Cytotherapy, 2015; 0: 1e8

Ascorbic acid promotes proliferation of natural killer cell populationsin culture systems applicable for natural killer cell therapy

MIRELLE J.A.J. HUIJSKENS1,*, MATEUSZ WALCZAK1,*, SUBHASHIS SARKAR1,FLORANCE ATRAFI1, BIRGIT L.M.G. SENDEN-GIJSBERS1, MARCEL G.J. TILANUS2,GERARD M.J. BOS1, LOTTE WIETEN2,y & WILFRED T.V. GERMERAAD1,y

1Department of Internal Medicine, Division of Haematology, Maastricht University Medical Center, Maastricht, TheNetherlands, and 2Department of Transplantation Immunology, Maastricht University Medical Center, Maastricht, TheNetherlands

AbstractBackground aims.Natural killer (NK)cellebased immunotherapy is apromising treatment for a varietyofmalignancies.However,generating sufficient cell numbers for therapy remains a challenge. To achieve this, optimization of protocols is required. Methods.Mature NK cells were expanded from peripheral blood mononuclear cells PBMCs in the presence of anti-CD3 monoclonalantibody and interleukin-2. Additionally, NK-cell progenitors were generated from CD34þ hematopoietic stem cells ordifferent T/NK-cell progenitor populations. Generated NK cells were extensively phenotyped, and functionality was deter-mined by means of cytotoxicity assay. Results. Addition of ascorbic acid (AA) resulted in more proliferation of NK cells withoutinfluencing NK-cell functionality. In more detail, PBMC-derived NK cells expanded 2362-fold (median, range: 90e31,351) inthe presence of AA and were capable of killing tumor cells under normoxia and hypoxia. Moreover, hematopoietic stemcellederived progenitors appeared to mature faster in the presence of AA, which was also observed in the NK-cell differen-tiation from early T/NK-cell progenitors. Conclusions. Mature NK cells proliferate faster in the presence of phospho-L-AA,resulting in higher cell numbers with accurate functional capacity, which is required for adoptive immunotherapy.

Key Words: ascorbic acid, cancer, cellular immunotherapy, hematopoietic stem cell, NK cell

Introduction

Natural killer (NK) cellebased immunotherapy is apromising approach for treatment of malignancies.NK cells can kill cancer cells without the need forprior direct sensitization [1e3] and without causinggraft-versus-host disease (GVHD) [4]. NK cells are asubset of cytokine-producing cytotoxic innatelymphoid cells that express CD56 and lack expres-sion of CD3 [4]. A well-balanced mechanism torecognize and kill infected or malignant cells whiletolerating healthy cells makes infusion of NK cells afeasible method to treat malignancies [5,6].

Preclinical and clinical studies have shown that ifinfused in large numbers, NK cells can be used toeliminate malignant cells [7]. Both autologous andallogeneic NK cells are able to effectively eliminatecancer cells in vitro [8,9]. Furthermore, it has beenshown that interleukin (IL)-2eactivated NK cells

*These authors contributed equally to this work.yThese authors contributed equally to this work.Correspondence: Wilfred T.V. Germeraad, PhD, Department of Internal Medic6229ER Maastricht, The Netherlands. E-mail: w.germeraad@maastrichtuniversity

(Received 22 August 2014; accepted 6 January 2015)

ISSN 1465-3249 Copyright � 2015, International Society for Cellular Therapy. Phttp://dx.doi.org/10.1016/j.jcyt.2015.01.004

from patients are effective against autologous cancercells in vivo in a mouse model [10]. The beneficialeffect on prolongation of survival in a tumor-bearingmouse model has been demonstrated to be depen-dent on the number of infused syngeneic NK cells[11]. Although infusion of autologous NK cells hasbeen shown to be safe, only limited efficacy is shownin several clinical trials [7,12,13]. A more powerfulapproach proposed by Ruggeri et al. [14] and Velardiet al. [15] showed that allogeneic NK-cell transferleads to higher tumor cytotoxicity in acute myeloidleukemia because of killer immunoglobulin-like re-ceptor (KIR)-ligand mismatching, which lowers theactivation threshold of NK cells. Similar effects havealso been seen in responses to solid tumors in both amouse model and in patients [16,17]. Takentogether, adoptive NK-cell studies demonstrate the

ine, Division of Haematology, Maastricht University, Universiteitssingel 50,.nl

ublished by Elsevier Inc. All rights reserved.

2 M. J.A.J. Huijskens et al.

possibilities of NK-cell therapy for a variety of can-cers, but many indicate the necessity of sufficient cellnumbers because of limited in vivo responses [18,19].

NK cells for immunotherapy can be derived fromseveral sources, for example, from peripheral blood,(induced) stem cells, cord/placental blood or bonemarrow. Clinical studies that used enriched periph-eral blood NK cells have shown that it is not possibleto produce sufficient numbers of NK cells with theright purity and state of activation [20,21]. To over-come this problem, in vitro expansion of NK cells iscurrently under investigation. Different methods forexpansion have been developed to obtain NK cells forinfusion, which is extensively reviewed by Cheng et al.[7]. Although great improvement in differentmethods of NK-cell expansion and generation havebeen achieved, for example, with certain cytokinecocktails and culture instruments, clinical success isstill limited, which indicates that it is critical to searchfor additional approaches to improve NK cell matu-ration and/or proliferation to obtain sufficientnumbers of properly selected NK cells for immuno-therapy [7,19]. Besides the number of NK cells, alsopurity of the NK-cell product (to prevent GVHD),culture time (reducing cost of the clinical product)and phenotype of the NK cells are important areas forimprovement. The latter is essential because weshowed in multiple myeloma and in breast cancer thatKIR-ligandemismatched NK cells are much moreefficient in eliminating tumor cells than matched NKcells [16] (S. Sarkar, Maastricht University, unpub-lished observations, 2013). However, to obtain suffi-cient numbers of this relatively infrequent NK-cellpopulation requires efficient ex vivo expansion.

Previously, we have shown that T/NK-cell pro-genitors can develop from hematopoietic stem cells(HSCs) in co-culture with feeder cells [22]. Inanother recent study, we have demonstrated thatespecially phosphorylated ascorbic acid (AA) has apositive effect on T/NK-cell progenitors; T(/NK)-cellprogenitors mature further and, importantly, prolif-erate faster in the presence of AA [23]. Therefore, wehypothesized in the current study that AA can be usedto improve ex vivo NK-cell expansion protocols. Totest the hypothesis, mature NK cells were expandedfrom PBMCs. Additional methods in which HSCswere cultured with feeder cells in NK-celleskewingconditions and in which NK cells were generatedfrom T/NK-cell progenitors were also investigated.

Methods

Cell lines

TSt-4 cells, an earlier gift of Prof Dr H. Kawamoto(Kyoto University, Japan), were grown in standard

Roswell Park Memorial Institute (RPMI) 1640medium (Sigma-Aldrich Co) containing 5% fetalbovine serum (Greiner Bio One), 1% penicillin-streptomycin, 1 mmol/L sodium pyruvate,0.1 mmol/L minimum essential medium non-essential amino acids and 5 � 10�5 mol/L 2-mercaptoethanol (all from Invitrogen Ltd), referredto as complete RPMI medium. K562 cells (ATCC)were cultured in RPMI 1640 medium (Gibco) sup-plemented with 10% fetal bovine serum (Integro),100 U/mL penicillin (Gibco) and 100 mg/mLstreptomycin (Gibco). All cultures were maintainedat 37�C in humidified air containing 5% CO2.

NK-cell expansion from peripheral blood mononuclearcells

Peripheral blood mononuclear cells (PBMCs) wereisolated from healthy adult donors with the use ofLymphoprep (Axis Shield). PBMCs, containing NKcells, were expanded in CellGro stem cell growthmedium (CellGenix) supplemented with 5% humanserum (BioWhittaker Lonza), 1000 U/mL penicillinand 100 mg/mL streptomycin (Gibco), 10 ng/mLanti-CD3 monoclonal antibody (OKT-3) and 1000IU/mL IL-2 (both from Miltenyi). Cells werecultured in six-well plates in a humidified incubatorat 37�C, 5% CO2, with an initial total culture volumeof 2 mL per well, with a total number of 2 � 106 cellsper well. Parts of the cultures were supplementedwith 50 mg/mL AA, a dose determined previously[23]. Every 3 to 4 days, medium was refreshedtogether with addition of freshly prepared AA. Whenrequired, cells were counted and replated to thestarting concentration. The cultures were maintainedfor 3 to 4 weeks.

Flow cytometry analysis of NK cells and NK-cellprogenitors

The following antibodies were used for phenotypicanalysis of NK cells and NK-cell progenitors:CD11a/lymphocyte functioneassociated antigen 1(LFA1) (HI111), CD16 (3G8), CD45 (HI30),CD57 (TB01, DAKO) and CD158e/3DL1 (DX9,Miltenyi Biotec), all fluorescein isothiocyanateelabeled; CD45 (HI30), CD158b (2DL2/3, Miltenyi),CD159c/NKG2C (134591, R&D Systems), CD244/2B4 (R&D systems), CD253/tumor necrosis fac-torerelated apoptosis-inducing ligand (TRAIL)(RIK-2.1, Miltenyi) and CD335/NKp46 (9E2), allphycoerythrin-labeled; CD45 (2D1) and CD159a/NKG2a (Z199, Beckman Coulter), peridinin chlo-rophyll proteinelabeled; CD45 (2D1), CD158a/2DL1 (143211, R&D systems), CD178/Fas ligand(NOK-1, Miltenyi), CD184/CXCR4 (12G5),

Figure 1. AA enhances proliferation of mature NK cells. NK cells were expanded in vitro from PBMCs in medium supplementedwith OKT-3 and IL-2 for 3 weeks in the presence of absence of AA. Percentage of CD56þCD3e NK cells and CD3þCD56e T cellson different days of culture from two representative donors are shown (A). Percentages of CD56þCD3e NK cells in the presence orabsence of AA are shown (n ¼ 7, P ¼ 0.0156, D0 versus D21 control; P ¼ 0.0313, D0 versus D21 AA; B). Fold expansions ofCD56þ NK cells generated in the presence or absence of AA on day 21 of culture are shown (n ¼ 7, P ¼ 0.0156; C). ns, notsignificant.

CD226/DNAX accessory molecule-1 (DNAM1)(102511, R&D Systems) and CD314/NKG2D(BAT221, Miltenyi), all allophycocyanin (APC)-labeled; CD45 (2D1) and CD3 (SK1), APC-Cy7elabeled; and CD45 (HI30) and CD56(B159), Horizon V450-labeled. Isotypes of theequivalent antibodies were used as control refer-ences. All antibodies, materials and equipment wereobtained from BD Biosciences unless stated other-wise. Samples were measured with the flow cytome-ter FACS Canto II (BD). Flow cytometric analysiswas performed with the use of BD FACS DIVAsoftware version 6.1.2 (BD) or FlowJo softwareversion 10.0.6 (Treestar). Living cells were gated onforward and sideward scatter pattern, with subse-quent doublet removal.

Cytotoxicity assay

The cytotoxic potential of NK cells was determined ina kill assay as described previously [24]. K562target cells were labeled with 3,30-dioctadecylox-acarbocyanine perchlorate (DiO, Sigma-Aldrich Co).

Both effector and target cells were individuallypre-incubated for 14 to 16 h at 21% or 0% to 2% O2

and subsequently combined at different effector-to-target ratios (E:T) in 96-well, round-bottomedplates in duplicate. After 4 h, samples weremeasured with the use of the Canto II (BD), and celldeath of DiOþ target cells was measured with propi-dium iodide (PI, Sigma-Aldrich Co). Specific cyto-toxicity was determined by means of the equation (%PIþ target cells e % spontaneous PIþ cells)/(100 e %spontaneous PIþ cells) � 100. The cytotoxic potentialof NK-cell progenitors was only studied in normoxicconditions.

Statistical analysis

Data are presented as median. All statistical analyseswere performed with the use of the Prism program(GraphPad Software Inc). Differences betweenexperimental conditions were evaluated for statisticalsignificance by means of the non-parametric Wil-coxon matched pairs test. Significance was acceptedat the levels of P < 0.05 and P < 0.01.

Figure 2. NK cells expanded in the presence of AA express all required activation and inhibitory molecules and are functional. NK cellswere expanded from PBMCs in medium supplemented with OKT-3 and IL-2 and in the presence or absence of AA. After 3 weeks ofculture, expanded NK cells were stained with phenotypic markers. Combined data from four different donors are shown. Significance isindicated as *P < 0.05 and **P < 0.01 (A). Specific cytotoxicity of NK cells expanded with or without AA was measured on day 21 ofculture. NK-cell progenitors (effector) and K562 leukemic target cells were mixed in different E:T ratios, and killing of target cells wasmeasured after 4-h incubation. Specific cytotoxicity of NK cells both under normoxia (21% O2; B) and hypoxia (0.2% O2) is shown (C). Pvalues indicate no significant differences: P ¼ 0.5, normoxia 1:1; P ¼ 0.75, normoxia 1:10; P ¼ 0.85, hypoxia 1:1; P ¼ 0.25, hypoxia 1:10.

4 M. J.A.J. Huijskens et al.

Results

Ex vivo expansion of NK cells is enhanced by additionof AA

To investigate whether AA has an effect on theproliferation of mature NK cells, NK cells fromtotal PBMCs were expanded in the presence of AA,IL-2 and OKT-3 (anti-CD3). At the start of theculture, 2.9% to 20.4% of CD56þ cells were present(Figure 1A,B). Although variation existed amongdifferent donors, the percentage of CD56þ cells forall donors in both conditions was higher at day 21 ofthe culture compared with day 0 (Figure 1B).Generally, two different culture phenotypes were

observed. One resulted in a high percentage of NKcells, whereas the other showed comparable per-centages of CD3þ cells, and CD56þCD3þ cellswere present (Figure 1A,B). Addition of AA did notinfluence the percentage of CD56þ NK cells ascompared with day 21 in the absence of AA. How-ever, NK cells proliferated more efficiently in thepresence of AA, and the total fold expansion(number of NK cells at day 21/number of NK cellsat day 0) was higher in AA-supplemented culturesthan in cultures without AA (Figure 1C; P ¼0.0156). In summary, these data show that AAimproves ex vivo proliferation of mature NK cells,resulting in a 2362-fold (median, range,

Ascorbic acid promotes expansion of NK cells 5

90e31,351) expansion of NK cells in 3 weeks ofculture.

Ex vivoeexpanded NK cells produced in the presence ofAA are functional

To characterize the functionality of ex vivoeexpandedmature NK cells, both phenotype and cytotoxiccapacity of these cells were studied. Detailed analysisof several activation, inhibitory, functional, homingand maturation receptors was performed by means offlow cytometry on day 0 and day 21 of culture.Representative histograms from a selected donor forall analyzed receptors are shown in SupplementaryFigure 1. Expression of NKp46, CD16, KIR3DL1and KIR2DL1 were not influenced by in vitro culture(Figure 2A). However, expanded cells expressedmore of the activation receptors NKG2D, DNAM-1and 2B4 compared with day 0 NK cells. Further-more, higher percentages of NKG2Cþ cells werepresent at the end of the culture. Additionally, higherpercentages of CD56þ cells expressed the inhibitoryreceptor NKG2A, whereas the expression ofKIR2DL2/3 differed after culture (Figure 2A).Expanded NK cells were characterized by lowerexpression of CD57. Moreover, TRAIL, LFA-1,CXCR4 and LFA1 expression were increased onCD56þ cells after culture. Comparison of receptorspresent on NK cells cultured in the presence orabsence of AA revealed that AA did not influence thephenotype of the expanded NK cells (Figure 2A).

The ex vivoeexpanded NK cells were able toefficiently lyse leukemic K562 cells, and the amountof lysis was comparable for NK cells expanded withor without AA (Figure 2B). For clinical application,NK-cell products should be able to kill tumor cells ina suppressive tumor environment. Hypoxia is atumor-associated factor, and we recently publishedthat hypoxia can severely reduce the killing capacityof unactivated NK cells; however, on IL-2 activation,NK cells did kill tumor cells in the hypoxic envi-ronment [24]. To investigate whether supplementa-tion with AA influenced the killing capacity of NKcells, we performed killing experiments under nor-moxia and hypoxia. Also, under hypoxic conditions,IL-2eexpanded NK cells were able to lyse targetcells, which indicates that expanded NK cells areefficient in eliminating tumor cells under morephysiological conditions (Figure 2C). There was nosignificant difference in the cytotoxic capacity of NKcells cultured in the absence or presence of AA(Figure 2C). Taken together, these data indicate thatin the presence of AA during the expansion of NKcells, the phenotype was not influenced. Importantly,NK cells expanded in the presence of AA expressedall required receptors and were fully functional.

AA improves the generation and expansion of NK-cellprogenitors from HSCs

Besides the expansion of mature NK cells, NK cellscan also be generated from stem or progenitor cells.To investigate whether AA has an effect on the gen-eration of NK-cell progenitors from HSCs, NK-cellprogenitors were generated from CD34þCD38e/dim

HSCs co-cultured with TSt-4 feeder-cells in NKcelleskewing conditions in the presence or absenceof AA in two independent experiments. In thepresence of AA, cells acquired CD56 expressionfaster compared with the cells cultured without AA.After 35 days of culture, the vast majority of cellsexpressed CD56 (78.3%) in the presence of AA,whereas the percentage of CD56þ only marginallyincreased to 15.8% in the absence of AA. Further-more, NK-cell progenitors generated in the presenceof AA had higher expression of activating receptorsincluding NKp46 and NKG2D, whereas expressionof inhibitory receptors (KIR2DL1, KIR2DL2/3,KIR3DL1 or NKG2A) was comparable in bothconditions as indicated by mean fluorescence in-tensity (Supplementary Figure 3B). At day 35 ofculture, AA supplementation had also resulted in ahigher fold expansion (809-fold) of CD56þ NK-cellprogenitors (number of CD56þ cells at day 35/number of CD56þ cells at day 0) compared with thecontrol condition without AA supplementation (78-fold; Supplementary Figure 3C).

AA improves the generation of NK-cell progenitors fromT/NK-cell progenitors

To provide further support for the beneficial role ofAA in NK-cell expansion systems, the influence ofAA on the capacity to generate NK-cell progenitorsfrom T/NK-cell progenitors was studied. T/NK-cellprogenitors were produced from HSCs in co-culture with TSt-4/DLL4 feeder cells in two inde-pendent experiments. HSCs were first co-culturedwith TSt-4/DLL4 feeder cells in T celleskewingconditions. After 21 days of culture, three pop-ulations of T-cell progenitors were sorted on thebasis of expression of CD7 and CD5 (CD7eCD5e,CD7þCD5e and CD7þCD5þ) and were depleted ofCD56þ cells. The purity of all obtained progenitorpopulations was higher than 97% (SupplementaryFigure 4). Hereafter, the different progenitor pop-ulations were cultured in NK-celleskewing condi-tions. CD56þ NK-cell progenitors were generatedfrom both CD7eCD5e and CD7þCD5e progenitors(Supplementary Figure 5A). AA increased the per-centage of CD56þ NK-cell progenitors obtainedfrom both early T/NK-cell progenitor populations.Remarkably, NK-cell progenitors could barely be

6 M. J.A.J. Huijskens et al.

generated from the most mature CD7þCD5þ pop-ulation. Regardless of the presence or absence of AA,these progenitors died when cultured in NK celleskewing conditions (Supplementary Figure 5A).

Phenotypic analysis of generated NK-cell pro-genitors from the CD7�CD5� and CD7þCD5�

progenitors revealed that cells generated in thepresence of AA expressed slightly more of the acti-vating receptors NKp46 and NKG2D and inhibitoryreceptor NKG2A (Supplementary Figure 5B).Furthermore, proliferation of NK-cell progenitorsgenerated from the CD7eCD5e and CD7þCD5e

progenitors was positively influenced by addition ofAA (Supplementary Figure 5C). Moreover, gener-ated NK-cell progenitors, whether cultured in thepresence or absence of AA, were able to kill K562leukemic cells (Supplementary Figure 4D).

Taking the results of experiments on differentia-tion to NK cells from early progenitors (HSCs andCD5eCD7e/CD5-CD7þ T/NK-cell progenitors)provided in Supplementary Figures 3 and 5 togethersuggests that the positive influence of AA is notmerely on proliferation but also on differentiation.

Discussion

This study demonstrates an important role forphosphorylated AA on the expansion of NK cells.We show that the expansion of NK cells from pe-ripheral blood is significantly increased in the pres-ence of AA, resulting in functional NK cells capableof killing tumor cells under normoxia and hypoxia.On the basis of this observation, we additionallyinvestigated whether AA has a positive influence onNK cells generated from HSCs in two differentsystems. Although the sample size was limited, thedata obtained from HSCs were in line with the dataobtained with peripheral blood NK cells and furthersupport the concept that AA has a positive effect onNK-cell proliferation and suggest a positive influenceon NK-cell differentiation. Together, these resultsindicate a positive role for AA in in vitro NK-cellculture systems. Because AA is an inexpensive andreadily available compound, our data provide proofof concept that AA supplementation can be an easyway to improve NK-cell expansion protocols result-ing in higher cell numbers as required for adoptiveimmunotherapy.

Different culture methods for the production ofNK cells for immunotherapy exist and are currentlyunder investigation [7,19]. These methods canroughly be divided into mature NK-cell products(eg, PBMC-derived) or immature NK-cell products(eg, from HSCs or progenitor cells). The expansionmethod for mature NK cells started with PBMCs

cultured in the presence of IL-2 and OKT-3.Whereas IL-2 can act on both T and NK cells,OKT-3 (anti-CD3) presumably first exerts its effecton the T cells present in the culture, which subse-quently create a milieu favorable for NK-cellexpansion, that is, in a contact- or cytokine-dependent manner. T cells disappear from thesecultures, probably because of exhaustion (Wietenet al., unpublished observations, 2012). In combi-nation with a mature phenotype of the expanded cellsand the ability of these cells to kill cancer cells, thispopulation is suggested to be the most suitable fortherapy. However, donor variation must be takeninto account. In our hands, expansions from somedonors resulted in high NK-cell percentages andrelatively low CD3þ T-cell percentages, whereasothers resulted in increased though less than 50%NK cells in the presence of high numbers of CD3þ Tand CD3þCD56þ NKT cells. To prevent GVHD,CD3þ T cells should be excluded from the product.Nonetheless, AA improved expansion in all donorsto an average of w6000-fold. Others have shown inclinical trials that PBMC-derived, expanded NKcells are safe and capable of exerting anti-tumor ef-fects [7]. The benefit of the use of mature NK-cellproducts is that donors and patients can be mis-matched for their KIR receptors, promising for agreater anti-tumor effect [25].

Because of the positive effect on mature NK-cellexpansion, we additionally investigated the influenceof AA in other NK culture systems to find proof ofconcept. Also in directing cells in the NK-cell lineagefrom both HSCs and T-cell progenitors, a positiveinfluence of AA was observed. On the one hand, itappears that addition of AA results in faster genera-tion of NK cells; on the other hand, even after longculture in the absence of AA, fewer NK cells areobserved. This suggests that AA makes more cellssensitive for differentiation into the NK-cell lineagein a proper NK induction environment.

NK and T cells share a common progenitor [26].We previously showed that cells in a mixed T/NK-cell progenitor population derived from HSCs hadthe potency to become NK cells [22]. In this latterstudy, the complete progenitor population was used.Only very few NK cells derived from theCD7þCD5þ could be detected, which suggests thatthese cells have lost most of their NK-cell potentialand are more committed to the T-cell lineagecompared with the more immature populations.However, more extensive research is required tomake this conclusion.

Currently, NK-cell progenitors produced fromHSCs and T/NK-cell progenitors are co-culturedwith feeder cells, which require selection of NKcells before clinical use in patients. We recently

Ascorbic acid promotes expansion of NK cells 7

established a feeder-free culture system in which, inthe presence of AA, T cells can be generated andadapting the growth factors to an NK-celleskewingprofile may be applicable [23]. The high purity of theNK cells obtained in these systems as the result of thepresence of AA is a great advantage. Moreover, in thepresence of AA, the expansion of NK cells fromHSCs is increased by 100-fold. The generated NKcells in these systems are still progenitors, becausethe majority does not express KIR receptors. How-ever, a recent study revealed that NK cells generatedfrom HSCs in a clinical setting do have the capacityto inhibit growth of leukemic cells, resulting in pro-longed survival in a murine tumor model [27].Furthermore, it has been shown that CD34þ HSC-derived NK cells do not show acute side effects oninjection in patients; unfortunately, no anti-tumoractivity of the NK cells was measured, probablybecause of insufficient NK-cell numbers [28].Another advantage of the generation of NK cellsfrom HSCs is that a minor fraction of the same HSCsinjected for HSCT in the patient could be used forthe generation of NK cells, minimizing the risk ofrejection of the NK cells.

Regardless of the source or preparation of the NKcells, they must be able to kill tumor cells in vivo. Formany years, there have been investigations in thetumor micro-environment, and abundant knowledgehas been generated on escape mechanisms from theimmune system by tumor cells. One of these factorsof the tumor micro-environment is hypoxia, whichhas been shown to contribute to therapy resistance ofmalignant cells [29,30]. Previously, we showed thatthe killing capacity of NK cells is reduced underhypoxia, which can be restored by IL-2 activation ofthe NK cells [24]. In the present study, we show thatIL-2eactivated, expanded mature NK cells kill can-cer cells under hypoxia also in the presence of AA,which confirms the functionality in a more physio-logical relevant setting.

Others have succeeded in generating clinical-grade NK cells applicable for therapy, as exten-sively reviewed by Cheng et al. [7]. Media regularlyused for NK-cell expansions are Glycostem, RPMI1640, Dulbecco’s modified Eagle’s medium andCellGro stem cell growth medium [7,19]. Althoughin general some commercially available cell culturemedia already contain AA, most of the media usedfor NK-cell expansion do not. Furthermore, if pre-sent, AA is in the non-phosphorylated form, which isless stable and has far less potency to stimulate cellsto proliferate than does the phosphorylated com-pound [23,31]. Therefore, our observations stronglysuggest that the controlled addition of phosphory-lated AA in any NK-cell expansion system will beadvantageous, and, if already present, optimized

concentrations are recommended. Potentially, thiscould result in faster expansion of NK cells, resultingin higher yields of NK cells in shorter culture time.This will lead to a marked cost reduction and theavailability of a higher number of NK cells forimmunotherapy. Our recent results show that onlysubpopulations of the NK-cell pool are efficienteffector cells (unpublished observations, S. Sarkar,Maastricht University, 2013). Because these subsetshave a relatively low frequency, very high cellnumbers are needed to be infused into patients.Fortunately, AA is available in clinical grade andcould immediately be added to existing NK-cellculture methods.

In summary, AA promotes the proliferation ofNK-cell populations without affecting their func-tionality. These findings are relevant for theimprovement of methods to generate sufficient NKcells for adoptive therapy.

Acknowledgments

This work was supported by a grant from the DutchCancer Society KWF: UM2010-4671 and with finan-cial support from the Cancer Research Fund Limburgof theHealth FoundationLimburg. S.S. was supportedby a PhD student grant from GROW, School ofOncology and Developmental Biology, MaastrichtUniversity Medical Center. L.W. was supported by apersonal grant from the Dutch Cancer Society.

Disclosure of interests: The authors have no com-mercial, proprietary, or financial interest in the productsor companies described in this article. Gerard Bos andWilfred Germeraad are founders of CiMaas BV, astartup company beginning March 1, 2015.

References

[1] Caligiuri MA. Human natural killer cells. Blood 2008;112:461e9.

[2] Lanier LL. Up on the tightrope: natural killer cell activationand inhibition. Nat Immunol 2008;9:495e502.

[3] Moretta L, Moretta A. Unravelling natural killer cell func-tion: triggering and inhibitory human NK receptors. EMBO J2004;23:255e9.

[4] Kang L, Voskinarian-Berse V, Law E, Reddin T, Bhatia M,Hariri A, et al. Characterization and ex vivo Expansion ofHuman Placenta-Derived Natural Killer Cells for CancerImmunotherapy. Front Immunol 2013;4:101.

[5] Karre K, Ljunggren HG, Piontek G, Kiessling R. Selectiverejection of H-2-deficient lymphoma variants suggestsalternative immune defence strategy. Nature 1986;319:675e8.

[6] Moretta A, Bottino C, Vitale M, Pende D, Cantoni C,Mingari MC, et al. Activating receptors and coreceptorsinvolved in human natural killer cell-mediated cytolysis.Annu Rev Immunol 2001;19:197e223.

8 M. J.A.J. Huijskens et al.

[7] Cheng M, Chen Y, Xiao W, Sun R, Tian Z. NK cell-basedimmunotherapy for malignant diseases. Cell Mol Immunol2013;10:230e52.

[8] Carbone E, Neri P, Mesuraca M, Fulciniti MT, Otsuki T,Pende D, et al. HLA class I, NKG2D, and natural cytotox-icity receptors regulate multiple myeloma cell recognition bynatural killer cells. Blood 2005;105:251e8.

[9] Shi J, Tricot G, Szmania S, Rosen N, Garg TK,Malaviarachchi PA, et al. Infusion of haplo-identical killerimmunoglobulin-like receptor ligand mismatched NK cellsfor relapsed myeloma in the setting of autologous stem celltransplantation. Br J Haematol 2008;143:641e53.

[10] Siegler U, Kalberer CP, Nowbakht P, Sendelov S, Meyer-Monard S, Wodnar-Filipowicz A. Activated natural killercells from patients with acute myeloid leukemia are cytotoxicagainst autologous leukemic blasts in NOD/SCID mice.Leukemia 2005;19:2215e22.

[11] Alici E, Konstantinidis KV, Sutlu T, Aints A, Gahrton G,Ljunggren HG, et al. Anti-myeloma activity of endogenousand adoptively transferred activated natural killer cells inexperimental multiple myeloma model. Exp Hematol 2007;35:1839e46.

[12] Ishikawa E, Tsuboi K, Saijo K, Harada H, Takano S,Nose T, et al. Autologous natural killer cell therapy for hu-man recurrent malignant glioma. Anticancer Res 2004;24:1861e71.

[13] deMagalhaes-Silverman M, Donnenberg A, Lembersky B,Elder E, Lister J, Rybka W, et al. Posttransplant adoptiveimmunotherapy with activated natural killer cells in patientswith metastatic breast cancer. J Immunother 2000;23:154e60.

[14] Ruggeri L, Capanni M, Mancusi A, Aversa F, Martelli MF,Velardi A. Natural killer cells as a therapeutic tool in mis-matched transplantation. Best Pract Res Clin Haematol2004;17:427e38.

[15] Velardi A, Ruggeri L, Mancusi A, Burchielli E, Perruccio K,Aversa F, et al. Clinical impact of natural killer cell recon-stitution after allogeneic hematopoietic transplantation.Semin Immunopathol 2008;30:489e503.

[16] Frings PW, Van Elssen CH, Wieten L, Matos C,Hupperets PS, Schouten HC, et al. Elimination of thechemotherapy resistant subpopulation of 4T1 mouse breastcancer by haploidentical NK cells cures the vast majority ofmice. Breast Cancer Res Treat 2011;130:773e81.

[17] Krause SW, Gastpar R, Andreesen R, Gross C, Ullrich H,Thonigs G, et al. Treatment of colon and lung cancer pa-tients with ex vivo heat shock protein 70-peptide-activated,autologous natural killer cells: a clinical phase i trial. ClinCancer Res 2004;10:3699e707.

[18] Stern M, Passweg JR, Meyer-Monard S, Esser R, Tonn T,Soerensen J, et al. Pre-emptive immunotherapy with purifiednatural killer cells after haploidentical SCT: a prospectivephase II study in two centers. Bone Marrow Transplant2013;48:433e8.

[19] Sutlu T, Alici E. Natural killer cell-based immunotherapy incancer: current insights and future prospects. J Intern Med2009;266:154e81.

[20] Luhm J, Brand JM, Koritke P, Hoppner M, Kirchner H,Frohn C. Large-scale generation of natural killer lympho-cytes for clinical application. J Hematother Stem Cell Res2002;11:651e7.

[21] McKenna DH Jr, Sumstad D, Bostrom N, Kadidlo DM,Fautsch S, McNearney S, et al. Good manufacturing prac-tices production of natural killer cells for immunotherapy: asix-year single-institution experience. Transfusion 2007;47:520e8.

[22] Meek B, Cloosen S, Borsotti C, Van Elssen CH,Vanderlocht J, Schnijderberg MC, et al. In vitro-differenti-ated T/natural killer-cell progenitors derived from humanCD34þ cells mature in the thymus. Blood 2010;115:261e4.

[23] Huijskens MJ, Walczak M, Koller N, Bried�e J, Senden-Gijsbers BL, Schnijderberg M, et al. Ascorbic acid inducesdevelopment of double-positive T cells from human he-matopoietic stem cells in the absence of stromal cells.J Leukoc Biol 2014;96:1165e75.

[24] Sarkar S, Germeraad WT, Rouschop KM, Steeghs EM, vanGelder M, Bos GM, et al. Hypoxia induced impairment ofNK cell cytotoxicity against multiple myeloma can be over-come by IL-2 activation of the NK cells. PloSone 2013;8:e64835.

[25] Miller JS, Soignier Y, Panoskaltsis-Mortari A,McNearney SA, Yun GH, Fautsch SK, et al. Successfuladoptive transfer and in vivo expansion of human hap-loidentical NK cells in patients with cancer. Blood 2005;105:3051e7.

[26] Ikawa T, Kawamoto H, Fujimoto S, Katsura Y. Commit-ment of common T/Natural killer (NK) progenitors to uni-potent T and NK progenitors in the murine fetal thymusrevealed by a single progenitor assay. J Exp Med 1999;190:1617e26.

[27] Cany J, van der Waart AB, Tordoir M, Franssen GM,Hangalapura BN, de Vries J, et al. Natural killer cellsgenerated from cord blood hematopoietic progenitor cellsefficiently target bone marrow-residing human leukemia cellsin NOD/SCID/IL2Rg(null) mice. PloS one 2013;8:e64384.

[28] Yoon SR, Lee YS, Yang SH, Ahn KH, Lee JH, Kim DY,et al. Generation of donor natural killer cells from CD34(þ)progenitor cells and subsequent infusion after HLA-mismatched allogeneic hematopoietic cell transplantation: afeasibility study. Bone Marrow Transplant 2010;45:1038e46.

[29] Rohwer N, Cramer T. Hypoxia-mediated drug resistance:novel insights on the functional interaction of HIFs and celldeath pathways. Drug Resist Updat 2011;14:191e201.

[30] Harris AL. Hypoxiaea key regulatory factor in tumourgrowth. Nat Rev Cancer 2002;2:38e47.

[31] May JM. The SLC23 family of ascorbate transporters:ensuring that you get and keep your daily dose of vitamin C.Br J Pharmacol 2011;164:1793e801.

Supplementary data

Supplementary data related to this article can befound at http://dx.doi.org/10.1016/j.jcyt.2015.01.004.