure 7D). However, FcεRIγ was missing in all subsets of cultured PB T cells, including naïve cells, whereas it was abundant in NK cells and detectable in UCB T cells (figure 7E).
In this study, we demonstrate for the first time that a small number of freshly isolated UCB T cells express NKp30. Following culture with either IL-2 or IL-15, UCB T cells acquire NKp30, NKp44 and NKp46. IL-7 induced NKp30, but not the other NCRs. In contrast, the NCRs were not observed when UCB T cells were cultured with IL-4. These cytotoxic triggering receptors were mainly expressed by expanded CD8+ T cells that co-expressed CD56. CD3+CD56+ T cells account for a minor fraction of freshly isolated UCB T cells, however after culture with IL-2 and/or IL-15 T cells can acquire CD56 after activation and proliferation (30, 52). Similarly, NKp30, -44 and -46 were expressed mainly on UCB T cells after proliferation. Consistent with the results of others (48, 53, 54), we observed essentially no expression of NKp30, NKp44 or NKp46 on freshly isolated adult PB T cells. In most experiments with adult PB we could, however, detect a small fraction of NCR+ T cells after 14 days of culture with IL-15. The percentage of such cells varied depending upon the individual donor and receptor tested (see figures 1B and and7B).7B). Given that UCB is mainly composed of naïve T cells (4), we reasoned that naïve PB T cells might also have the capacity to acquire NCRs following cytokine stimulation and that these naïve PB T cells might be overgrown by the expansion of effector T cells. Indeed, purified naïve PB T cells contained a small, but reproducible CD3+NKp30+ fraction after 14 days of culture with IL-15. In some donors we could also detect NKp44 expressing cells in the naïve fraction. In contrast to naïve cells, purified EMRA+, EMRA- and CM PB T cell populations showed essentially no NCR acquisition after culture. Such results are remarkable considering that other NK cell-associated receptors (KIR, NKG2A/C, and NKG2D) are preferentially expressed by antigen-experienced T cells and that such cells increase with age (28, 31-33). In contrast, NCRs appear to be acquired by UCB T cells and naïve PB T cells.
Using cytokine secretion, degranulation assays, redirected killing and AICD, we show that NKp30 on UCB T cells is functional. Interestingly, CD3 triggering induced much higher cytotoxicity in redirected killing compared to NKp30. Such results are not surprising given that all the cells express CD3, while only a fraction (typically 30-50%) expressed NKp30. Interestingly, the difference between these two stimuli (CD3 vs. NKp30) was not apparent following plate-bound mAb induced degranulation. This may be due to the magnitude of response induced by CD3 stimulation in the context of a cellular target (i.e., redirected killing) compared to isolated receptor triggering by platebound mAb. In fact, our results with CD107a show that it is a rare subset of T cells (~10%) that degranulate in response to single receptor triggering (either CD3 or NKp30). Upon further investigation we found that most of the cells that degranulate following CD3 ligation are in fact, NKp30+ (not shown). Thus, NKp30 may mark activated T cells, poised to degranulate.
The proximal adapter protein for NKp44, DAP12, is only minimally expressed by IL-15 expanded UCB T cells, likely explaining the lack of NKp44 signaling. The lack of NKp46 function is more difficult to understand. The percentage of NKp46 expressing UCB T cells was consistently lower than those expressing NKp30 and NKp44, however, we would have expected to see some evidence of NKp46 function in degranulation, IFN-γ or redirected killing assays, but this was not observed. The proximal adapter proteins used by NKp46 (CD3ζ and FcεRIγ) are expressed by UCB T cells at levels apparently sufficient to permit signaling via NKp30. Since UCB T cells consistently showed lower levels of FcεRIγ compared to NK cells it can be hypothesized that FcεRIγ is relatively more important for NKp46 signaling as compared to NKp30. Alternatively, NKp46 could require distinct, and yet undefined, down-stream signaling components that are present in NK cells, but not in UCB T cells. Lastly, naïve T cells from adult PB can also acquire NCRs, but at a lower proportion than their UCB counterparts. However, these receptors (on naïve PB T cells) were not functional, possibly owing to the absence of FcεRIγ in these cells.
At present, we do not understand the mechanism for the observed differences between UCB and PB T cells with respect to NCR acquisition and function. One explanation may be that the hormonal and/or cytokine milieu found in placental tissues favors NCR acquisition. Whether certain placental derived factors “prime” UCB T cells to express NCRs is conceivable. However, we were unable to induce NCR expression following culture of PB T cells in cord blood sera and IL-15 (not shown). Likewise, it is possible that the hormonal changes that occur at the time of child birth or the stress associated with delivery may in some way influence the ability of UCB T cells to acquire NCRs. Because UCB units were donated in a de-identified manner, we were not able to determine whether the mode of child birth (caesarean section vs. vaginal delivery) played any role in either de novo NKp30 expression (i.e., day 0) or NCR expression after cytokine stimulation. Since NCRs may be involved in antiviral immunity (55-57), another possible explanation may be that the ability to acquire NCRs by UCB T cells is a form of evolutionary protection against infectious organisms. Such findings would support the concept that newborns rely more upon their innate immune system until the adaptive immune system fully develops Whether infants or young children also express NKp30 or acquire NCRs following IL-15 stimulation is not known. Another possible explanation for the differences between NCR expression from PB and UCB is that the cells present in UCB are not the true counterparts of those found in PB. In fact, T cells other than those commonly found in the PB have recently been show to express NCRs. For instance, the Jabri group (58) has demonstrated that intestinal epithelial lymphocytes (T cells) express functional NKp44 and NKp46 in patients suffering from celiac disease. Interestingly, while PB T cells only express CD3ζ, thymic independent T cells isolated from the murine gastrointestinal tract express adapters both for NKp30 and NKp46 (CD3ζ and FcεRIγ)(59), perhaps supporting this view. In summary, the ability to acquire NCRs appears to be dependent upon the maturational status of the T cells. As stated above, such findings are in stark contrast to the expression of other NK cell-associated receptors on T cells. Collectively, these results show that there are fundamental differences between PB and UCB T cells with respect to NCR acquisition and function. Such findings challenge the dogma that NCRs are uniquely expressed by NK cells.
This work was supported by Children's Cancer Research Fund (MRV), P01 CA65493 (JSM), R01 HL55417 (JSM), R01 AI34495 (BRB) and R01 CA72669 (BRB). The authors have no competing financial interests to disclose.