calcium oscillations in human t and natural killer cells

0022-1 767/93/1507-2620$02.00/0

Copyright 0 1993 by The American Association of Immunologists The Journal of Immunology Vol 150. 2620-2633, No. 7, April 1, 1993

Printed in U.S.A.

Calcium Oscillations in Human T and Natural Killer Cells Depend upon Membrane Potential and Calcium Influx' -

Stephen D. Hess,2* Marga O~rtgiesen,~' and Michael D. Cahalan4*

*Department of Physiology and Biophysics, University of California, Irvine, CA 9271 7, and +Research Institute of Toxicology, University of Utrecht, NL-3508 TD Utrecht, The Netherlands

ABSTRACT. With the use of single-cell digital imaging of the fluorescent Ca2+ indicator dye fura-2 we investigated

Ca2+ signaling in human T lymphocytes and NK cells during activation by a variety of stimuli. A low percentage

of resting T cells or T cell blasts displayed oscillations in cytosolic Ca2+ when stimulated with the mitogenic lectin

PHAor by the addition of OKT3 mAb followed by a secondary cross-linking antibody. Lymphokine-activated T killer

cells were more responsive than resting cells. A comparison of PHA, cross-linked anti-CD3, and a heteroconjugate

mAb showed that at least 20% of the cells from these T cell preparations oscillated. Addition of PHA or cross-linked

anti-CDl6 caused NK cells to oscillate. In contrast, thapsigargin, a microsomal ATPase blocker, resulted in a

relatively uniform, slowly rising and sustained Ca2+ response in all cell types studied. The maintenance of both thapsigargin- and receptor-induced responses required Ca2+ influx driven by a negative membrane potential.

Because Ca2+ oscillations occurred in response to stimuli which mimic the normal activation of lymphocytes, and

inasmuch as the percentage of oscillating cells increases with state of activation, these oscillations may play an

important role in mitogenic activation. journal of /rnrnuno/ogy, 1993, 150: 2620.

C alcium plays a crucial role in the activation of T lymphocytes. Cross-linking surface proteins with mitogenic lectins or mAb to the CD3/TCR com-

plex elicits a series of intracellular biochemical events (reviewed in Ref. l) , including the activation of tyrosine kinases (for reviews see Refs. 2 4 ) and PLC' (5-7). Both inositol- 1,4,5-triphosphate (IP,) and diacylglycerol (8) are generated, resulting in the activation of protein kinase C and release of Ca2+ from internal stores (8-10). In addition, Ca2+ influx from the external medium is triggered (1 1-1 3) through voltage-independent Ca2+ channels (14, 15). Both

Received for publication July 7, 1992. Accepted for publication December 31, 1992.

The costs of publication of this article were defrayed in part by the payment of

accordance with 18 U.S.C. Section 1734 solely to indicate this fact. page charges. This article must therefore be hereby marked advertisement in

and GM41514 and by a grant from Pfizer Inr. ' This work was supported by National Institutes of Health Grants NS14609

Current address: SIBIA, INC., 505 Coast Boulevard South, Suite 300, La Jolla, CA 92037.

Supported by a fellowship from the Royal Netherlands Academy of Arts and Sciences.

Ca2+ release and influx contribute to the rise in cytosolic Ca2+ concentration ([Ca2'Ii) (15-17).

The early [Ca2+Ii transient is necessary for a variety of later events in T cell activation, and is temporally correlated with the transcriptional activation of so-called immediate genes including c-fos and c-myc (18), as well as subsequent activation of c-myb, IL-2 and its receptor (reviewed in Ref. 19). Calcium may stimulate the activity of PLC (20, 21) (but see also Ref. 22), in addition to activating protein kinase C and calcium-calmodulin-dependent protein kinases. Blocking Ca2+ influx prevents normal activation of T cells (23), underscoring the requirement for this ion in the signaling cascade.

To date most studies of [Ca2+Ji transients in T cells have employed intracellular Ca2+ indicator dyes such as indo- 1,

4Address correspondence and reprint requests to Michael D. Cahalan, Department of Phys~ology and Biophysics, University of California, Irvine. CA 92717.

Abbreviations used in this paper: PLC, phospholipase C; ICa2+I,, intracellular calcium concentration; FFT, fast Fourier transform; GAM, goat anti-mouse; T-LAK, lymphokine-activated killer T cell.

2620

Journal of Immunology 262 1

quin-2 (24, 25), or fura-2 (26) in conjunction with flow cytometry or cuvette measurements. These methods sample the average response of the population and do not provide details of single-cell responses. When fura-2 measurements are coupled with digital fluorescence ratio-imaging methods, [@+Ii signals in individual cells are resolved (15, 16, 27). With the use of the human Jurkat leukemic T cell line, Lewis and Cahalan (1 5) found that PHA-stimulated Jurkat cells display oscillations in [Ca2+Ii that are dependent on Ca2+ influx and a negative membrane potential and are correlated with the current carried through a mitogen-gated Ca'+-selective conductance. Similar [Ca2+Ii oscillations have been observed in a human CD4+ T cell clone after contact with APC (27). Such oscillations may be important in signal transduction (22, 28). Our goal therefore was to extend the approach of single-cell [Ca2+Ii imaging to describe the temporal and spatial patterning of [Ca2+Ii transients in normal human T cells.

In the present study, we used digital fluorescence ratio imaging of fura-2 to describe the dynamics of CaZt signals in individual resting and activated human T cells in response to addition of PHA or mAb to appropriate surface proteins. We report that [Ca2+Ii oscillations occur in freshly isolated T cells, peripheral blast cells, and T-LAK cells stimulated through the CD3/TCR complex, and in CD3- NK cells stimulated with PHA or cross-linked anti-CD16 mAb. These findings suggest that [Ca'+], oscillations may play an important role in the activation of a variety of human lymphocyte populations.

Materials and Methods Human lymphocyte preparations

Lymphocytes were purified from peripheral blood obtained from consenting healthy adult donors. Venous blood was collected in heparinized tubes and diluted 50% with RPMI 1640 (GIBCO/BRL, Gaithersburg, MD) containing 25 mM HEPES. This suspension was then centrifuged at 400 X g through a Ficoll-Paque density gradient (Pharmacia LKB, Piscataway, NJ) for 30 min at room temperature (25-27°C). The interface was removed, washed three times with RPMI containing 20% FCS (JR Scientific, Woodland, CA), and applied to a sterile nylon wool column pre-equilibrated with RPMI/20% FCS. After 45 min at 37"C, the column was eluted with prewarmed RPMI/20% FCS, and the eluted cells were washed three times with this medium. FACS analysis of cells isolated from four healthy donors showed that 76.9% of the cells were CD3+ T cells, 7.6% were CD19+ B cells, and 0.6% were CD14+ monocytes. The cells were used within 6 h for experiments with freshly isolated cells or cultured at an initial density of 0.3 X IO5 cells/ml in RPMI/lO% FCS in a humidified incubator at 37°C with 5% C02.

A T cell blast preparation was obtained by culturing fresh

T cells in RPMI/lO% FCS with 10 pg/ml PHA (Difco Lab- oratories, Detroit, MI). After 72 to 96 h in culture, the medium was changed to fresh RPMI/10% FCS with PHA. These cells were usually used after 4 to 8 days in culture.

NK cells were isolated from fresh blood by negative- selection panning (29) of cells eluted from a nylon wool column as described above. Plastic petri plates were coated overnight at 4°C with 10 pg/ml GAM IgG (Caltag, South San Francisco, CA). The plates were washed twice with PBS, allowed to stand at room temperature for 30 min in PBS containing 0.2% BSA, and then washed once again with PBS. Approximately 20 X 10'T cells were suspended in 280 pl of RPMI/10% FCS along with 40 pl each of anti-CD3, CD4, and CD8 mAb (Becton Dickinson, San Jose, CA) for 45 min on ice. The excess antibodies were washed off five times with ice-cold RPMI/10% FCS. The cells were then resuspended at 5 to 10 X 10' cells/ml, and 4 ml of this suspension were allowed to settle on a precoated plate for 45 min at 40"C, with a gentle swirl after 25 min. The plate was swirled again, and the cells in suspension were removed by pipetting, washed twice with PBS/BSA, and either stained with anti-CD16 and anti-CD3 mAb as above or resuspended in RPMIFCS and used in experiments within 6 h. These cells were 75 to 85% CD16+ and CD3-.

T-LAK cells were obtained from PBMC by culturing for 48 to 72 h with 200 U/ml human rIL-2 (Hoffmann LaRoche, Nutley, NJ) and 400 ng/ml PHA (Burroughs Wellcome, Research Triangle Park, NC) followed by growth in rIL-2 alone by using methods described elsewhere (30).

Reagents and antibodies

PHA-P (Difco or Burroughs Wellcome) was stored at -20°C in aliquots at 10 mg/ml, then thawed and diluted immediately before use in experiments. Unconjugated OKT3 mAb (Ortho Diagnostics, Raritan, NJ) and conjugated anti-CD3, anti-CD4, anti-CD8, and anti-CD16 mAb (Becton Dickinson) were kept refrigerated until just before use. Goat anti-mouse IgG (Organon Teknika, West Chester, PA) was used to cross-link primary mAb and kept cold until just before use. Heteroconjugate antibodies anti-CD3/CD4 and CD3/CD8 were a kind gift of Dr. Jeffrey Ledbetter (Bristol Myers Squibb, Seattle, WA), and were stored as aliquots containing 1 mg/ml BSA at -70°C until just before use.

Digital video-imaging

All experiments were performed on an IM-35 inverted microscope (Zeiss, Oberkochen, Germany). Cells were loaded with fura-2 AM (Molecular Probes, Eugene, OR) at 1 pM in RPMI/lO% FCS for 25 min at room temperature in the dark. Longer loading times or loading at higher tempera- tures resulted in loading of dye into intracellular organelles

2622 Ca2+ OSCILLATIONS IN HUMAN LYMPHOCYTES

that did not exhibit the changes in [Ca2+Ji seen in the bulk cytoplasm. The cells were washed thrice with RPMWCS and used within 4 h. Separate whole-cell patch-clamp experiments with pipettes containing fura-2 free acid showed that the concentration of fura-2 AM in cells following the above loading procedure was 50 to 100 pM. Lymphocytes were then plated at 3 to 4 X lo6 cells/ml on poly-o-lysine- coated glass coverslips, and rinsed to remove nonadherent cells with mammalian Ringer solution containing (in mM): 160 NaCl, 4.5 KCI, 2 CaCl,, 1 MgC12, 5 HEPES-NaOH (pH 7.4), 11 glucose.

The experiments were controlled, recorded, and analyzed with the use of VideoProbe software and hardware (ETM Systems, Irvine, CA) with a 386 computer (AST Research, Irvine, CA). Cells were alternately illuminated at 350 and 385 nm by a xenon light source (Zeiss) passing through excitation filters (Omega Optical, Brattleboro, VT) housed in a motorized filter wheel and electronic shutter (Sutter Instruments, Novato, CA). The light beam was then reflected by a 405-nm dichroic mirror and passed through a 63X oil immersion objective (Zeiss) to illuminate the cells. Emitted light at greater than 510 nm was then measured with a SIT camera (Hamamatsu Photonics, Bridge- water, NJ).

Typically, [Ca2+Ji was estimated every 10 s by dividing the background-subtracted 350- and 385-nm raw fluorescence images pixel-by-pixel, and diplaying the result with a pseudocolor scale. Data were analyzed in more detail offline by averaging the nonzero 3501385 ratio values within an approximately 20 X 20 pixel rectangle positioned over each cell in the field (usually 90-150 cells with the 63X objective). Drift of the cells during the 20-min experiments was usually not a problem; deliberately posi- tioning rectangles at the edges of cells did not affect the analysis until only a very small area of the cytoplasm (<lo%) remained under the rectangle. Cells that drifted substantially during the experiment were not included in the analysis. To obtain information on the period of the oscillations, the [Ca2+Ii responses of cells that produced sustained oscillations (5 or more peaks in [Ca2+li during the experiment) were further analyzed using a FFT algorithm implemented in a commercial spreadsheettanalysis pro- gram (Igor v. 1.2, WaveMetrics, Inc., Lake Oswego, OR).

In some cases, cells were phenotyped at the end of the experiment by the addition of 20 pl of anti-CD4 and anti- CD8 mAb conjugated to FITC or phycoerythrin directly to cells on the microscope stage. The excess antibodies were washed off after 8 min, and the images obtained by exciting phycoerythrin and FITC separately were later superimposed over the fura-2 images to score the phenotype of cells. Whenever a GAM secondary antibody had been used as a cross-linker, excess antibody was first adsorbed by the addition of excess mouse IgG or mouse serum.

Fura-2 calibration

Fura-2 signals were calibrated with an intact-cell method. Estimates of Rmin were obtained by exposing loaded cells to Ringer solution containing 10 mM EGTA, and 1 pM ionomycin, but no added CaZf. After collection of these images, the Ringer solution was changed to one containing 10 mM Ca2+ and 1 pM ionomycin, and the cells were rinsed twice before obtaining the R,,, estimates. Typical values obtained were 0.7 for Rmin, 10 for R,,,, and 9 for the scale factor. Using a Kd of 350 nM obtained for fura-2 inside cells (31), the resting [Ca2+Ji of cells was usually 60 to 120 nM. This calibration technique has the advantage of measuring dye responses in situ, although the necessary condition that no Ca2+ be bound to the dye at Rmin is difficult to achieve in a living cell. Therefore, these calibrations are reasonable estimates of [Ca2+Ii.

Results [Ca2+li signaling induced by a mitogenic lectin

Video-imaging experiments have previously demonstrated that Jurkat cells often exhibit oscillations in [Ca2+Ii after PHA stimulation (15). To extend these results to normal human T cells, we examined [Ca2+Ii responses at the single-cell level in resting T cells and in a population of T-LAK cells. Freshly isolated human T cells responded to PHA stimulation with an increase in [Ca2+Ji after approximately 60 s at 25°C. In the experiment shown in Figure 1, 29% of the cells demonstrated [Ca2+Ji oscillations, but the pattern of the [Ca2+Ii signal was quite variable between individual cells. Some cells displayed no response or only a transient increase in [Ca2'Ii (Fig. 1, A, and B ) , whereas others showed a continual increase in [Ca2+Ii after activation (Fig. IC). Figure 1, D and E, shows two examples of sustained oscillations; the peak [Ca2+Ii in the cells varied from 400 to 850 nM. When the population average is examined (Fig. l F ) , the lack of response in some cells, variable latency to the first response, and asynchrony of individual cell [Ca2+Ii oscillations together produce a smooth rise in [Ca2+li, similar to that seen in flow cytometry or cuvette experiments.

To determine if [Ca2+Ji oscillations induced by PHA are a general property of T cell subsets, the [Ca2+Ii transients in a preparation of T-LAK cells prepared by culturing blast cells in PHA and IL-2 were also examined, as illustrated in Figure 2. In this experiment, 33% of the cells showed no response to 10 pglml PHA (Fig. 2 A ) , whereas others showed a large (1 pM) rise in [Ca2+Ii (Fig. 2 B ) . The [Ca2'Ii level of 26% of the cells in this experiment oscillated (Fig. 2, C, D, and E ) . We conclude that [Ca2+Ii oscillations can be induced in normal T cells by PHA. The results also point to the possibility that [Ca2+Ii oscillations occur more fre- quently in cells that have been previously activated. This

Journal of Immunology

4 A

2623

B -t q 4 w

.- +- a zoo 2

FIGURE 1. Responses of human resting T cells to addition of 10 pg/rnl PHA (arrow). (A-E) intracellular [Ca2+Ii in single cells measured with fura-2. Here and in the following figures the examples were chosen to approximate the proportion of cells in the total population that displayed similar responses. ( F ) Average response of the cells in the experiment.

- I I I I I I

O 4PPlmP&) 800 1000

I C

N +- c : b Y 9 200

100

0 200

conclusion is substantiated in experiments with mAb stimulation (described below).

[Ca2+li signaling induced by mAb to surface receptors: correlation with state of activation

Although mitogenic lectins provide a convenient method for activating T cells, their binding specificity is not clearly defined (1). Recent results have focused attention on the role of the 5-chain of the CD3RCR complex in signal transduction (1,32-34). In addition to being expressed in T cells, the 5-chain is present in NK cells in association with CD16 (35). We have examined [Ca2'Ii signals in these cell types in response to more specific stimuli believed to be physiologically relevant. We have also compared [Ca2+Ii signaling in cell types at various stages of their activation to determine whether the tendency to produce [Ca2+Ii oscillations increases with the state of activation. Resting T cells. Figure 3 illustrates responses in resting T

4 . 600 , 800 1000 %me (s)

O 2oo 4%rnesops) 800 1000

I D

-I-

500 t F

l W E ! X z 0 0 200 400 Timtys) 800 1000

cells to soluble OKT3 mAb followed by GAM H and L chain secondary cross-linking antibody. Addition of soluble OKT3 alone had little effect on [Ca2+Ii in individual cells or in the average of the 120 cells in the field (Fig. 3F). Upon addition of the secondary cross-linking GAM, several cells displayed an abrupt increase in [Ca2+li which approached 700 nM (Fig. 3, B-D); 16% of the cells oscillated (see Fig. 3E for an example). The apparent decrease in [Ca2+Ii following addition of the GAM mAb (Fig. 3F) is an occasional artifact caused by a change in background fluorescence at one excitation wavelength.

Table I summarizes the experiments on fresh T cells. For PHA stimulation at 25"C, the percentage of cells showing a [Ca2+Ii increase exhibited a weak dose dependence, although the percentage of cells that oscillated appeared to peak at 25 pg/ml PHA. Soluble OKT3 mAb applied alone was not very effective in eliciting [Ca2+Ii transients. Cross- linking the OKT3 mAb with a GAM secondary antibody

2624 CaLf OSCILLATIONS IN H U M A N LYMPHOCYTES

1200 - 1000 -

800 - Y

N +&- 600

- PHA 0 400 ([I

-

1

A

0 200 400 600 800 1000 1200 Time (s)

lZoo r C

loo0 t FIGURE 2. Addition of 10 pg/ml PHA (arrow) triggers [Ca2+], increases in individual T-LAK cells (B-E) and in the average response ( F ) .

v z +i;

0 N ([I

0 200 400 600 800 1000 1200 Time (s)

T v

.- +- m N

0

400 600 800 1000 1200 Time (s)

0 200

1200 r B

z +- ([I

Y

N

0

0 200 400 600 800 1000 1200 Time (s)

F

800

600

400k 200 0 1 0 200 400 600 800 1000 1200

Time (s)

was approximately as effective as 10 pg/ml PHAin eliciting [&*+Ii transients. Fifty to sixty percent of the cells in this mixed population responded to PHA or cross-linking CD3 with a [Ca2+Ii increase. Activated T cells. To correlate the occurrence of [Ca2+Ii oscillations with the state of activation of T cells, we examined [CazfIi transients in a primary blast cell culture prepared by growing human T cells in medium containing 10 pg/ml PHA. These cells exhibit increased numbers of voltage-gated Kt channels (36) and Ca*+-dependent K t channels (37). Both channel types might be expected to influence [Ca2+Ii signaling in normal human T cells, as has been shown in Jurkat T cells (1 5,38). When soluble OKT3 mAb was applied alone to these cells, fewer than 1 % of the cells oscillated and 46% showed an increase in [Ca2+Ii (Table 11). Cross-linking CD3 on these cells with a GAM secondary antibody caused 20% of the cells to oscillate, and 77% of the cells showed an increase in [Ca2+Ii (Table 11). Comparison of Tables I and I1 shows that although cross-

linking CD3 was apparently more effective in eliciting both increases in [Ca2+Ii and oscillations in blast cells than in fresh cells, the differences were not statistically significant (for oscillationsp = 0.446, t = -0.781; for [Ca2+Ii increase

Because CD4 or CD8, in conjunction with CD3, is nor- mally engaged in a complex by peptide bound to MHC during Ag presentation, we examined the ability of two heteroconjugate mAb to stimulate blast cells. Table I1 shows that the CD3/CD4 heteroconjugate applied at 1 pg/ml caused 16% of blast cells to oscillate, and 54% showed an increase in [Ca2+Ii. An experiment in which the conjugate was particularly effective is shown in Figure 4. After addition of the CD3/CD4 conjugate, 82% of the cells in this experiment responded with an abrupt increase in [Ca2+Ii which approached 700 nM in some cells (Fig. 4, A-E). The types of responses we observed included one or more large peaks in [Ca2+Ii that progressively decayed (Fig. 4, A and B), a large peak followed by several smaller

p = 0.095, t -1.78).


6wl 500

A 6ool 500

2625

B

FIGURE 3. Addition of mAb to CD3 (OKT3, 2.5 pg/ml) elicits little [Ca2+li change in individual human resting T cells (A-E) or in the average response ( F ) . Ad- dition of a cross-linking secondary antibody (GAM, 1 .O mg/ml) elicits transient responses in some cells (6 and C) and oscillations in others ( D and E ) .

N PLi m

2 2w 100

0 0 200 400 Time 600 (s80c 1000 1200

-t C

500

600 I E

lo0l I \ nucysh- 2i)O 4hO 6b 860 1dOO 1 i O O

Time (s)

0 200 400 600 800 1000 1200 Time (s)

F

v

6-300 - .- O R 3

m oy200 -

100 - 100

i 260 400 600 800 1000 1200 Time (s)

0 200 400 600 800 1000 1200 Time (s)

Table I htracellular Ca2+ responses in resting human T cells

ICa2+l, ICa2+I,

Stimulus Total No. oscillations, increase, of cells percent percent

of cells" of cells"

Table I I htracellular Caz+ transients in blast cells

ICaz+I, ICaZ+l, Total No. oscillations, increase, of cells percent percent

of cellsd of cells"

Stimulus

PHA 10 pg/rnl 5 359 25 p g h l

17.3e4.5 46.6e10.7 2 191 32.5k2.9 54.8k1.6

50 pg/rnl 7 790 21.9e3.4 63.3t3.0 OKT3, 2.5 p g h l 7 417 G A M after OKT3 6 366

4.2r3.9 28.3k9.4 12.8e3.6 50.9e8.1

OKT3, 2.5 pg/rnl 6 617 0.6k0.4 46.6t11.9 G A M 1/10, after OKT3 9 939 20.8t7.9 77.1e10.7

CD3/CDB, 1 pg/rnl CD3/CD4, 1 p g h l 14 1045 16.7t3.7 54.1k6.0

8 711 2.2k0.9 16.0t6.6

Mean t SEM.

Mean 2 S E M

oscillations (Fig. 4, C and D ) , or oscillations of relatively constant amplitude that were superimposed on a slowly declining baseline (Fig. 4E). The amplitude of these oscillations could approach 250 nM (Fig. 4E) and could out- last our normal recording duration (1 200 s). The average of the 5 2 cells in the field also shows a sharp increase in [Ca2+Ii (Fig. 4F).

Table I1 also shows that the CD3/CD8 heteroconjugate was much less effective than the CD3/CD4 conjugate in eliciting [Ca2+Ii transients. To explore possible differences in subset responsiveness, as has been reported for fresh human T cells activated with anti-CD3 mAb (39), we phenotyped cells at the end of experiments by adding anti-CD4 mAb conjugated to phycoerythrin and anti-CDS mAb conjugated to FITC. We found that 94% of the CD4+ blast cells showed a [Ca2+li transient in response to cross-linking


FIGURE 4. The addition of a heteroconjugate of CD3 and CD4 mAb elicits large increases in [Ca2+li in individual human blast cells (A-E) and in the average response (F). Several cells ( B - E ) displayed [Ca2+Ii oscillations.

500 - -

600 - 500 -

gm- v

._ CD3/CD4

m

100 -

I B

0 260 4kO 6bO 8 W ldoo l;00 Time (s)

0 1 1 1 1 1 1 1 0 xx) 4%m62(s,800 1000 1200 0 200 400 600 800 1000 1200

Time (s)

0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200 Time (s) Time (s)

CD3, whereas 97% of the CD8+ cells responded. This suggests that the CD3/CD8 heteroconjugate antibody is not as effective as the CD3/CD4 antibody in triggering [Ca2+Ii transients, since the CD8+ blast cells were competent to respond to CD3 cross-linking. In addition, we found that in 99% of the cells the phenotype, as determined with anti- CD4 and anti-CD8 mAb, corresponded exactly to the heteroconjugate antibody that elicited a response in an individual cell. This correlation demonstrates appropriate specificity of the heteroconjugate antibodies and allowed us to use both heteroconjugates in a single experiment to stimulate both helperhnducer and cytotoxic/suppressor subsets. T-LAK. PHA was an effective stimulus in eliciting [Ca2+Ii transients in T-LAK cells (Fig. 2, Table 111). Although either purified OKT3 or mAb obtained from a hybridoma supernatant applied alone caused [Ca2+Ii transients and even some oscillations, both were much more effective when cross-linked with a GAM secondary antibody (Table 111).

We examined the subset responsiveness of T-LAK cells stimulated by PHA or cross-linked CD3. We found no significant difference between CD4+ and CD8+ subsets in either the fraction of responsive cells or the fraction of cells that exhibited [Ca2'Ii oscillations when stimulated with PHA, the OKT3 hybridoma supernatant alone, or the supernatant followed by a GAM secondary antibody (not shown). The increased tendency of T-LAK cells to respond with [Ca2+Ii oscillations, compared with resting or PHA- activated T cells, suggests that culturing these cells in IL-2 indirectly potentiates [Ca2+Ii signaling. We did not exam- ine this possibility directly, but it has been reported that IL-2 does not by itself stimulate phosphatidylinositol turnover (40).

Because 20 to 60% of these T-LAK cells are CDI 6+ (30), and a l-chain similar to that associated with CD3 is coupled to CD16 in NK cells (33, we determined whether CD16 could transduce [Ca2'Ii signals in T-LAK cells by adding an anti-CD16 mAb to the cells. As shown in Table 111, CDI 6

Journal of Immunology 2627

Table 111 Ca" transients in T-LAK cells

ICa'*l, lCaL+I, Total No. oscillations, increase, " of cells percent percent

of cells.' of cells"

Stimulus

PHA 5 Pdmi 10 868 64.4k4.1 86.9k2.8 10 pdml 5 628 48.8+11.9 78.4e6.1

OKT3 hybridoma 6 500 10.0e6.2 29.5k11.3 supernatant

G A M after OKT3 7 608 62.2k7.9 88.8k4.3

OKT3, 2.5 ps/ml supernatant

3 350 25.4213.6 62.7k28.3 G A M 1/10, after OKT3 3 340 49.0218.6 84.8k12.1 Anti-CD16 2 132 0 1.5k0.1 G A M after antiLCD1 6 2 132 20.9k1.7 61.3e32.1

Mean t SEM.

mAb alone was not effective in eliciting [Ca"], transients. After addition of a secondary GAM antibody, 20% of the cells oscillated, and 61% showed an increase in [Ca2'Ii. This result suggests that the CD16 molecule in T-LAK cells is able to transduce signals which lead to elevations in [ca2+li. N K cells. A population of NK cells was obtained from human peripheral blood by negative-selection panning with anti-CD3, CD4, and CD8 mAb. These cells were approximately 80% CD-16+. Figure 5 shows that application of an anti-CD16 mAb (before time 0) followed by a GAM secondary cross-linking antibody elicited [Ca2+Ii transients in these cells. In this experiment, 73% of the cells showed an increase in [Ca2+Ii (Fig. 5 , A and B ) and 47% of the cells oscillated (Fig. 5 , C, D, and E ) with the peak [Ca2+Ii ex- ceeding 500 nM in some cells. The average of the 193 cells in the field is shown in Figure 5F.

When a GAM secondary antibody was used to cross-link CD16 on the surface of NK cells, 5 1 2 7% of the cells oscillated, and 72 -t 11% of the cells showed an increase in [Ca2"Ii (mean 5 SEM, n = 5 experiments, 644 cells). PHA at 25 yg/ml was also an effective stimulus, as 58 5 7% of the cells oscillated and 70 t 8% had increased [Ca2+Ii ( n = 2 experiments, 174 cells). These results show that NK cells display [Ca2+Ii oscillations when activated either with PHA or with anti-CD16 mAb.

Analysis of [Ca2+li oscillations

We examined the relationship between the latency of the first response and the period of the subsequent oscillations to test whether these two processes might share a common mechanism. To measure the period of the oscillations, we analyzed the Fourier transform of 640-s segments of the data. Figure 6 shows examples of the [Ca2+Ii responses and the complex conjugate of the FFT in two T cell blasts responding to the CD3/CD4 heteroconjugate antibody. In this

analysis, the low frequency peak at a period of 640 s cor- responds to a slowly relaxing component of the signal. At higher frequencies, one or more clear peaks of variable amplitude occurred at periods corresponding to approximately 90 to 160 s. For six blasts cells, the mean period was 143 t 21 s, whereas the latency to the first response after antibody application was 57 8 s. For both T cell blasts (Fig. 6E) and for resting T cells (data not shown), no correlation exists between the latency and the period (? = 0.006).

We also compared the latency and timing of [Ca2+Ii transients from those cells in which [Ca2+Ii oscillated when stimulated by PHAor by mAb. In resting T cells, the latency to the first [Ca2+Ii response was 67 -t 9 s (mean t SEM of 6 cells) when cross-linked OKT3 was the stimulus and 63 t 12 s when PHA was used. Fourier analysis yielded similar peaks corresponding to periods of 137 -+ 18 s for OKT3-stimulated cells, and 164 -+ 20 s for PHA-treated cells. The similarity in the latencies and periods suggests that both stimuli produce a rise in [Ca2+Ii through the same biochemical pathway. For comparison, [Ca2+Ii oscillations in Jurkat cells and an Ag-specific T cell clone P28D have been reported to exhibit periods of 92 2 12 s and 116 28 s, respectively (15, 27).

Temperature dependence

Additional experiments on resting T cells were performed at 37°C to evaluate the temperature dependence of the response latency and oscillation period. Raising the temperature decreased the latency after cross-linking CD3 with GAM antibody from 67 to 24.0 2 5.1 s, but the mean period of the oscillations (121 t 14 s, n = 6) and the percentage of oscillating cells were not significantly affected. Together these data suggest that the steps coupling receptor stimulation to the initial rise in [Ca2+li are more temperature dependent than are the mechanisms that control the subsequent [Ca" l i oscillations.

Calcium influx is reduced by depolarization of the membrane potential

The results described above demonstrate a heterogeneous response pattern in individual resting and activated T cells, as well as in NK and T-LAK cells. Oscillations in [Ca2+Ii can occur in each of these cell types in response to appropriate stimuli, but the tendency to oscillate increases in activated cells. Because the initial large increase in [Cazfli occurs in the absence of added external Ca2+, release of Ca2+ from internal stores is thought to underlie the initial transient increase in [Ca2+Ii in lymphocytes (9, 10). In Jurkat T cells and P28D cells examined with single-cell Ca2+ imaging, release from internal stores has been shown to produce the initial rise in [Ca2+Ii, but sustained responses and oscillations require Ca2+ entry from the outside ( 1 5 , 16,

2628 CaL+ OSCILLATIONS IN HUMAN LYMPHOCYTES

FIGURE 5. [Ca'+], responses in human NK cells mediated through CD16. After a 1 O-min incubation with a CD16 mAb (before time O), addition of GAM secondary antibody (arrow) caused [Gal'], increases in individual cells (A-E) and in the population average ( F ) . In this experiment, 47% of the cells oscillated (C, D, and E) .

A 600 B

0 200 400 Time 600 (sy 1000 1200 0 200 400 600 800 1000 1200 Time (s)

0 200 400 Time 600 (sy 1000 1200 0 200 400 Time 600 (sy 1000 1200

6oo r E 6oo r F

0 200 400 Time 600 (sy 1000 1200 0 200 400 600 800 1000 1200 Time (s)

27). To determine the relative contribution of Ca2+ from internal stores versus that obtained from influx in peripheral blood T cells, we removed external Ca2+ during the oscillations. As seen in Figure 7, A and B, for blast cells, changing the bath solution from one containing 2.0 mM Ca2+ to a solution containing 10 mM EGTA, with no added Ca2+, caused the oscillations to cease immediately and the [Ca2+Ii to decrease to below that seen at rest. Upon return to normal Ringer solution, the [Ca2+Ii level showed an abrupt transient increase similar to the initial response seen in other cells (e.g., Fig. 1E). Because this concentration of EGTA empties thapsigargin-sensitive internal stores within 90 s in these cells (see below), the large increase in [Ca2+Ii upon return to normal Ringer solution may reflect an increased plasma membrane permeability to Ca2+ due to depletion of intracellular pools (4144). After the initial [Ca2+Ii increase, the oscillations in some cells resumed after addition of normal (Ca2+-containing) Ringer solution. These results show that influx of external Ca2+ sustains the

oscillations, and suggest that the pathway by which Ca2+ enters the cells, perhaps a mitogen-gated Ca2+ channel similar to one described in Jurkat cells (13, remains activated after cross-linking of CD3. Similar results were obtained for both resting and activated T cells.

A negative membrane potential facilitates Ca2+ entry in many nonexcitable cells including mast cells (45) and Jur- kat T lymphocytes (1 5). We therefore examined the role of membrane potential in the maintenance of the oscillations in normal T cells. Soon after exposure of the cells to 160 mM Kt Ringer solution, which should depolarize human T cells to 0 mV, the [Ca2+Ii declined in each of the cell types investigated here, as illustrated for activated T cells in Fig- ure 7, C and D. The overshoot in [Ca2+Ii upon return to normal Ringer is similar to that seen after removal of external Ca2+ (Fig. 7, A and B ) . Comparable results were seen with resting T cells, NK and T-LAK cells (not shown). These results suggest that the normal membrane potential facilitates Ca2+ influx in human lymphocytes.

journal of Immunology

0-

'0° [ c

0 1 0 1200

Time, s 220

A A A A A

180 - Period, s A A A A A A

140 - A A A

A 100 - A A

A A

60 I 0 " " ' 40 80 110 160

Latency, s

FIGURE 6. Analysis of the periodicity of the [Ca2+li oscillations in human blast cells. Examples of [Ca2+li responses ( A and C) and their resulting major periods ( B and D). To obtain the periods, a 640-s segment was analyzed with a FFT. The peaks in the complex conjugate of the transform indicate major periods of the oscillations. Latencies were measured from the time of stimulation to the initial upstroke of the [CaL+I, signal. A plot ( E ) of the latency to the first [Ca2+], increase versus the period of the subsequent oscillations illustrates a lack of correlation. Each point is from one of 23 sustained oscillators from the experiment shown in Figure 3.

Thapsigargin induces a smooth rise in [Ca2+li, sustained by the membrane potential and inhibited by the K+-channel blocker, charybdotoxin

Thapsigargin, a specific blocker of the Ca2+ ATPase which pumps Ca2+ ions into intracellular stores, is a useful tool in the analysis of [Ca2+li regulation (46,47). In contrast to the highly variable [Ca'+Ii responses obtained by stimulation through surface receptors, application of thapsigargin results in a smoothly graded, relatively uniform, and long lasting rise of [@'Ii in individual T cells (Fig. 8). The rise of [Ca2+Ii after thapsigargin treatment, at an average rate of 2 to 3 nM/sec, suggests an ongoing "leak" and re-uptake of Ca'+ ions moving between intracellular stores and cytoplasm in resting cells. Rapid turnover of the intracellular stores is also indicated by the rapid (<90 s) depletion of the pools after incubating cells in Ringer solution containing EGTA with no added Ca2+ (data not shown). The slow decline of [Ca2+], after treatment with thapsigargin may

2629

reflect either increased activity of the Ca2+ pump at the surface membrane or inactivation of Ca2+ influx. Figure 8 also shows that the rise in [Ca2+li induced by thapsigargin is inhibited by membrane depolarization, in a manner similar to results shown in Figure 7 for T cells stimulated through the TCR complex.

Both voltage-gated and calcium-activated Kt channels may participate in the generation of [Ca2'], signals (48,49) by maintaining the membrane potential and thereby providing the electrical driving force for Ca'+ entry. Because the Ca2+ signals induced by receptor stimulation are highly variable from cell to cell and are spiky in individual cells, we have used the relatively consistent thapsigargin-induced [Ca2'], signal to address the role of K + channels in con- trolling Ca2+ influx. Charybdotoxin, a peptide scorpion toxin, blocks both voltage- and calcium-activated K+ channels in human T cells with an affinity in the 1 to 5 nM range (37, 50, 51). Figure 9 demonstrates that 100 nM charybdotoxin lowers [Ca'+], following thapsigargin stimulation. These results are consistent with the idea that thapsigargin treatment, by emptying intracellular stores, results in calcium influx down an electrochemical gradient sustained, in part, by charybdotoxin-sensitive K' channels.

Discussion

With video-imaging of [Ca2'Ii, we have characterized the responses of normal human resting and activated T lymphocytes, T-LAK cells, and NK cells to a variety of activation stimuli at the single-cell level. Our results indicate that human T lymphocytes display a temporally heterogeneous pattern of [Ca2'Ii responses when stimulated with PHA or mAb. Individual cells within each cell type are capable of displaying oscillations in [Ca2+Ii. We found that the proportion of I-LAK cells that display both increases in [Ca2+Ii and [Ca2'Ii oscillations when activated with crosslinked anti-CD3 was greater than for fresh cells. In the T cell preparations, we compared the [Ca2'li responses to various stimuli and found that PHA, cross-linked anti-CD3, and heteroconjugate CD3/CD4 mAb each produced oscillations in at least 20% of the responding cells. Human NK cells also displayed [Ca2+li increases and oscillations when triggered with PHA or specifically through CD16. These oscillations are dependent on the presence of external Ca2+ and are reduced by membrane depolarization. Because [Ca2'li oscillations can occur in all of the subsets we examined in response to stimuli designed to mimic the normal parameters of in vivo activation, and the expression of the oscillations increases with the activational state of the cells, the oscillations in [Ca2+Ii may well play an important role in early activation events in T cells and NK cells.

Our results on human T cells are generally consistent with previous investigations with cell lines. Similar response amplitudes of [Ca2+li into the low micromolar


A C 1400 1 Ringer 1400 1 Ringer

FIGURE 7. [Ca2+li oscillations are dependent upon a membrane potential-sensitive Ca2+ influx. Oscillations triggered by cross-linking CD3 with GAM antibody (arrows) terminated abruptly upon changing the external medium to one with no added Ca2+ and 10 mM EGTA (arrow). [Ca2'l, rapidly increased upon return to normal Ringer solution (ar row) in the example cell ( A ) and in the average (B ) . (C and D) When the Ringer solution was changed to one containing 160 mM K+ (arrow), the ICaZ+li of individual cells (C) and the average (D) quickly decreased. After return to normal Ringer solution, the oscillations resumed in some cells, and an overshoot in [CaZ+li was seen in the cell and the average response.

160 mM K+ z ._ r 2

R1

0' 0 1

B D

0 1 0 1 0

range and oscillation periods of 90 to 160 s are observed in fresh T cells stimulated with PHA or cross-linking antibodies (Figs. 1 and 3), Jurkat cells stimulated with PHA (15), and a T cell clone in response to contact with APC (27). Our results on the effects of temperature are different from those of Donnadieu et al. (27), who reported a sub- stantial temperature dependence of the periodicity of APC- induced oscillations in P28D cells. In addition, response latencies in human T cells are 20 to 40% of those reported for P28D cells in response to soluble antibody stimulation (27). It is unclear whether these differences reflect the properties of the cell line or the mode of stimulation. Mechanism of [Ca2+Ii oscillations. Cross-linking of surface receptors associated with the 5-chain, either the CD3/ TCR complex or CD16, results in [Ca2+Ii signals that are often oscillatory in lymphoid cell lines. Recent evidence using chimeric proteins that link the intracellular domain of the {-chain of CD3 to the extracellular and transmembrane regions of CD8 suggests that the ("chain functionally cou- ples the CD3/TCR complex in Jurkat T cells (33). Anti- bodies which cross-link the CD3/TCR complex or which bring together CD4 and CD3 were especially effective stimuli in eliciting [Ca2+Ii responses and oscillations (Figs. 3 and 4). Since CD16 is associated with the 5-chain in CD3- NK cells (33, we investigated whether stimulation through CD16 could cause [Ca2+Ii oscillations. We found that cross-linking CD16 on both NK and T-LAK cells elicited [Ca2+Ii transients and oscillations (Fig. 5). These findings suggest that CD16, possibly through the [-chain, interacts with intracellular enzymes necessary for [Ca2+Ii signaling in these cell types.

The pattern of [Ca2'Ii signaling stimulated through surface receptors is very different from that induced by thapsi-

600 1200 1800 0 600 1200 1800

Time, s Time, s

gargin, an inhibitor of the intracellular Ca-ATPase (Fig. 8). Thapsigargin treatment results in a relatively smooth and uniform rise in [Ca2+Ii, compared with the fluctuating and highly variable [Ca2+Ii signals when monoclonal antibodies or lectins are used as ligands. Stimulation via surface receptors appears to be necessary for eliciting oscillatory [Ca2+li responses.

The oscillations observed in human lymphocytes typically followed an abrupt increase in [Ca2+Ii due to release of Ca2+ from internal stores, and were dependent upon both influx of Ca2+ from the external medium and the maintenance of a negative membrane potential. Because Ca2+ influx in T lymphocytes is inhibited by depolarization (Fig. 7), it has been proposed that potassium channels, by maintaining a negative membrane potential, play a crucial role in T cell activation (reviewed in Refs. 48 and 49). Several lines of evidence indicate that voltage-gated type n channels control the resting membrane potential of T cells (reviewed in Refs. 48 and 49). After activation, the subsequent rise in [Ca2+li would open Ca2+-dependent K' channels, hyperpolarize the membrane potential, and potentiate Ca2+ entry (52,53). Recent patch-clamp experiments have demonstrated a dramatic increase in the expression of Ca2+- dependent K+ channels in PHA-activated blast cells relative to resting cells (37). The additional hyperpolarizing outward current provided by these channels may provide an element of positive feedback to augment Ca2+ influx. K-channel blockers, including charybdotoxin, have been reported to inhibit T cell activation (reviewed in Refs. 48 and 49) (5 1,54) (but see Refs. 55 and 56). Several of these blockers, including charybdotoxin, inhibit both voltage- gated and calcium-activated K+ channels (37, 50, 51). Ca2+ influx stimulated by thapsigargin is similarly inhib-


Reslmg Human T Cells Resting Human T Cells

2631

100 nM CTX

1 uM Tg Ringer

[Ca2+]i, nM

l 6 O 0 1 0

2500 1

"*ii---" 0

1600 1

0 7 16001

Average I /--" 0 1 0 500 1000 1500 2000

Time, s

FIGURE 8. Thapsigargin-induced rise in [Ca2+li. The top five traces show the response of individual cells; the average response is shown in the bottom trace. Addition of 1 pM thapsigargin produced a prolonged elevation in [Ca2+], in a majority of the cells. Addition of 40 mM K+-containing Ringer solution (arrow) caused a transient decrease in [Ca2'l, which was reversed by restoring the bath solution to normal Ringer solution.

ited by depolarization with 40 mM K+ or by charybdotoxin (Figs. 8 and 9); both treatments reduce [Ca2+li to a plateau level between resting and maximal [Ca2+li levels. These observations suggest that K+ channels play a crucial role by modulating Ca2+ influx in human T cells and therefore the subsequent Ca2+-dependent events in activation and proliferation.

The lack of correlation between the latency to the first [Ca2+Ii increase and the period of the subsequent oscillations in T cells suggests that different biochemical pathways regulate these events (Fig. 6) (see also Refs. 15 and 57). Lewis and Cahalan (15) concluded that a mitogen- gated Ca2+i conductance, which waxed and waned pre- ceding [Ca2+Ii changes in Jurkat T cells, induces the oscillations and that [Ca2+Ii regulates both the Ca2+ influx pathway and the oscillations. Some of the possible feedback loops include the modulation by [Ca2+Ii of the Ca2+ influx pathway (15) or the ability of IP3 to release internal stores (58) , the existence of multiple internal stores of Ca2+ with different [Ca2+Ii sensitivities (58), or Ca2+ stimulation of PLC activity (21, 59).

U I I I I

2500 1

25001

0 w 2500L 0 0 400 800 1200 1600

Time, s FIGURE 9. Inhibition of the thapsigargin-induced [Ca2+], signal by charybdotoxin (100 nM) added at the indicated times. The lower trace shows the average response. A further decrease in [Ca2+], was observed upon raising [K+], to 160 mM and was reversed by changing the solution to Ringer solution.

Are [Ca2+li oscillations physiologically important? Our results have demonstrated the ubiquity of [Ca2+Ii oscillations in lymphoid cells in response to a variety of in vitro stimuli. Similar oscillations in [Ca2+Ii have been observed in T cell clones that are dependent on contact with APC (27) (S. D. Hess, D. Choquet, and M. D. Cahalan, manuscript in preparation). These data suggest that the oscillations observed in human T cells may occur during antigen presentation in vivo.

Despite speculations about the physiological role of [Ca2+Ii oscillations in a variety of cell types, few corre- lations between the amplitude or frequency of the oscillations and [Ca2+Ii-dependent events have been demonstrated (22, 28, 60). It is possible that signaling using

2632

[Ca2+], oscillations may represent an energy savings compared to using changes in the basal level of [Ca2'li (28) and may increase the fidelity of the signal (20). In our experiments, the rise in [Ca2+], appeared to be uniform through- out the cell, including the nucleus. Using the single-cell approach it should be possible to correlate functional con- sequences of different types of [Ca2+Ii signals on IL-2 se- cretion or early gene activation events in T cells, using a reporter gene (61). Our characterization of [Ca2+], signals in individual human lymphocytes provides an important framework for future attempts to correlate these events with single-cell functional assays of [Ca'+li-dependent events in the activation cascade.

Acknowledgments

We thank J . Ledbetter for providing heteroconjugate antibodies, E. K. Ininns and G. Granger for help in preparing T-LAK cells, P. Steinbach for programming assistance, K. G. Chandy, R. Davis, and J. Ledbetter for helpful comments on the manuscript, and L. Forrest and A. Nguyen for valuable technical assistance.

References I .

2.

3.

4.

5.

6.

7.

8.

9.

10.

Klausner, R. D., and L. E. Samelson. 1991. T cell ,antigen receptor activation pathways: The tyrosine kinase connection. Cell 64:875. Alexander, D. R., and D. A. Cantrell. 1989. Kinases and phos- phatases in T-cell activation. Zmmunol. Today 10:200. Mustelin, T., and A. Altman. 1989. Do CD4 and CD8 control T-cell activation via a specific tyrosine protein kinase? Irnrnunol. Today 10:189. Rudd, C. E. 1990. CD4, CD8 and the TCR-CD3 complex: a novel class of protein-tyrosine kinase receptor. Zmmunol. To- day 11:990. Mustelin, T., K. M. Coggeshall, N. Isakov, and A. Altman. 1990. T cell antigen receptor-mediated activation of phospholipase C requires tyrosine phosphorylation. Science 247: 1584. Augustine, J. A,, J. P. Secrist, J. K. Daniels, P. J. Leibson, and R. T. Abraham. 199 1. Signal transduction through the T cell antigen receptor. Activation of phospholipase C through a G protein-independent coupling mechanism. J . Immunol. 146: 2889. Graves, J. D., and D. A. Cantrell. 1991. An analysis of the role of guanine nucleotide binding proteins in antigen receptor/ CD3 antigen coupling to phospholipase C. J. Irnmunol. 146: 2102. Berridge, M. J. 1987. Inositol triphosphate and diacylglycerol: two interacting second messengers. Annu. Rev. Biochem. 56: 159. June, C. H., J. A. Ledbetter, P. S. Rabinovitch, P. J. Martin, P. G. Beatty, and J. A. Hansen. 1986. Distinct patterns of transmembrane calcium flux and intracellular calcium mobilization after differentiation antigen cluster 2 (E rosette receptor) or 3 (T3) stimulation of human lymphocytes. J. Clin. Invest. 77:1224. Ledbetter, J. A., C. H. June, L. S. Grosmaire, and P. S. Rabino- vitch. 1987. Crosslinking of surface antigens causes mobi-

C a L + OSCILLATIONS IN HUMAN LYMPHOCYTES

lization of intracellular ionized calcium in T lymphocytes. Proc. Natl. Acud. Sci. USA 84:1384.

1 1. Freedman, M. H., M. C. Raff, and B. Gomperts. 1975. In- duction of increased calcium uptake in mouse T lymphocytes by concanavalin A and its modulation by cyclic nucleotides. Nature 255:378.

12. Deutsch, C., and M. A. Price. 1982. Cell calcium in human peripheral blood lymphocytes and the effect of mitogen. Bio- chim. Biophgs. Acta 687:211.

13. Gelfand, E. W., R. K. Cheung, and S. Grinstein. 1984. Role of membrane potential in the regulation of lectin-induced calcium uptake. J. Cell. Physiol. 121:533.

14. Kuno, M., and P. Gardner. 1987. Ion channels activated by inositol 1,4,5-trisphosphate in plasma membrane of human T-lymphocytes. Nature 326:301.

15. Lewis, R. S., and M. D. Cahalan. 1989. Mitogen-induced oscillations of cytosolic Ca2+ and transmembrane Ca'+ current in human leukemic T cells. Cell Regul. 1:99.

16. Lewis, R. S., and M. D. Cahalan. 1990. Ion channels and calcium signalling in single mitogen-stimulated T lymphocytes. UCLA Syrnp. Mol . Cell. Biol. 113:125.

17. Weiss, M. J., J. F. Daley, J. C. Hodgdon, and E. L. Reinherz. 1984. Calcium dependency of antigen-specific (T3-Ti) and alternative (TI 1) pathways of human T-cell activation. Proc. Nutl. Acad. Sci. USA 81:6836.

18. Reed, J. C., J. D. Alpers, P. C. Nowell, and R. G. Hoover. 1986. Sequential expression of protooncogenes during lectin- stimulated mitogenesis of normal human lymphocytes. Proc. Nutl. Acad. Sci. USA 83:3982.

19. Crabtree, G. R. 1989. Contingent genetic regulatory events in T lymphocyte activation. Science 243:355.

20. Meyer, T., and L. Stryer. 1988. Molecular model for receptor- stimulated calcium spiking. Proc. Natl. Acad. Sci. USA 85: 5051.

21. Eberhard, D. A,, and R. W. Holz. 1988. Intracellular Ca2+ activates phospholipase C. Trends. Neurosci. 11:517.

22. Rapp, P. E., and M. J. Berridge. 1981. The control of trans- epithelial potential oscillations in the salivary gland of Cal- liphora erythrocephala. J. Exp. Biol. 93:119.

23. Whitney, R. B., and R. M. Sutherland. 1972. Requirement for calcium ions in lymphocyte transformation stimulated by phytohemagglutinin. J . Cell. Physiol. 80:329.

24. Tsien, R. Y., T. Pozzan, and T. J. Rink. 1982. Calcium home- ostasis in intact lymphocytes: cytoplasmic free calcium mon- itored with a new, intracellularly trapped fluorescent indicator. J . Cell Biol. 94:325.

25. Tsien, R. Y., T. Pozzan, and T. J. Rink. 1982. T-cell mitogens cause early changes in cytoplasmic free Ca2+ and membrane potential in lymphocytes. Nature 295:68.

26. Grynkiewicz, G., M. Poenie, and R. Y. Tsien. 1985. A new generation of Ca2+ indicators with greatly improved fluorescence properties. J. Bid . Chem. 260:3440.

27. Donnadieu, E., D. Cefai, Y. P. Tan, G. Paresys, G. Bismuth, and A. Trautmann. 1992. Imaging early steps of human T cell activation by antigen-presenting cells. J. Irnrnunol. 148:2643.

28. Berridge, M. J., P. H. Cobbold, and K. S. R. Cuthbertson. 1988. Spatial and temporal aspects of cell signalling. Phil. Trans. R. Soc. Lond. B 320:325.

29. Mason, D. W., W. J. Penhale, and J. D. Sedgwick. 1987. Prep- aration of lymphocyte subsets. In Lymphocytes. G.G.B. Klaus, ed. IRL Press, Oxford. p. 48.

Journal of Immunology 2633

30. Yamamoto, R. S., J. Coss, B. Vayuvegula, S. Gupta, Y. Beamer, S. Jacques, E. W. B. Jeffes 111, W. E. Carson 111, J. Jakowatz, and G. A. Granger. 1991. Generation of stimulated, lymphokine-activated T killer (T-LAK) cells from the peripheral blood of normal donors and adult patients with re- current glioblastoma. J. Immunol. Methods 137:225.

3 I. Negulescu, P. A,, and T. E. Machen. 1991. Intracellular ion activities and membrane transport in parietal cells measured with fluorescent dyes. Methods Enzymol. 192:58.

32. Sussman, J. J., Bonifacio, J. S., Lippincott-Schwartz, J., Weissman, A. M., Saito, T., Klausner, R. D., and Ashwell, J. D. 1988. Failure to synthesize the T cell CD3-5 chain: structure and function of a partial T cell receptor complex. Cell 52:85.

33. Irving, B. A,, and A. Weiss. 1991. The cytoplasmic domain of the T cell receptor 5 chain is sufficient to couple to receptor- associated signal transduction pathways. Cell 64.891.

34. Romeo, C., M. Amiot, and B. Seed. 1992. Sequence require- ments for induction of cytolysis by the T cell antigenPC receptor ( chain. Cell 68:889.

35. Lanier, L. L., G. Yu, and J. H. Phillips. 1989. Co-association of CD3 (with a receptor (CD16) for IgG Fc on human natural killer cells. Nature 342:803.

36. Deutsch, C., D. Krause, and S. C. Lee. 1986. Voltage-gated potassium conductance in human T lymphocytes stimulated with phorbol ester. J. Physiol. 372:405.

37. Grissmer, S., and Cahalan, M. D. 1991. Ca’+-activated K+ channels in resting and activated human peripheral T lymphocytes. Biophys. J. 59:213a.

38. Grissmer, S., R. S. Lewis, and M. D. Cahalan. 1992. Ca2+- activated K+ channels in human leukemic T cells. J. Gen. Physiol. 99:63.

39. Rabinovitch, P. S., C. H. June, A. Grossman, and J. A. Led- better. 1986. Heterogeneity among T cells in intracellular free calcium responses after mitogen stimulation with PHA or anti-CD3. Simultaneous use of indo-1 and immunofluores- cence with flow cytometry. J. Immunol. 137:952.

40. Mills, G . B., D. J. Stewart, A. Mellors, and E. W. Gelfand. 1986. Interleukin 2 does not induce phosphatidylinositol hy- drolysis in activated T cells. J. Immunol. 136:3019.

4 I . Putney, J. W. 1986. A model for receptor-regulated calcium entry. Cell Calcium 7:l.

42. Sarkadi, B., A. Tordai, L. Homolya, 0. Scharff, and G. Gar- dos. 1991. Calcium influx and intracellular calcium release in anti-CD3 antibody-stimulated and thapsigargin-treated human T lymphoblasts. J. Membr. Biol. 123:9.

43. Gouy, H., D. Cefai, S. B. Christensen, P. Debre, and G. Bis- muth. 1990. Ca2+ influx in human T lymphocytes is induced independently of inositol phosphate production by mobilization of intracellular Ca2+ stores. A study with the Ca2+ endoplasmic reticulum-ATPase inhibitor thapsigargin. Eur. J. Immunol. 20:2269.

44. Mason, M. J., M. P. Mahaut-Smith, and S. Grinstein. 1991. The role of intracellular Ca2+ in the regulation of the plasma membrane Ca2+ permeability of unstimulated rat lymphocytes. J. Biol. Chem. 266:10872.

45. Penner, R., G. Matthews, and E. Neher. 1988. Regulation of calcium influx by second messengers in rat mast cells. Nature 334:499.

46. Thastrup, O., P. J . Cullen, B. K. Drobak, M. R. Hanley, and A. P. Dawson. 1990. Thapsigargin, a tumor promoter, dis- charges intracel!ular Ca2+ stores by specific inhibition of the endoplasmic reticulum Ca2+-ATPase. Proc. Nutl. Acud. Sci. USA 87:2466.

47. Putney, J. W. 1991. The capacitative model for receptor-

48. Lewis, R. S., and M. D. Cahalan. 1990. Ion channels and signal transduction in lymphocytes. Annu. Rev. Physiol. 52: 415.

49. Cahalan. M. D., K. G. Chandy, and S. Grissmer. 1991. Po- tassium channels in development, activation, and disease in T lymphocytes. Curc Top. Membr. 39:357.

50. Sands, S. B., R. S. Lewis, and M. D. Cahalan. 1989. Charyb- dotoxin blocks voltage-gated K+ channels in human and mu- rine T lymphocytes. J . Gen. Physiol. 93:1061.

51. Price, M., S. C. Lee, and C. Deutsch. 1989. Charybdotoxin inhibits proliferation and interleukin 2 production in human peripheral blood lymphocytes. Proc. Nutl. Acad. Sci. USA 86:10171.

52. Grinstein, S., and S. J. Dixon. 1989. Ion transport, membrane potential, and cytoplasmic pH in lymphocytes: changes during activation. Physiol. Rev. 69:417.

53. Mahaut-Smith, M. P., and L. C. Schlichter. 1989. Ca2+- activated K+ channel> in human B lymphocytes and rat thy- mocytes. J. Physiol. 415:69.

54. Slaughter, R. S., J. L. Shevell, J. P. Felix, C. S. Lin, N. H. Sigal, and G. J. Kaczorowski. 199 I . Inhibition by toxins of charybdotoxin binding to the voltage-gated potassium channel of lymphocytes: correlation with block of activation of human peripheral T-lymphocytes. Biophys. J. 39:213u.

55. Gelfand, E. W., and R. Or. 1991. Charybdotoxin-sensitive, Ca’+-dependent membrane potential changes are not in- volved in human T or B cell activation and proliferation. J. Immunol. 147:3452.

56. Attali, B., G. Romey, E. Honore, A. Schmid-Alliana, M-G. Mattel, F. Lesage, P. Ricardi, J. Barhanin, and M. Lazdunski. 1992. Cloning, functional expression, and regulation of two K+ channels in human T lymphocytes. J. Biol. Chem. 267: 8650.

57. Niklinska, B. B., H. Yamada, J. J. O’Shea, C. H. June, and J. D. Ashwell. 1992. Tyrosine kinase-regulated and inositol phosphate-independent Ca2+ elevation and mobilization in T cells. J . Biol. Chem. 267:7154.

58. Bezprozvanny, I., J. Watras, and B. E. Ehrlich. 1991. Bell- shaped calcium-response curves of Ins( I ,4,5)P; and calcium- gated channels from endoplasmic reticulum of cerebellum. Nature 351:751.

59. Harootunian, A. T., J. P. Y. Kao, S. Paranjape, and R. Y. Tsien. 1991. Generation of calcium oscillations in fibroblasts by positive feedback between calcium and IP,. Science 251:75.

60. Woods, N. M., K. S. R. Cuthbertson, and P. H. Cobbold. 1987. Agonist-induced oscillations in cytoplasmic free calcium concentration in single rat hepatocytes. Cell Calcium 8:79.

61. Karttunen, J., and N. Shastri. 1991. Measurement of ligand- induced activation in single viable T cells using the lac2 reporter gene. Proc. Natl. Acad. Sci. USA 88:3972.

activated calcium entry. Adv. Pharmucol. 22:251.

calcium oscillations in human t and natural killer cells

Documents