HUMAN CELL(Hum Cell) Copyright Q 2004 by The Japan Human Cell Society

Vol. 17 No. 1 Printed in Japan

- Original Article

A GFP-Transfected HFWT Cell Line, GHINK-1, as a Novel Target for Non-RI Activated Natural Killer Cytotoxicity Assay

Hideki HARADA', Kaoru SA.IJO', Isamu ISHIWATA', and Tadao OHNO'

& s h a h An anchoragedependent Wilms' tumor cell line, HFWT, has been found to be extremely susceptible to human natural killer (NK) cells. Here we established a transfectant of HFWT with the green fluorescence protein (GFP) gene, designated GHINK-1 cells, to apply to the activated NK cytotoxicity assay without radioisotope labeling. After beiig coaltured with CD3-CD56' NK cells, GHINK-1 cells released GFP into the medium. The intensity of the fluorescence from the released G F P correlated almost exactly with the radioactivity of a standard "Cr-release assay performed with suspensioncultured K562 cells. Because it does not require separation of the remaining live target cells by cenhifugation, the non-radioisotopic GFP release assay with GHINK-1 cells is a convenient alternative for monitoring human activated NK killing activity.

Hey words : Human activated NK cell, GFF, Qtotoxicily assay, HFYCTT cell line, non-RI [HUMAN CELL 17(1) : 43 - 48,20041

Introduction

The s'Cr-release assay has been generally adopted to assess cytotoxicity of immune effector cells such as cytotoxic T lymphocytes (CTL) or natural killer (NK) cells because of its prominent sensitivity and reproducibility. However, handling of radioactive 5'Cr is complicated and, therefore, non-radioisotopic (RI) cytolytic assays have been proposed as alternatives. For example, europium and fluorescent substrates have been attempted".*'. Ohmori and coworkers3' developed a beta-galactosidase gene-transfected K562 cell line for the non-RI assay of NK and lymphokine

~

1: RIKEN Cell Bank, RIKEN m e Institute of Physical and Chemical Research), Koyadai 3-1-1, Tsukuba Science City, 3050074. Japan.

2: Ishiwata Obstetrics and Gynecology Hospital. Kamimito, Mito. 310-0041. Japan.

activated killer 0 activity. Cytotoxic activity of NK cells has also been detected by flow cytometry') and an ELISPOT assay developed to detect IFN-g production5'. These alternative non-RI methods, however, need extra reagents such as antibodies or fluorescent substrates to quantify the killing activity.

Green fluorescence protein (GFP) is a well-known jelly fishderived protein with strong fluorescence. The GFP gene is widely used a s a reporter gene in experiments of DNA transfection and construction of variant cell lines because GFP is highly hydrophilic. chemically and physically stable, and has low cytotoxicity toward host cells". A non-RI cytotoxicity assay would be easy to performed with GFP gene- transfected target cells because killed target cells would release GFP into the culture supernatant To our knowledge, however, none of the GFP-expressing target cell lines, especially an anchoragedependent human NK-sensitive cell line, has been reported for

43

routine assessment of NK activity. In a previous report, we described a human cell

line, HFWT that was extremely sensitive to activated N K cells. The activated NK activity could be determined by colorimetry of surviving (those firmly attached to the culture surface) HFWT cells after co- culture with activated NK cells and staining with crystal violet (CV)“. However, this assay was not necessarily applicable at an effector/target @/T) ratio higher than 10 because the sticky (resistant to washing off) activated NK cells that were adhering to the culture surface caused considerable determination errors. Thus, we established a GFP gene-transfected HFWT cell line, which was a more convenient target cell line for the non-FU assay of human activated NK cells.

Materials and Methods

1. Cells and Culture Condition The HFWT cell line” was provided by Cell

Medicine Inc., (Tsukuba, Japan). The HFWT and K562 cells were maintained using Ham’s F12 medium supplemented with 15% and 10% (v/v) FBS, respectively. in a humidified 5% C02 incubator at 37C. Human peripheral blood mononuclear cells (PBMC) from healthy volunteers (after obtaining informed consent) were prepared by a ficoll-isopaque gradient using a conventional preparation kit (Lymphoprep tube, Nycomed Pharma AS., Norway). 2. GFP Transfec t ion and Selec t ion of CFP

Transfectant Transfection of the GFP gene was achieved by

exactly following the manufacturer’s instruction manual for the reagent Fugene 6 (Roche, Indianapolis, IN). The plasmid containing the GFP gene (pEGFP-N1, BD Sciences, San Diego, CA) was added to the reagent mixture and incubated at room temperature for 15 min. This mixture was then added to 2x10’ HFWT cells in a 35mm dish and incubated at 37°C for 2 days. A transient transfectant of the GFP gene was further maintained in the selection medium containing G418 (500mg/ p 1) for 7 days. Any remaining cells were routinely cloned by limiting dilution.

3. Expansion of Activated NK, L4K and T-L4K Cells from PBMC PBMC were cultured in RHAMa medium”, which

was prepared for lymphocyte cultivation by the RIKEN Cell Bank and is a mixture of RPMI1640, Ham/F12 and MEMa media (3:l:l in volume), supplemented with 0.02 mg/l a -tocopherol, 0.002 mg/l sodium selenite, 0.004 mg/l linoleic acid, 0.2 mg/l cholesterol, 500 mg/l human serum albumin, 0.6 mg/l 2-mercaptoethanol. 0.6 mg/l ethanolamine, 5 mg/l recombinant human insulin, 5 mg/l transferring. We further supplemented this medium with 5% (v/v) autologous plasma and 200 U/ml of interleukin-2 (IL-2). For studying the expansion of activated NK cells, X106 of PBMC were cocultured with lX10’ HFWT cells in 24 well plates“. LAK and T-LAK cells were cultured without any stimulator cells, although the T-LAK cells were expanded in an anti-CD3 antibody (NU-T3, Nichuei Co. Ltd., Tokyo)coated culture plate’”. Half of the culture medium was changed every other day. For 7-day incubations, these cells were analyzed by flow cytometery before being used in experiments. 4. Antibodies and FACS Analysis

monoclonal antibodies were purchased from Dako Japan Inc. (Tokyo). Suspended lymphocytes (1x106) were washed three times with Dulbecco’s calcium-, magnesium-free PBS [PBS(-)I before being stained for 30 min with the above-mentioned monoclonal antibodies. After washing with PBS(-) supplemented with 4% FBS, and resuspended in the same buffer at a concentration of 1X106 cells/ml, the cells were immediately analyzed by FACsCan flow cytometer (BD Sciences, San Diego, CA). GFP gene-transfected HFWT cells were also analyzed by FACScan flow cytometery .

Anti-CD3/FITC (UCHT1) and CD56/PE (MOC-1)

5. CytotoxicityAssay In the GFP release assay, MEM medium without

phenol red (200p I/well, Nissui Pharmaceuticals, Co., Tokyo, Japan) was used after being supplemented with 5% plasma protein fraction (Bayer Yakuhin, Ltd., Osaka, Japan) for the co-culture of effectors and targets. This was done to avoid influencing the measurement of GFP fluorescence intensity. Effector cells suspended in lOOpl of the abovementioned assay

44

HUMAN CELLVol. 17 No. 1 (2004)

medium were plated onto GHINK-1 cells (3x10' cells/100 p l/well) that had been pre-cultured overnight. Effector-to-target (E/T) ratios were determined by two-fold serial dilutions. Sodium dodecyl sulfate was added at a final concentration of l%(w/v) to control wells that contained target cells without effector cells to completely release the GFP from the cells. After 4hr incubation, 5Opl of the supernatant was sampled and measured by a fluorescence plate reader (Wallac 1420 ARVOsx, Amersham Pharmacia Biotech). Cytotoxic activity was calculated as follows: (A-B)/(C-B) XlOO (%I, where A represents fluorescence intensity of the medium from

assay as well as in the GFP-release assay. After 4hr incubation, the radioactivity of the supernatant was measured with a scintillation counter, and the cytotoxicity calculated. Both assays were performed in triplicated wells. 6. Statistical analysis

Linear regression analysis was performed to evaluate the correlation between the "Cr-release assay and the GFP-release assay using activated NK cells as effector cells, and 5'Cr-labeled K562 cells and GFP- expressing GHINK-1 cells as target cells, respectively. Sixteen paired sets of data were obtained from four independent experiments.

the well with effectors and targets; B, fluorescence intensity of the medium from the well to determine

Results and Discussion

spontaneous release of GFP 6.e., E/T ratio is 0); and C. fluorescence intensity of the medium from the well to which sodium dodecyl sulfate was added. In the 51Cr- release assay, 1X106 of K562 cells were labeled with lOOml(50pCi/ pl) of Naz5'Cr04 (Amersham-Pharmacia. UK) for lhr. After washing the target cells, they were adjusted to l X l O ' cells/well and used in the "Cr-release

We have established a clone, designated GHINK-1, derived from GFP genetransfected HFWT cells. The intensity of GFP fluorescence in GHINK-1 cells was stable for at least two years following multiple passages. The selection reagent, G418, was not required for maintenance of GFP expression. Although an EGFP-transfected K562 cell l ine has been

l A K (no stimulator) .................... 7-LAK (amti-CD3 mAb) ---. ............. -

I

CD3

I zu Fig. 1: Sensitivity of GHINK-1 cells to activated NK cells. (a) Effector cells. i.e.. (1) NK (Hm. (2) LAK (no stimulator), and (3) T-IAK (antiCD3 mAb) were induced by culturing PBMC (1) onto feeder HFWT cells in medium containing IL2. (2) in the absence of stimulator feeder cells but containing IL2 in the medium. and (3) on immobilized antiCD3 mAb with IL2 in the medium, respectively. The percentages of CD3-CD56* cells in the effectors were 88.9%. 12.1% and 2.1%. respectively. (b) Sensitivity of GHINK-1 cells to activated NK (A), IAK(.) and T- LAK (+) effectors. Cytotoxicity was calculated by measuring GFP released from GHINK-1 cells when they were killed by effector cells, as described in the Material and Methods. Each point is a mean of observations from three wells and vertical bars show the standard deviation. Results from a representative experiment of three are shown.

b

IUU -

Y

c

0 2 + 6 0

Em reti-

45

established for the target using a flow cytometory- based assay measuring NK activity"', the transfected cell line needed G418 to maintain EGFP expression. The original cell line, K562. has been widely utilized as the standard in the "Cr-release assay to determine NK activity.

To elucidate the susceptibility of GHINK-1 cells to effector cells, we induced the cell populations NK, w(

and T-LAK from human PBMC. The percentages of CD3-CD56' cells in these lymphocyte populations were 88.9%, 12.1% and 2.1%, respectively (Fig. la). GHINK-1 cells showed high, medium and low susceptibility to NK, LAK and T-LAK cell populations, respectively (Fig. lb); i. e., the susceptibility was proportional to the percentages of CD3-CD56' NK cells.

Then, we defined "activated NK cells" as the NK cells in PBMC cultured in medium containing IL2, and

a r5 r

X

C

loo 80 i 0

a a

I ,' 0 a

a

0,' 9

+ * a -

# r .t

on the HFWT feeder cells. When the susceptibility of GHINK-1 cells was directly compared with that of K562 cells and to the activated-NK cells, there was no critical difference between data obtained from either the "Cr- or the GFP- release assay (Fig. 2). Cytotoxicity was almost equally detected by both assay systems in PBMC cultured for 4 days (Fig. 2% week cytotoxicity) and 7 days Fig. 2b. strong cytotoxicity) on the feeder HFWT cells grown in medium containing IL2. Actual fluorescence intensities observed in four experiments are shown in Table 1. Spontaneous release of GFP from GHINK-1 was similar to that of "Cr. The correlation coefficient between the 51Cr- and GFP-release assays was 0.9436 (p<O.OOOl) (Fig. 2 ~ ) .

The advantages of the GFP-release assay as compared with the standard "Cr (or other fluorescent reagent)-release assay are as follows: there is (1) no need for dangerous RI ("'Cr), expensive monoclonal

eo r b 6 v 60

0 ' I I

Q 2 ratim

Fig. 2: Comparison of "Cr- and GFP-release assays, Effector cells were 112-stimulated and feeder HFWT-stimulated/activated NK cells cultured for 4 days (a) and 7 days (b), respectively. Similar results were obtained from two other experiments. X and 0 indicate results from the "Cr-release assay with target K562 cells, and the GFP-release assay using target GHINK-1 cells, respectively. Each point is a mean of three wells and the vertical bar shows the standard deviation. (c) Correlation between the "Cr- and GFP-release assays. Each point is a mean of 16 paired- observations and each symbol corresponds to the results from different experiments. The correlation coefficient was calculated from four independent experiments. (b) Sensitivity of GHINK-1 cells to activated NK (A), LAK(H) and T-LAK (+) effectors. Cytotoxicity was calculated by measuring GFP released from GHINK-1 cells when they were killed by effector cells, as described in the Material and Methods. Each point is a mean of observations from three wells and vertical bars show the standard deviation. Results from a representative experiment of three are shown.

46

HUMAN CELLVol. 17 No. 1 (2004)

antibodies, fluorescent reagents and/or substrates, (2)

no reduction of GFP concentration in GHINK-1 cells as observed following passage for two years, and no requirement of G418 selection reagent to maintain fluorescence intensity, (3) no need to use centrifugation for separation of the suspension cultured target cells from the supernatant, and (4) low spontaneous release of measurement markers from the target cells.

The results described here strongly suggest that GHINK-1 cells are very useful for assay of cytotoxic activity of human activated NK cells. Although we have tried to determine the sensitivity of GHINK-1 cells to "resting" NK cells prepared from peripheral blood of adult subjects, we could not observe a high sensitivity of target GHINK-1 cells to the NK cells. The present method will also be useful to determine the activity of human activated NK cells.

Acknowledgement

This work was partly supported by the Special Coordination Fund of the Science and Technology Agency of Japan.

References

Volgmann T, Klein-Sbucheier A and Mohr H: A fluorescence-based assay for quantitation of lymphokine-activated kiIler cell activity. J. Immunol. Methods. 119: 4551,1989. Blomberg K, Hautala R, Lovgren J, et al.: Tirne- resolved fluorometric assay for natural killer activity using target cells labeled with a florescence enhancing ligand. J. Immunol. Methods. 193: 199206,1996. Ohmori H, Ikeda H, Tanigawa T, et al.: Enzyme release assay of human NK cell activity using beta- galactosidaseexpressing K562 target cell line. J. Immunol. Methods. 164: 131-135.1993. Radosevic K, Ganitsen HS. Van Graft M, et al.: A simple and sensitive flow cytometric assay for the determination of the cytotoxic activity of human natural killer cells. J. Immunol. Methods. 135: 81- 89,1990. Derby EG, Reddy V, Nelson EL, et al.: Correlation of human CD56' cell cytotoxicity and I F N g production. Cytokine. 13: 85-90,2001. Chalfie M, Tu Y, Euskirchen G, e t al.: Green fluorescent protein as a marker for gene

Table 1: Fluorescence intensity of supernatants from GHINK-1 cells co-cultured with effector cells EffeCtorlTarget Ratio

Effectorflarget Ratio v 0.5 1 2 4 8

Ave. +SD Ave. +SD Ave. *SD Ave. fSD Ave. *SD

Exp.1 6919.7 98.4 - - 9106.0 788.1 13122.7 1528.5 17627.7 676.2 23264.3 3120.2 Exp.2 6279.0 283.8 - - 20299.0 1077.4 28015.0 4375.9 32373.0 3167.5 39116.0 2513.2 Exp.3 6806.7 313.8 - - 8483.3 155.9 11233.3 1035.8 13446.0 669.6 20493.0 420.1 Exp.4 6694.3 755.1 7524.3 53.1 8328.3 658.4 10167.7 770.9 13581.7 939.6 - -

Spontaneous Maxgb Blank' '' release (%I '

Ave. *SD Ave. *SD GFP 51Cr

Exp.1 37471.3 5468.3 4566.7 119.3 7.2 6.9 Exp.2 40831.0 2581.6 4130.0 388.4 5.9 8.0

Exp.3 35160.3 288.7 4613.7 234.1 8.6 13.7 Exp.4 53353.0 3691.2 4278.7 42.0 5.2 7.7

'GHINK-1 cells were lysed with 1% SDS. "Blank indicates fluorescence intensity of the medium. "'Spontaneous release of GFP was calculated as follows: (ac)/(bc) X 100. The raw data of "Cr-release assay were not shown.

47

expression. Science. 263: 802-805,1994. 7) Harada H, Watanabe S , Saijo K, et al.: Selective

expansion of human natural killer cells from peripheral blood mononuclear cells cultured on an anchoragedependent Wilms tumor cell line. Jpn. J. Cancer Res. 9 3 313-319.2002.

8) Ishiwata I, Ono I, Ishiwata C, et al.: Carcinoembryonic proteins produced by Wilms’ tumor cells in uiw and in v iva Exp. Pathol. 41: 1- 9,1991.

9) Liu SQ, Saijo K, Todoroki K, et al.: Induction of human autologous cytotoxic T lymphocytes on formalin-fixed and paraffin-embedded tumour

sections. Nature Med. 1: 267-271,1995. 10) Sekine T, Shiraiwa H, Yamazaki T, et al.: A feasible

method for expansion of peripheral blood lymphocytes by culture with immobilized anti-CD3 monoclonal antibody and interleukin-2 for use in adoptive immunotherapy of cancer patients. Biomed. phannacother. 4 7 73-78,1993.

11) Kantakamalakul W, Jaroenpool J and Pattanapanyasat I<: A novel enhanced green fluorescent protein (EGFP)-K562 flowcytometric method for measuring natural killer (NK) cell cytotoxic activity. J. Immunol. Methods. 272: 189 197,2003.

Received 2004.3.3. Accepted 2004.3.20. Addmas correspondence and reprint requests to : Dr. T. Ohno. RIKEN Cell Bank, RIKEN m e Institute of Physical and Chemical Research), Koyadai 3-1-1, Tsukuba Science City, 3050074. Japan. Tek +81-2!&363611 Fax: +81-2983&9130 Email address: tad-ohnoOrtc.riken.go.jp