Activation of NK cell cytotoxicity by thenatural compound 2,3-butanediol

Hsin-Chih Lai,*,† Chih-Jung Chang,‡ Chun-Hung Yang,*,§ Ya-Jing Hsu,* Chang-Chieh Chen,*Chuan-Sheng Lin,* Yu-Huan Tsai,* Tsung-Teng Huang,*,† David M. Ojcius,†,�

Ying-Huang Tsai,¶ and Chia-Chen Lu#,1

*Department of Medical Biotechnology and Laboratory Science, †Center for Pathogenic Bacteria and Center for Molecular andClinical Immunology, ‡Department of Microbiology and Immunology, College of Medicine, Chang Gung University, Taiwan,Republic of China; §Division of Applied Toxicology, Taiwan Agricultural Chemicals and Toxic Substances Research Institute,

Council of Agriculture, Taiwan, Republic of China; �Health Sciences Research Institute and School of Natural Sciences,University of California Merced, Merced, California, USA; ¶Department of Pulmonary and Critical Care Chang Gung Memorial

Hospital, Chia Yi Branch, Taiwan, Republic of China; and #Department of Respiratory Therapy, College of Medicine, Fu JenCatholic University, Taiwan, Republic of China

RECEIVED JANUARY 19, 2012; REVISED JUNE 7, 2012; ACCEPTED JUNE 29, 2012. DOI: 10.1189/jlb.0112024

ABSTRACTThe natural compound 2,3-BTD has diverse physiologi-cal effects in a range of organisms, including acting asa detoxifying product of liver alcohol metabolism in hu-mans and ameliorating endotoxin-induced acute lunginjury in rats. In this study, we reveal that 2,3-BTD en-hances NK cell cytotoxic activity in human pNK cellsand NK92 cells. Treatment of NK cells with 2,3-BTD in-creased perforin expression in a dose-dependent man-ner. This was accompanied by elevated JNK andERK1/2 MAPK activities and enhanced expression ofNKG2D/NCRs, upstream signaling molecules of theMAPK pathways. The 2,3-BTD effect was inhibited bypretreatment with inhibitors of JNK (SP) or ERK1/2 (PD)or by depleting NKG2D/NCRs or JNK1 or ERK2 withsiRNA. These results indicate that 2,3-BTD activatesNK cell cytotoxicity by NKG2D/NCR pathways and rep-resent the first report of the 2,3-BTD effect on activa-tion of innate immunity cells. J. Leukoc. Biol. 92:000–000; 2012.

IntroductionThe low molecular-weight compound 2,3-BTD, which is widelysynthesized in humans [1], yeast [2], and bacteria [3], is involvedin a variety of biological activities. These include homeostasis ofenvironmental pH when bacteria grow to high cell density [4],stimulation of bacterial biofilm formation [5], enhancement ofplant growth and systemic resistance against bacteria [6], toler-ance to drought in Arabidopsis thaliana [7], activity as a beetlepheromone [8], and potent CNS-depressant effect in rats [9].

This remarkable functional repertoire suggests that 2,3-BTD mayact as a signaling molecule in a wide variety of species [6, 8]. Re-cently, we reported that 2,3-BTD ameliorates endotoxin-inducedacute lung injury in rats [10]. We also characterized the effect ofRSV on NK cell NKG2D/NCR signaling and cytotoxic activity[11]. However, whether 2,3-BTD plays a role in modulating im-mune activity via regulation of NK cell cytotoxicity activity re-mains to be characterized.

The NK cells are important for early host defense againstinfection and tumors [12–14]. NK cells are with the capabilityof granule exocytosis by releasing granule proteins, such asperforin, granzymes, and granulysin [15]. The NK cell cyto-toxic activity is controlled by coordinated signals generatedfrom the ligation of inhibitory and activating receptors [16].The major activating receptors are the NCRs, comprising con-stitutively expressed NKp46 [17] and NKp30 [18], and the in-duced NKp44 [19]. Moreover, the NKG2D receptor, which isalso identified in human T cells [20], also involves activatingNK cell cytotoxicity [21]. The ligation of the activating recep-tors lead to activation of a cascade of intracellular signaling,resulting in polarization and exocytosis of granules to lyse theTS [16, 22]. In NK cells, ERK, JNK, and p38 are intermediatesof the important signaling MAPK pathways that regulate gran-ule polarization, which is mediated by reorientation of the mi-crotubule organizing center to the synapse [23, 24].

In the present study, we show that the cytotoxic activity of thepNK cells and NK92 cells with 2,3-BTD pretreatment is enhancedsignificantly compared with that of the untreated groups. Subse-quently, we investigate the potential signaling pathways involvedin 2,3-BTD-stimulated NK cell cytotoxicity. Our results indicatethat the effect is mediated through activation of NCR/NKG2Dpathways, which, in turn, leads to increased cytotoxic activity.

1. Correspondence: Dept. of Respiratory Therapy, College of Medicine, FuJen Catholic University, No. 510, Zhongzheng Rd., Xinzhung Dist., NewTaipei City 24205, Taiwan, Republic of China. E-mail: 082385@mail.fju.edu.tw

Abbreviations: 2,3-BTD�2,3-butanediol, ATCC�American Type CultureCollection, ES�effector cell, MICA/B�MHC class I chain-related gene A/B,NCR�natural cytotoxicity receptor, NKG2D�NK group 2, member D,PD�PD98059, pNK�primary NK, qPCR�quantitative PCR, RNAi�RNA in-terference, RSV�resveratrol, SB�SB203580, siRNA�small interfering RNA,SP�SP600125, TM�maximum release, TS�target cell

Article

0741-5400/12/0092-0001 © Society for Leukocyte Biology Volume 92, October 2012 Journal of Leukocyte Biology 1

Epub ahead of print July 16, 2012 - doi:10.1189/jlb.0112024

Copyright 2012 by The Society for Leukocyte Biology.

MATERIALS AND METHODS

Cell isolation and cultureThe human NK cell line NK92 was derived from a non-Hodgkin’s lym-phoma (ATCC CRL-2407) [25]. NK92 cells were maintained in culturewith 100 IU/ml IL-2 (PeproTech, Rocky Hill, NJ, USA). pNK cells werecollected from healthy individuals using the protocol by Favier et al. [26].Briefly, pNK cells were isolated from peripheral blood using the EasySepNK cell depletion kit, as recommended by the manufacturer (Stem Cells,Grenoble, France). The purity of pNK cells was verified systematically byflow cytometry based on CD16, CD56, and CD3 expression. pNK cells wereused only when cell preparations contained �80% CD16�CD56� and weredevoid of CD3� cells. pNK cells were activated by adding 200 UI/ml IL-2(PeproTech, Rocky Hill, NJ, USA) to the culture medium for 36–48 h. Allcell lines were cultured in �MEM (Gibco, Grand Island, NY, USA), contain-ing 2 mM l-glutamine, 1.5 g/L sodium bicarbonate, 12.5% horse serum,and 100 U/ml IL-2. The K562 cell line (ATCC) was maintained in RPMI-1640 medium containing 10% FBS.

Chemical reagents and media�MEM without ribonucleosides and deoxyribonucleosides and RPMI-1640 me-dium without phenol red, horse serum, and FBS were from Invitrogen (Carls-bad, CA, USA). Inositol, 2-ME, folic acid, glutamine, JNK inhibitor SP, ERK1/2inhibitor PD, and p38 MAPK blocker SB were purchased from Sigma-Aldrich(St. Louis, MO, USA). Inhibitors were stored in DMSO at �20°C. 2,3-BTD(Fluka, Sigma-Aldrich) was dissolved in saline and PBS for further assays.

AntibodiesHuman rIL-2 was purchased from PeproTech (London, UK). Mouse poly-clonal antibodies against perforin were from Abnova (Taipei, Taiwan);mouse anti-human NKG2D, NKp30, NKp44, and NKp46 antibodies werefrom R&D Systems (Minneapolis, MN, USA); rabbit polyclonal anti-ERK,JNK, p38, antiphospho-ERK (Thr202/Tyr204), antiphospho-JNK (Thr183/Tyr185), and antiphospho-p38 (Thr180/Tyr182) were from Cell SignalingTechnology (St. Quentin en Yvelines, France), and mouse anti-GAPDHmAb was from Chemicon International (Temecula, CA, USA).

Flow cytometric analysis of NKG2D/NCRsThe cells (1�106 cells/well) were collected by centrifugation at 500 g at4°C, washed once with PBS, and resuspended in 500 ml PBS. The cellswere then incubated with 2% BSA to block nonspecific antigens for 30min, followed by anti-human NKG2D NKp30 and NKp46 antibodies (R&DSystems) and anti-human NKp44 (Becton Dickinson, San Jose, CA, USA)for another 30 min on ice. The binding cells were washed with PBS twice.Finally, the samples were analyzed by flow cytometry using a FACScan flowcytometer (Becton Dickinson).

siRNA transfectionThe siRNAs against human p38 (sc-29433), JNK1 (sc-29380), ERK2 (sc-35335),NKG2D (sc-42948), NKp44 (sc-72170), NKp46 (sc-63344), and NKp30 (sc-42950) and a control siRNA (sc-37007) were purchased from Santa Cruz Bio-technology (Santa Cruz, CA, USA). Transfection of siRNAs into NK92 cells wasperformed according to instructions (Amaxa, Gaithersburg, MD, USA).

Cytotoxicity assaysStress-induced cytotoxicity was evaluated by a CytoTox 96 nonradioactivecytotoxicity assay [27] (Promega, Madison, WI, USA), based on the colori-metric measurement of LDH, a stable cytosolic enzyme released upon celllysis. NK ES (5�104, 2.5�104, and 5�103 cells/well in U-bottom 96-micro-well plates; Corning, Corning, NY, USA) were pretreated with or withoutJNK or ERK1/2 inhibitor for 30 min, followed by treatment with 2,3-BTDfor 16 h. Cells were then washed and resuspended in RPMI 1640 (withoutphenol red), supplemented with 2% FBS. To these were added a fixed

number of K562 TS (5�103/ well) at the E:T ratio of 1:1, 5:1, or 10:1. Themicroplates were then spun for 4 min at 250 g to settle the cell mixturesbefore incubation for 4 h at 37°C in 5% CO2. After this coincubation, su-pernatant (50 �l) was collected from each well, added to 50 �l reconsti-tuted substrate mix, and incubated for 30 min in the dark at room temper-ature. The enzymatic reaction was stopped by Stop Solution (50 �l), andthe absorbance at 490 nm was measured. TM was determined by lysing TSwith 10 �l lysis solution. Spontaneous release by the TS or ES was deter-mined following incubation alone at the respective cell concentrations. Re-sults were expressed as percentage cytotoxicity, calculated as [(experimen-tal�ES�TS)/(TM�TS)] � 100.

RNA isolation and real-time PCRmRNA was isolated from pNK cells using the Qiagen RNeasy kit (Qiagen, Va-lencia, CA, USA), following the manufacturer’s instructions, and total RNA wasconverted into cDNA by standard reverse transcription with the Taqman RTkit (Applied Biosystems, Foster City, CA, USA). qPCR was performed with1/50 of the cDNA preparation in an Mx3000P (Stratagene, La Jolla, CA, USA)in a 25-ml final volume with Brilliant QPCR Master Mix (Stratagene). Thereal-time PCR included an initial denaturation at 95°C for 10 min, followed by40 cycles of 95°C for 30 s, 55°C for 1 min, and 72°C for 1 min and one cycleof 95°C for 1 min, 55°C for 30 s, and 95°C for 30 s.

Western blot analysis of proteinsNK cells (106/flask) were plated for 24 h in IL-2-free medium containing12.5% horse serum and 12.5% FBS. Cells were then treated with the indicatedinhibitors for 30 min, followed by a specific dose of 2,3-BTD for 16 h. Washedcells were lysed in mammalian protein extraction reagent (Pierce Chemical,Thermo Scientific, Rockford, IL, USA). Total lysate protein samples (40 �g/lane) were fractionated on a 10% SDS polyacrylamide gel and blotted ontoPVDF membranes (Immobilon-P, Millipore, Billerica, MA, USA). Membraneswere blocked with 5% nonfat milk for 1 h at room temperature in TBST (Tris10 mM, NaCl 150 mM, pH 7.6, containing 0.1% Tween 20) and probed withprimary antibodies (1:1000 for antiperforin, anti-NKG2D, anti-NKp30, anti-NKp44, anti-NKp46, anti-ERK, anti-JNK, anti-p38, antiphospho-ERK, antiphos-pho-JNK, and antiphospho-p38 and 1:10,000 for anti-GAPDH) overnight at4°C. Membranes were then incubated with appropriate HRP-conjugated sec-ondary antibodies (1:5000). Immunoreactive protein bands were developed byECL (Amersham Pharmacia Biotech, Germany).

Data analysisStatistical analysis was performed using ANOVA (SPSS 12.0 software, SPSS,Chicago, IL, USA) with a correction for multiple comparisons. A differencebetween results of two assay conditions that gave a P value � 0.05 was con-sidered to be significant.

RESULTS

2,3-BTD increases NK92 cell and human pNK cellcytotoxicityWhether 2,3-BTD has any effect on modulation of NK cell im-mune activity was first addressed. We treated NK92 cells with2,3-BTD at different concentrations, followed by coincubationwith K562 cells at an E:T ratio of 1, 5, and 10, respectively. Asshown in Fig. 1A, the cytotoxic activity of NK92 cells was stimu-lated by 2,3-BTD in a dose-dependent manner. At a concentra-tion of 10 �M, the activity was increased by 150% comparedwith that of the untreated cells at the same E:T ratio of 10. Sub-sequently, pNK cells were pretreated with or without desireddoses of 2,3-BTD, followed by measuring NK cell cytotoxic activ-ity. The cytotoxic activity of 2,3-BTD-treated pNK cells was en-hanced dose-dependently compared with that of the nontreated

2 Journal of Leukocyte Biology Volume 92, October 2012 www.jleukbio.org

groups (Fig. 1B; P�0.01). These results indicate that 2,3-BTDtreatment increases NK cell cytotoxicity activity.

2,3-BTD enhances perforin expression in NK cellsThe possible effect of 2,3-BTD on increasing NK cell cytotoxicactivity might be through enhancing perforin and/or granulysinexpression. To determine whether this may be possible, cellswere treated with 2,3-BTD at increasing concentrations for 2 h,and the perforin and granulysin mRNA levels were then quanti-fied by real-time PCR. As shown in Fig. 1C, the abundance ofperforin mRNA was increased by 2,3-BTD treatment in a dose-dependent manner. At a 2,3-BTD concentration of 10 �M, theperforin mRNA levels were increased by 6.8-fold compared withthat of the untreated cells. By contrast, granulysin mRNA expres-sion was not obviously affected. Furthermore, Western blot analy-sis indicated that production of perforin, but not granulysin, wasincreased by 2,3-BTD treatment in a dose-dependent manner(Fig. 1D). At 10 �M, the stimulation by 2,3-BTD was up to seven-fold higher. Intriguingly, the expression of granzyme B was alsoincreased by 2,3-BTD treatment in a dose-dependent manner inpNK cells (Supplemental Fig. 1). 2,3-BTD enhanced the expres-sion and polarization of perforin when the stained pNK cellswere detected by fluorescence microscopy (Supplemental Fig. 2).

Briefly, 2,3-BTD enhances perforin expression and cytotoxicity inNK cells.

2,3-BTD activates perforin expression throughNKG2D/NCRsThe potential effect of 2,3-BTD on mRNA and protein levelexpression of NKG2D and NCRs, which act upstream of perfo-rin, was next evaluated. As shown in Fig. 2A, the NKG2D,NKp44, and NKp46 receptor mRNA levels were clearly en-hanced by 2,3-BTD in a dose-dependent manner. Concomi-tantly, protein production of the three receptors was also en-hanced after 2,3-BTD treatment, and an increase of 4.9-fold inNKp46 and 12.7-fold in NKp44 was most significant (Fig. 2B).In comparison, NKp30 expression was not affected by 2,3-BTD.To see whether NKG2D, NKp30, NKp44, and NKp46 were in-volved in 2,3-BTD activation of perforin production, the pro-tein levels of these receptors were knocked down by RNAi, fol-lowed by measurement of perforin production. As shown inFig. 2C–F, the NKG2D, NKp30, NKp44, and NKp46 proteinlevels were reduced significantly by RNAi treatment as com-pared with the scramble control groups. Simultaneously, 2,3-BTD-stimulated expression of perforin was attenuated signifi-cantly by transfection with the siRNA. Together, 2,3-BTD en-

Figure 1. 2,3-BTD enhances perfo-rin expression and NK cell cyto-toxicity. Cytotoxicity of NK92 cellsand pNK cells against K562 cells(E:T�10:1) was activated by 2,3-BTD in a dose-dependent manner(A and B), also under differentE:T ratios of 1, 5, and 10. Cytotox-icity was measured by LDH release,and data are expressed as percentcytotoxicity. NK cells were treatedwith 2,3-BTD at doses indicated.The mRNA (C) and protein (D)levels of perforin and granulysinwere measured. Data were pre-sented as the means of triplicatedeterminations from three inde-pendent experiments. (†P�0.05;¥P�0.01; #P�0.001).

Lai et al. 2,3-BTD enhances NK cell cytotoxicity

www.jleukbio.org Volume 92, October 2012 Journal of Leukocyte Biology 3

hances perforin expression via increased expression ofNKG2D, NKp44, and NKp46.

2,3-BTD stimulates NKG2D/NCR-mediated ERK1/2and JNK MAPK pathwaysMAPK signaling is implicated in NK cell-mediated tumor celllysis [28, 29]. In addition, inhibition of JNK and/or ERK1/2signaling interferes with NKG2D-mediated cytotoxicity [23].To examine whether 2,3-BTD affects ERK1/2, JNK, and p38activities, the NK cells were treated with different doses of 2,3-BTD, and the levels of phospho-ERK (Thr202/Tyr204), phos-pho-JNK (Thr183/Tyr185), and phospho-p38 (Thr180/Tyr182) were examined by Western blot analysis. As shown in

Fig. 3A, 2,3-BTD stimulated the phosphorylation of ERK1/2and JNK in a dose-dependent manner, whereas it exerted littleor no effect on p38 phosphorylation. To examine whichMAPK signaling pathway is involved in the 2,3-BTD (10 �M)NKG2D/NCR-mediated cytolytic activity, specific inhibitors toindividual MAPKs were used. Inhibition of JNK activity by SP(10 �M) and ERK1/2 activity by PD (10 �M) reduced cytotox-icity by �57% and �23%, respectively. In contrast, inhibitionof p38 by SB (10 �M) had little effect (6%; Fig. 3B). We alsoobserved that there was no difference in cytotoxicity after ERKand JNK inhibitor treatment for 30 min or 16 h (Supplemen-tal Fig. 3). In brief, 2,3-BTD activates NKG2D/NCR-dependentERK1/2 and JNK MAPK activities.

Figure 2. 2,3-BTD enhances perforin production via stimulation of NKG2D and NCR expression. After 2,3-BTD treatment, total cellular mRNAsfrom pNK cells were prepared for quantification of NKG2D and NCR mRNA levels by real-time PCR (A). pNK cells were also subjected to recep-tor protein quantification by flow cytometry analysis (B). For receptor knockdown assay, NK92 cells were transfected with NKG2D (C), NKp30 (D),NKp44 (E), NKp46 (F) or control siRNA (Scramble), followed by knockdown validation and evaluation of perforin production. Data are represen-tative of three independent experiments (†P�0.05; ¥P�0.01; #P�0.001). Allophycocyanin (APC), PE, and FITC, Fluorophore used for labeling re-ceptor-specific antibodies.

4 Journal of Leukocyte Biology Volume 92, October 2012 www.jleukbio.org

ERK activity involves 2,3-BTD-stimulated NK cellperforin production and cytotoxicityTo ascertain whether ERK1/2 activation plays a role in perfo-rin-mediated cytotoxicity in 2,3-BTD-treated NK cells, the cellswere treated with increasing doses of the ERK1/2 inhibitor PD

for 30 min before 2,3-BTD treatment. As shown in Fig. 4A, PDat 5–20 �M significantly attenuated 2,3-BTD (10 �M)-stimu-lated ERK1/2 kinase activity. Furthermore, the dose-depen-dent activation of ERK1/2 phosphorylation and perforin pro-duction by 2,3-BTD was attenuated concurrently by PD at 10

Figure 3. MAPK pathway is acti-vated in 2,3-BTD-treated NK cells.(A) pNK cells were dose-depend-ently treated with 2,3-BTD for 30min before measuring phosphory-lation levels of ERK (P-ERK), JNK(P-JNK), and p38 (P-p38). (B) Ef-fect of MAPK inhibitors for p38(SB, 20 �M), MEK1/2 (PD, 10�M), or JNK (SP, 10 �M) on thecytotoxicity of pNK cells pretreatedwith 10 �M 2,3-BTD is shown. C,Control. Results shown are repre-sentative of three independent ex-periments (¥P�0.01; #P�0.001).

Figure 4. ERK activity involves 2,3-BTD-stimulatedNK cell perforin production and cytotoxicity. (A)Effect of pretreatment of different concentrationsof PD on ERK phosphorylation in pNK cells fol-lowed by 10 �M 2,3-BTD treatment. (B) Effect ofpretreatment with 10 �M PD or buffer control onERK phosphorylation and perforin production inpNK cells treated with different concentrations of2,3-BTD. (C) pNK cells were pretreated with orwithout ERK inhibitor PD (10 �M) for 30 min,followed by treatment with or without differentconcentrations of 2,3-BTD using a 4-h nonradioac-tive LDH release assay. (D) Effect of ERK siRNAon ERK phosphorylation and perforin productionin NK92 cells treated with 10 �M 2,3-BTD. (E)NK92 cells were transfected with ERK2 siRNA orcontrol siRNA, followed by stimulation with or

without 10 ��M 2,3-BTD before the cytotoxicity assay. Results shown are representative of three independent experiments (†P�0.05;¥P�0.01; #P�0.001).

Lai et al. 2,3-BTD enhances NK cell cytotoxicity

www.jleukbio.org Volume 92, October 2012 Journal of Leukocyte Biology 5

�M (Fig. 4B). We next evaluated whether 2,3-BTD-activatedcytotoxic activity could be reduced by inhibition of ERK activ-ity. NK cells were treated with/without 10 �M PD and/or in-creasing 2,3-BTD concentrations, as described, and thenmixed with K562 cells at an E:T ratio of 10:1. As shown in Fig.4C, 2,3-BTD treatment resulted in a significant increase of thecytotoxic activity compared with that of the untreated cells. Asexpected, pretreatment with PD significantly attenuated thecytotoxic activity of the NK cells activated by 2,3-BTD stimula-tion. To further characterize the ERK involvement, NK cellswere transfected with siRNA against ERK2 to determinewhether perforin production stimulated by 2,3-BTD treatmentwas affected. Figure 4D shows that transfection of ERK2 siRNAmarkedly reduced the production of ERK2 and perforin pro-tein after treatment of NK cells with 10 �M 2,3-BTD. More-over, transfection of NK92 cells with ERK2 siRNA markedlyreduced cytotoxicity against K562 cells (Fig. 4E). In contrast,knockdown of p38 did not exhibit any significant effects (datanot shown). In conclusion, ERK activity mediated the 2,3-BTD-stimulated perforin production and cytotoxicity in NK cells.

JNK activity involves 2,3-BTD-stimulated perforinproduction and cytotoxicityThe JNK pathway also involves regulation of perforin expres-sion and cytotoxic activity of NK cells [23]. Whether the JNKinhibitor SP attenuates 2,3-BTD-dependent JNK activation andperforin production was studied next. NK cells were incubatedwith increasing concentrations of SP for 30 min, followed by

2,3-BTD treatment. As shown in Fig. 5A, the JNK inhibitor re-duced 2,3-BTD-stimulated JNK phosphorylation in a dose-de-pendent manner. Concomitant with reduced JNK pathway ac-tivity, 2,3-BTD stimulation of perforin production was also at-tenuated by SP at 10 �M (Fig. 5B). In a second series ofexperiments, NK cells were treated with or without SP for 30min, followed by increasing doses of 2,3-BTD treatment foranother 30 min. The cytotoxic activity was then assayed at anE:T ratio of 10. As shown in Fig. 5C, 2,3-BTD exerted signifi-cant stimulation of the cytotoxic activity against K562 cells.Pretreatment with SP also significantly attenuated the cytotoxicactivity of the NK cells, irrespective of 2,3-BTD stimulation.Furthermore, 2,3-BTD-dependent JNK phosphorylation andperforin production were attenuated significantly by transfec-tion with JNK1 siRNA (Fig. 5D). Moreover, transfection withJNK1 siRNA also led to a significant decline of cytotoxicity(Fig. 5E). Briefly, JNK plays a role in 2,3-BTD-stimulated perfo-rin production and cytotoxicity in NK cells.

Increased NK cell cytotoxicity by 2,3-BTD stimulationis not target–cell-specificTo evaluate whether the increased NK cell cytotoxic activity as aresult of 2,3-BTD treatment is specific for the TS, NK cells weretreated with increasing concentrations of 2,3-BTD, followed bycoculture with HepG2 or A549 cells. Figure 6A and B showedthat NK cell cytotoxic activity against HepG2 or A549 cells is stim-ulated by 2,3-BTD in a dose-dependent manner at the E:T ratioof 10. We used Western blot analysis to determine whether the

Figure 5. Implication of JNK in perforin-mediated cytotoxicityin 2,3-BTD-treated NK cells. (A) Effect of pretreatment of dif-ferent concentrations of SP on JNK phosphorylation in pNKcells treated with 10 �M 2,3-BTD. (B) Effect of pretreatment of10 �M SP or control buffer on JNK phosphorylation and perfo-rin production in pNK cells, followed by treatment with differ-ent concentrations of 2,3-BTD. (C) pNK cells were pretreatedwith or without JNK inhibitor SP (10 �M), followed by treat-ment with or without different concentrations of 2,3-BTD usinga 4-h nonradioactive LDH release assay. (D) Effect of JNKsiRNA on JNK phosphorylation and perforin production inNK92 cells treated with 10 �M 2,3-BTD. (E) NK92 cells weretransfected with JNK1 or control siRNA, followed by treatmentwith or without 10 �M 2,3-BTD before the cytotoxicity assay.Results shown are representatives of three independent experi-ments (†P�0.05; ¥P�0.01; #P�0.001).

6 Journal of Leukocyte Biology Volume 92, October 2012 www.jleukbio.org

common NKG2D ligand-MICA/B is expressed on the three TSlines [30]. The results showed that MICA/B proteins are ex-pressed by all three tumor cell lines tested (Fig. 6C).

DISCUSSION

In this study, we show that 2,3-BTD stimulates NK cell cytotoxicactivity. The activation appears to be a result of activation of themajor NK cell-activating receptors NKG2D, NKp44, and NKp46,leading to the activation of the ERK1/2 and JNK pathways,which, in turn, mediate the up-regulation of perforin productionand perforin-mediated cytotoxicity. Although 2,3-BTD plays multi-ple roles in regulation of physiological functions in bacteria,plants, beetles, and rats, to our knowledge, this is the first reportabout the effect of 2,3-BTD on activation of innate immune cells.The enhanced NK cytotoxic activity is effective, not only againstthe K562 but also A549 and HepG2 cell lines, indicating a gen-eral ability of 2,3-BTD to activate NK cells against tumor cells.Expression of another cytotoxic protein, granulysin, was not af-fected by 2,3-BTD treatment. This phenomenon suggests thatexpression of granulysin might be under control by a differentpathway than perforin. Alternatively, as granulysin is expressed3–5 days after NK cell activation [31], a longer time might beneeded for effects of 2,3-BTD on granulysin expression.

The mechanism of 2,3-BTD effect on activation of NK cellsremains not fully understood. 2,3-BTD is reported to act as asignaling molecule in a wide variety of species. It acts as amale pheromone emitted by the Dynast beetle, Scapanes austra-lis [8]. Furthermore, studies with Arabidopsis mutant lines indi-cated that 2,3-BTD induced systemic drought tolerancethrough the salicylic, ethylene, and jasmonic acid-signalingpathways [7]. Moreover, a role for cytokine-signaling pathwaysin growth promotion of Arabidopsis by bacterial 2,3-BTD wasalso reported [32]. In humans, 2,3-BTD, which is metabolizedfrom ethanol, is present in �5% of normal individuals withserum concentrations �5 �M [1]. The blood concentration of2,3-BTD in nonalcoholics was 0.5–0.8 �M. By contrast, 2 h af-ter alcohol ingestion, blood levels had risen the 2,3-BTD con-centration to 1.2 �M [33]. Results from this study suggestedthat 2,3-BTD involves control of signaling pathways, leading toNK cell activation. As expression of perforin is also under thecontrol of NF-�B in NK cells [34], it is likely that 2,3-BTDtreatment may enhance NF-�B translocation in NK cells. Incomparison, results from our previous in vivo study showedthat pretreatment of rat lungs with 2,3-BTD (10 �mole kg�1)ameliorates LPS-induced acute pneumonia symptoms [10].2,3-BTD appears to inhibit NF-�B translocation through inhibi-

tion of I�B� phosphorylation in rat lung tissue, resulting inreduced expression levels of TNF-�, IL-1�, and IL-6 and a de-crease in acute inflammation [10]. It is interesting that a smallmolecule such as 2,3-BTD can have such distinct and diversebiological effects. One possible explanation is that 2,3-BTDmay resemble or mimic an endogenous signaling molecule.Alternatively, after metabolism into different metabolic inter-mediates, the effects of 2,3-BTD may vary depending on thetiming, doses, and experimental systems to which it is added.Another possibility is it exerts its effects through a series offortuitous interactions with other molecules once inside a cell,ass we have also observed that Fas ligand or TRAIL expressionwas up-regulated in pNK cells with 2,3-BTD 10 �M treatment(Supplemental Fig. 4). All of these are important issuesneeded to be addressed. Based on these reports, 2,3-BTD thusseems to have an immune-modulatory effect. Treatment of 2,3-BTD may not only reduce inflammation of lung tissue cells,but it can also activate resting NK cells. Hence, 2,3-BTD mayaffect different signaling pathways in different TS.

An important future research direction will be the identificationof cell-signaling molecules targeted by 2,3-BTD. The structure of2,3-BTD is rather simple, comprising four carbon atoms and twohydroxyl (�OH) groups. Intriguingly, the 2,3-BTD structure is differ-ent from most compounds or hormones identified. Whether 2,3-BTD interacts directly with NKG2D or NCRs remains to be deter-mined. With the use of the human cell line NK92 as the studymodel, we have observed consistently that knockdown of each of theactivating receptors leads to decreased perforin production. We alsoobserved that in the case of 2,3-BTD treatment, the receptor redun-dancy did not appear to mask the effect of the knockdowns. Theywere reproducible experimental results. Furthermore, besides thepossibility of interacting with extracellular receptors, it is expectedthat 2,3-BTD may diffuse into NK cells and interact with intracellularsignaling mediators. We have reported previously that RSV, a naturalpolyphenolic compound purified from grape seeds, shows a similarability to activate NK cells [11]. An increase in the expression ofNKG2D, activity of MAPKs, and perforin production was also ob-served when NK cells were treated with RSV. Compared with RSV,which enhances only NKG2D production, 2,3-BTD enhancesNKG2D and NCR expression. Thus, although 2,3-BTD and RSV acti-vate NK cell cytotoxicity, the underlying activation mechanismsmight vary, which is worthy of further study.

Although the role of 2,3-BTD in diseases remains unclarified,the production of 2,3-BTD is associated closely with some liver-related diseases. The 2,3-BTD synthesis is increased in severelyalcoholic humans closely related to hepatocelular carcinoma [1].Besides, in patients whose livers were chronically infected with

Figure 6. The effect of 2,3-BTD onpNK cell cytotoxicity against other TS.Results of a 4-h nonradioactive LDHrelease assay at the E:T ratio of 10:1are shown (A and B). (C) Expressionof tumor antigen MICA/B in cell lineK562, A549, and HepG2. GAPDH wasused as an internal loading control.Data are representative of three inde-pendent experiments (#P�0.001).

Lai et al. 2,3-BTD enhances NK cell cytotoxicity

www.jleukbio.org Volume 92, October 2012 Journal of Leukocyte Biology 7

Schistosomiasis japonica, production of 2,3-BTD was also increasedin the patients’ sera [35]. Whether 2,3-BTD shows any effect onamelioration of liver damage is to be studied. It is possible thatincreased 2,3-BTD production may activate NK cells in responseto stresses against liver. This may lead to enhanced innate immu-nity and result in clearance of damaged liver cells. In conclusion,2,3-BTD activates NK cell NKG2D/NCR signaling pathways andenhances NK cell cytotoxicity. Thus, 2,3-BTD should be consid-ered as a compound for activating NK cell activity.

AUTHORSHIP

H-C.L., C-C.L., and C-J.C. designed and performed the re-search, analyzed data, and wrote the manuscript. C-H.Y.,Y-J.H., C-C.C., C-S.L., and D.M.O. helped interpret data andhelped with manuscript preparation. T-T.H. and Y-H.T. wereresponsible for statistical analysis and manuscript preparation.

ACKNOWLEDGMENTS

This work was supported by grants 100-SKH-FJU-17 andNSC100-2320-B-030-001-MY2, respectively, from Shin Kong WuHo-Su Memorial Hospital and National Science Council (Tai-wan) to C.-C.L. and by grants CMRPD-190502 and NSC98-2320-B-182-007-MY3, respectively, from Chang Gung MemorialHospital and National Science Council (Taiwan) to H.-C.L.

REFERENCES

1. Rutstein, D. D., Veech, R. L., Nickerson, R. J., Felver, M. E., Vernon,A. A., Needham, L. L., Kishore, P., Thacker, S. B. (1983) 2,3-Butanediol:an unusual metabolite in the serum of severely alcoholic men duringacute intoxication. Lancet 2, 534–537.

2. Gonzalez, E., Fernandez, M. R., Larroy, C., Pares, X., Biosca, J. A.(2001) Characterization and functional role of Saccharomyces cerevisiae2,3-butanediol dehydrogenase. Chem. Biol. Interact. 130–132, 425–434.

3. Syu, M. J. (2001) Biological production of 2,3-butanediol. Appl. Microbiol.Biotechnol. 55, 10–18.

4. Kovacikova, G., Lin, W., Skorupski, K. (2005) Dual regulation of genesinvolved in acetoin biosynthesis and motility/biofilm formation by thevirulence activator AphA and the acetate-responsive LysR-type regulatorAlsR in Vibrio cholerae. Mol. Microbiol. 57, 420–433.

5. Tsang, L. H., Cassat, J. E., Shaw, L. N., Beenken, K. E., Smeltzer, M. S.(2008) Factors contributing to the biofilm-deficient phenotype of Staphy-lococcus aureus sarA mutants. PLoS One 3, e3361.

6. Ryu, C. M., Farag, M. A., Hu, C. H., Reddy, M. S., Kloepper, J. W., Pare,P. W. (2004) Bacterial volatiles induce systemic resistance in Arabidopsis.Plant Physiol. 134, 1017–1026.

7. Cho, S. M., Kang, B. R., Han, S. H., Anderson, A. J., Park, J. Y., Lee,Y. H., Cho, B. H., Yang, K. Y., Ryu, C. M., Kim, Y. C. (2008) 2R,3R-Bu-tanediol, a bacterial volatile produced by Pseudomonas chlororaphis O6, isinvolved in induction of systemic tolerance to drought in Arabidopsisthaliana. Mol. Plant Microbe Interact. 21, 1067–1075.

8. Rochat, D., Morin, J. P., Kakul, T., Beaudoin-Ollivier, L., Prior, R., Re-nou, M., Malosse, I., Stathers, T., Embupa, S., Laup, S. (2002) Activity ofmale pheromone of Melanesian rhinoceros beetle Scapanes australis. J.Chem. Ecol. 28, 479–500.

9. Izumo, T., Ichiki, C., Saitou, K., Otsuka, M., Ohmori, S., Kamei, C.(1998) Effects of ethanol, acetoin and 2,3-butanediol on EEG powerspectra in conscious rats. Biol. Pharm. Bull. 21, 29–33.

10. Hsieh, S. C., Lu, C. C., Horng, Y. T., Soo, P. C., Chang, Y. L., Tsai,Y. H., Lin, C. S., Lai, H. C. (2007) The bacterial metabolite 2,3-butane-diol ameliorates endotoxin-induced acute lung injury in rats. MicrobesInfect. 9, 1402–1409.

11. Lu, C. C., Chen, J. K. (2010) Resveratrol enhances perforin expressionand NK cell cytotoxicity through NKG2D-dependent pathways. J. Cell.Physiol. 223, 343–351.

12. Elhaik-Goldman, S., Kafka, D., Yossef, R., Hadad, U., Elkabets, M., Val-lon-Eberhard, A., Hulihel, L., Jung, S., Ghadially, H., Braiman, A., Apte,

R. N., Mandelboim, O., Dagan, R., Mizrachi-Nebenzahl, Y., Porgador, A.(2011) The natural cytotoxicity receptor 1 contribution to early clear-ance of Streptococcus pneumoniae and to natural killer-macrophage crosstalk. PLoS One 6, e23472.

13. Yokoyama, W. M., Kim, S., French, A. R. (2004) The dynamic life of nat-ural killer cells. Annu. Rev. Immunol. 22, 405–429.

14. Brill, K. J., Li, Q., Larkin, R., Canaday, D. H., Kaplan, D. R., Boom,W. H., Silver, R. F. (2001) Human natural killer cells mediate killing ofintracellular Mycobacterium tuberculosis H37Rv via granule-independentmechanisms. Infect. Immun. 69, 1755–1765.

15. Trapani, J. A., Smyth, M. J. (2002) Functional significance of the perfo-rin/granzyme cell death pathway. Nat. Rev. Immunol. 2, 735–747.

16. de Saint Basile, G., Menasche, G., Fischer, A. (2010) Molecular mecha-nisms of biogenesis and exocytosis of cytotoxic granules. Nat. Rev. Immu-nol. 10, 568–579.

17. Elboim, M., Gazit, R., Gur, C., Ghadially, H., Betser-Cohen, G., Mandelboim,O. (2010) Tumor immunoediting by NKp46. J. Immunol. 184, 5637–5644.

18. Li, Y., Wang, Q., Mariuzza, R. A. (2011) Structure of the human activat-ing natural cytotoxicity receptor NKp30 bound to its tumor cell ligandB7-H6. J. Exp. Med. 208, 703–714.

19. Campbell, K. S., Yusa, S., Kikuchi-Maki, A., Catina, T. L. (2004) NKp44 triggersNK cell activation through DAP12 association that is not influenced by a puta-tive cytoplasmic inhibitory sequence. J. Immunol. 172, 899–906.

20. Barber, A., Sentman, C. L. (2011) NKG2D receptor regulates humaneffector T-cell cytokine production. Blood 117, 6571–6581.

21. Xu, X., Rao, G. S., Groh, V., Spies, T., Gattuso, P., Kaufman, H. L.,Plate, J., Prinz, R. A. (2011) Major histocompatibility complex class I-re-lated chain A/B (MICA/B) expression in tumor tissue and serum ofpancreatic cancer: role of uric acid accumulation in gemcitabine-in-duced MICA/B expression. BMC Cancer 11, 194.

22. Maghazachi, A. A. (2005) Insights into seven and single transmembrane-spanning domain receptors and their signaling pathways in human nat-ural killer cells. Pharmacol. Rev. 57, 339–357.

23. Li, C., Ge, B., Nicotra, M., Stern, J. N., Kopcow, H. D., Chen, X.,Strominger, J. L. (2008) JNK MAP kinase activation is required forMTOC and granule polarization in NKG2D-mediated NK cell cytotoxic-ity. Proc. Natl. Acad. Sci. USA 105, 3017–3022.

24. Chini, C. C., Boos, M. D., Dick, C. J., Schoon, R. A., Leibson, P. J.(2000) Regulation of p38 mitogen-activated protein kinase during NKcell activation. Eur. J. Immunol. 30, 2791–2798.

25. Gong, J. H., Maki, G., Klingemann, H. G. (1994) Characterization of ahuman cell line (NK-92) with phenotypical and functional characteris-tics of activated natural killer cells. Leukemia 8, 652–658.

26. Favier, B., Lemaoult, J., Lesport, E., Carosella, E. D. (2010) ILT2/HLA-G interaction impairs NK-cell functions through the inhibition ofthe late but not the early events of the NK-cell activating synapse. FASEBJ. 24, 689–699.

27. Bruno, A. P., Lautier, D., d’Orgeix, A. T., Laurent, G., Quillet-Mary, A.(2000) Acute myeloblastic leukemic cells acquire cellular cytotoxicityunder genotoxic stress: implication of granzyme B and perforin. Blood96, 1914–1920.

28. Jiang, K., Zhong, B., Gilvary, D. L., Corliss, B. C., Hong-Geller, E., Wei,S., Djeu, J. Y. (2000) Pivotal role of phosphoinositide-3 kinase in regula-tion of cytotoxicity in natural killer cells. Nat. Immunol. 1, 419–425.

29. Wei, S., Gamero, A. M., Liu, J. H., Daulton, A. A., Valkov, N. I., Trapani, J. A.,Larner, A. C., Weber, M. J., Djeu, J. Y. (1998) Control of lytic function by mito-gen-activated protein kinase/extracellular regulatory kinase 2 (ERK2) in a hu-man natural killer cell line: identification of perforin and granzyme B mobiliza-tion by functional ERK2. J. Exp. Med. 187, 1753–1765.

30. Borchers, M. T., Harris, N. L., Wesselkamper, S. C., Vitucci, M., Cos-man, D. (2006) NKG2D ligands are expressed on stressed human airwayepithelial cells. Am. J. Physiol. Lung Cell. Mol. Physiol. 291, L222–L231.

31. Krensky, A. M., Clayberger, C. (2005) Granulysin: a novel host defensemolecule. Am. J. Transplant. 5, 1789–1792.

32. Ryu, C. M., Farag, M. A., Hu, C. H., Reddy, M. S., Wei, H. X., Pare,P. W., Kloepper, J. W. (2003) Bacterial volatiles promote growth in Ara-bidopsis. Proc. Natl. Acad. Sci. USA 100, 4927–4932.

33. Montgomery, J. A., Jette, M., Brunengraber, H. (1990) Assay of physio-logical levels of 2,3-butanediol diastereomers in blood and urine by gaschromatography-mass spectrometry. Anal. Biochem. 185, 71–76.

34. Zhou, J., Zhang, J., Lichtenheld, M. G., Meadows, G. G. (2002) A rolefor NF-� B activation in perforin expression of NK cells upon IL-2 re-ceptor signaling. J. Immunol. 169, 1319–1325.

35. Nishina, M., Kato, K., Matsushita, K., Hayashi, M., Matsuda, H. (1997)Proton nuclear magnetic resonance detects leucine, 2,3-butanediol, anda prominent increase in the level of choline in the sera from patientschronically infected with Schistosomiasis japonica. Physiol. Chem. Phys. Med.NMR 29, 55–61.

KEY WORDS:perforin � NKG2D � NCR � resveratrol

8 Journal of Leukocyte Biology Volume 92, October 2012 www.jleukbio.org