Ammonium chloride exposure inhibits cytokine-mediated eosinophil survival

JOURNALOF IMMUNOLOGICAL METHOOS ELSEVIER

Journal of Immunological Methods 168 (1994) 187-196

Ammonium chloride exposure inhibits cytokine-mediated eosinophil survival Mikako Ide, Deborah Weiler, Hirohito Kita *, Gerald J. Gleich Department of Immunology and Division of Allergic Diseases and Medicine, Mayo Clinic and Mayo Foundation, Rochester, MN 55905, USA

(Received 10 September 1993, accepted 27 September 1993)

Abstract

To study human eosinophils, their efficient purification from peripheral blood is crucial. Although a number of purification procedures, including discontinuous Percoll and metrizamide density gradient centrifugation, have been used, it has been difficult to isolate eosinophils from normal donors with consistently high yields and purities. Recently, a new isolation technique called magnetic cell separation system (MACS) was reported. To evaluate this procedure, we isolated eosinophils from human peripheral blood using either MACS or the standard discontinuous Percoll density methods, and compared cellular viability, morphology, and response to degranulation stimuli. MACS gave a higher yield of eosinophils than Percoll density centrifugation; for example, 6.6 + 1.1 × 106 eosinophils were isolated from 20 ml of blood by MACS compared to 6.4 + 2.4 × 106 from 120 ml by Percoll density gradient. Further, the purity of eosinophils isolated by MACS was 97.1 + 0.5% (-~+ SEM) compared to 77.8 + 2.9% with Percoll. As part of the MACS protocol, erythrocytes are lysed with either 155 mM ammonium chloride or hypotonic lysis. With 155 mM ammonium chloride treatment, the eosinophils showed a striking reduction in cytokine mediated survival due to interleukin (IL)-3, IL-5 and granulocyte-macrophage colony-stimulating factor (GM-CSF), marked morphologic abnormalities and a reduced degranulation response. With hypotonic lysis, no differences were observed in survival and morphology between eosinophils purified by MACS and Percoll methods; the degranulation responses to stimuli were essentially the same between the two methods. Taken together, these observations suggest that the exposure of eosinophils to 155 mM ammonium chloride results in cellular damage. Therefore, MACS with hypotonic lysis is a useful technique to isolate eosinophils for biological study. Key words: Eosinophil; Purification; Magnetic cell separation; CD16

* Corresponding author. At: Department of Immunology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA. Tel.: (507) 284-4492; Fax: (507) 284-1086. Supported in part by grants from the National Institutes of Health, AI 15231 and AI 31155, and by the Mayo Foundation. Abbreviations: DCS, defined calf serum; EDN, eosinophilderived neurotoxin; GM-CSF, granulocyte-macrophage colony-stimulating factor; IL, interleukin; MACS, magnetic cell separation system; OVA, ovalbumin; PBS, phosphate buffered saline; slgA, secretory IgA; RIA, radioimmunoassay. Elsevier Science B.V. SSDI 0022-1759(93)E0255-G

1. Introduction

I n r e c e n t years, t h e r e c o g n i t i o n o f t h e e o s i n o p h i l ' s i n v o l v e m e n t in a v a r i e t y o f d i s e a s e s a s s o c i a t e d with e o s i n o p h i l i a , such as a s t h m a a n d o t h e r allergic c o n d i t i o n s , has i n c r e a s e d i n t e r e s t in e o s i n o p h i l s ( G l e i c h et al., 1992). E o s i n o p h i l i a has also o c c u r r e d as a side effect o f clinical t h e r a p y with c y t o k i n e s ( V a n H a e l s t P i s a n i et al., 1991).

188

M. Ide et al. /Journal of lmmunological Methods 168 (1994) 187-196

Because eosinophils usually constitute less than 5% of granulocytes, in vitro studies of eosinophils depend on their successful purification from whole blood. Various methods have been employed to purify eosinophils from human peripheral blood, including density gradients of metrizamide or Percoll, often coupled with other procedures to separate the neutrophils from the eosinophils because these two granulocytes have overlapping densities (Hansel et al., 1990). Recently, Hansel and colleagues developed a procedure combining a single Percoll density gradient centrifugation to isolate granulocytes and erythrocytes, followed by 155 mM ammonium chloride lysis of the erythrocytes. Neutrophils are then bound to micromagnetic beads coupled to CD16 antibody and removed with a magnetic cell separation system (MACS, Becton-Dickinson, San Jose, CA) (Hansel et al., 1991). The method is highly effective compared to density gradient centrifugation; commonly, eosinophil preparations are 99.5% pure with a yield of 81.7% from normal donors (Hansel et al., 1991). This remarkable efficiency means that eosinophils can be routinely isolated from normal healthy individuals, as well as from patients with atopic or other eosinophil associated diseases. To employ this method in our studies of eosinophils, we compared MACS purified cells to those purified on Percoll gradients for immunoglobulin-induced degranulation, survival responses to cytokines, and morphology. 2. Materials and methods

2.1. Percoll purification Purification of eosinophils on Percoll gradients was performed as previously described (AbuGhazaleh et al., 1989) with slight modifications. Briefly, five parts of heparinized blood obtained from normal volunteers or patients with mild asthma or atopy were incubated with one part of 6% hetastarch in 0.9% NaC1 (Hespan, Du Pont Pharmaceuticals, Wilmington, DE) for 45 min at 37°C in order to sediment erythrocytes. After two washes with Pipes (piperazine-N,N',-bis(2-ethane sulfonic acid)) buffer (25 mM Pipes, 50 mM NaC1, 5 mM KCI, 40 mM NaOH and 5.4 mM glucose,

pH 7.4), cells were suspended in a Percoll (Sigma Chemical Co., St. Louis, MO) solution with a density of 1.070 g / m l containing 5% heat-inactivated defined calf serum (DCS, HyClone, Logan, UT) and overlayered on discontinuous Percoil density gradients of densities 1.080, 1.085, 1.090, 1.100 and 1.120 g/ml. Percoll osmolarities ranged from 290 to 315 mosmol/kg with a pH of 7.4. Gradients were centrifuged at 4°C for 45 min at 1500 x g in a fixed angle rotor (J2-21, Beckman Instruments, Palo Alto, CA). Eosinophil-rich fractions were collected from the gradient using a polystaltic pump and microfractionator. Cells were washed with Pipes buffer with 1% DCS, and eosinophil numbers and purities were determined by staining with Randolph's phloxine-methylene blue. Normodense eosinophils with purities of 80% or greater were pooled and used for experiments. Contaminating erythrocytes were lysed either utilizing hypotonic lysis or ammonium chloride, as described below. Cells were washed twice with Hybri-Care medium (American Type Culture Collection, Rockville, MD) and resuspended in Hybri-Care with 10% DCS. Again, eosinophil numbers and purities were confirmed by Randolph's stain.

2.2. Immunomagnetic purification MACS was used with minor modifications of the method described by Hansel et al. (1991). Briefly, peripheral blood was obtained from the same donor on the same day and time as the blood for Percoll purification, anticoagulated with 130 mM trisodium citrate and diluted with an equal volume of phosphate buffered saline (PBS). Eosinophil and noneosinophil cell numbers were determined with Randolph's stain. Diluted blood was overlayered on an isotonic Percoll solution (density 1.082 g / m l ) (volume of diluted blood is twice the volume of Percoll) in 50 ml Falcon plastic tubes (Becton-Dickinson Labware, Lincoln Park, NJ), and centrifuged for 30 min, 1000 x g at 4°C (TJ-6, Beckman Instruments). The supernatant and the mononuclear cells at the interface were carefully aspirated and the inside wall of the tube was wiped with a sterile gauze to remove mononuclear ceils attached to the wall.

M. Ide et al. /Journal of lmmunological Methods 168 (1994) 187-196 To the pellet of granulocytes and erythrocytes, 5 vols. of ice-cold ammonium chloride lysis buffer (NHaCI 155 mM, KHCO 3 10 mM, and E D T A 0.1 mM) were added and erythrocytes were allowed to lyse for 15 min on wet ice. Alternatively, 20 ml of ice-cold sterile water (Baxter Healthcare, Deerfield, IL) were added to the pellet and mixed gently for 30 s after which 20 ml of 2 x Pipes buffer was added. If erythrocytes remained, the procedure was repeated again. After erythrocyte lysis, granulocytes were washed twice in PBS with 2% DCS and transferred to a 15 ml Falcon plastic tube. The supernatant was carefully aspirated, leaving the pellet nearly dry. The cell pellet was cooled on ice and an approximately equal volume of CD16 antibody conjugated to magnetic particles (Miltenyi Biotec, Sunnyvale, CA) was added. Cells and antibody-conjugated particles were incubated for 60 min on ice with occasional gentle mixing. Concurrently, the separation column was washed, incubated with PBS containing 2% DCS and positioned in the MACS strong magnetic field at room temperature. The column was flushed with ice-cold PBS with 2% DCS prior to the separation procedure. 5 ml of ice-cold PBS with 2% DCS were added to the cell/particle mixture and cells were gently resuspended. The cell suspension was loaded onto the top of the separation column. Cells were eluted with 15 ml of ice-cold PBS with 2% DCS in 5 ml fractions. Randolph's stain was used to count the number and purity of eosinophils. Fractions of eosinophils with purity 95% or greater were pooled. Cells were washed and resuspended in Hybri-Care. 2.3. Eosinophil suruival assay The survival assay was performed as described previously (Wallen et al., 1991) with minor modifications. 50 ~1 of cell suspension were plated into 96-well, flat bottom, half area microtiter plates (Costar, Cambridge, MA) at 1 x 104 cells/ well in Hybri-Care medium with 10% DCS. Serial dilutions of recombinant human interleukin-3 (IL-3) (Genzyme, Boston, MA), recombinant human IL-5 (a kind gift from S. Narula, Ph.D., Schering-Plough, Kenilworth, N J) or recombinant human granulocyte-macrophage colony-stimulat-

189

ing factor (GM-CSF, Genzyme) were added. To study the effect of ammonium chloride on eosinophil survival, 50/zl of cell suspension were plated at 1 x 104 cells/well and 25 ~1 of a serial dilution of IL-5 and varying concentrations of ammonium chloride were added. In these experiments, the final culture volume was 100 ~l/well. Cells were incubated for 4 days at 37°C, 5% CO 2 in a humidified incubator. After 4 days culture, 50/zl of supernatant was carefully removed from each well, and 10/~1 of fluorescein diacetate (0.2 m g / m l in PBS) were added. After 30 min at 4°C, 10/zl of propidium iodide (0.5 ~ g / m l in HybriCare medium) were added. After an additional 5 min, the numbers of live cells and dead cells were counted using a hemacytometer and a fluorescent microscope. Percent viability was calculated by dividing the number of live cells by the total number of ceils counted. 2.4. Morphologic analysis by electron microscopy The procedure used was based on a previous study (Peters et al., 1988). Samples from five normal or mildly atopic donors were randomly selected for study. Briefly, eosinophils were purified by the MACS or Percoll methods with erythrocyte lysis either by ammonium chloride or by hypotonic lysis. Cells were fixed in Trump's fixative overnight at 4°C and postfixed with 1% OsO 4 in 0.1 M phosphate buffer. After en bloc staining with 2% uranyl acetate and dehydration through a series of increasing concentrations of ethanol, cells were infiltrated and embedded in Spurr (Electron Microscopy Sciences, Fort Washington, PA). Thin sections were mounted on copper grids and examined with a Philips CM12 electron microscope. Two or three areas were randomly selected and photographed at a constant magnification ( x 1800). Eosinophils in the resulting photomicrographs were scored by three independent observers for the degree of granule lucency from 0 to 2, where 0 = none, 1 = moderate, and 2 = marked granule lucency. 2.5. Eosinophil degranulation Immunoglobulins coupled to cyanogen bromide-activated Sepharose 4B beads (Sigma) were

190

M. Ide et al. /Journal of Immunological Methods 168 (1994) 187-196

used as stimuli for eosinophil degranulation (Abu-Ghazaleh et al., 1989). The beads were coupled to human secretory I g A (slgA, Accurate Chemical Scientific Corp., Westbury, NY), hum a n serum I g G (Organon Teknika-Cappel, Malvern, PA), or the negative control protein, ovalbumin (OVA, Sigma) at a concentration of 5 mg protein per ml packed beads as previously described. Eosinophil degranulation was examined using a previously established procedure (Abu-Ghazaleh et al., 1989) with minor modifications. Purified eosinophils were washed twice with HybriCare medium containing 50 / z g / m l gentamicin and 0.1% h u m a n serum albumin. The washed cells were resuspended at 5 x 105 c e l l s / m l in the same medium. Aliquots of the cell suspension (100 /~1) were incubated for 60 min in 96 well round bottom tissue culture plates (Costar) with 50 /xl of Hybri-Care medium or IL-5, 1 n g / m l . Cellular degranulation was initiated by addition of 5 0 / z l of m e d i u m alone, O V A or immunoglobulin coated Sepharose beads. After incubation for 2 h at 37°C, 5% CO2, the plates were centrifuged at 1000 x g for 10 min at 4°C. Ceil-free supernatants were collected immediately and frozen at - 2 0 ° C . To quantitate eosinophil degranulation, the amount of released eosinophil-derived neurotoxin ( E D N ) was measured by radioimmunoassay ( R I A ) as previously described (Abu-Ghazaleh et al., 1989). Total cellular E D N was detected by

100] ~ - - - u

Table 1 Comparison of yield and purity of eosinophils between MACS and Percoil purification methods Method Blood no. of mean Mean % Mean% volume eosinophils yield purity (ml) MACS 20 6.6x 10 6 + 1.1 90.2+4.5 97.1+0.5 Percoll 120 6.4×106+2.4 19.5+5.6 77.8±2.9 Mean values (n = 15) are shown as the mean + SEM.

lysing cells with 0.5% Nonidet P-40. All assays were performed in duplicate. The R I A for E D N was a double antibody competition assay using rabbit anti-EDN antibody, 125I-labeled EDN, and burro antirabbit I g G (Ackerman et al., 1983). 2.6. Statistical analysis

D a t a are expressed as m e a n + SEM. Statistical significance of the differences between procedures was assessed by paired Student's t test.

3. Results

3.1. Purity and yield

Table 1 compares the Percoll and MACS methods (n = 15) for eosinophil purity and yield. Even though the Percoll method started with six times the volume of blood, the MACS method gave more eosinophils with higher purities. Contaminating cells were nearly always neutrophils.

""o'-..-_. o ...... o,, 3.2. Eosinophil survival ~

60

"'/

} '°t

',,\

,o

O

0

1

~

2

",, 3

4

5

Day

Fig. 1. Eosinophil spontaneous survival over 5 days at 4°C (dotted line) and 37°C (solid line) for cells purified by MACS using either water (o) or ammonium chloride ( • ) to produce lysis (n = 2).

Because lysis of erythrocytes is one of the key steps in the purification of eosinophils by MACS, we compared the survival of cells after erythrocyte lysis with water or ammonium chloride. Fig. 1 shows eosinophil survival in the absence of exogenous cytokine. At 4°C, cells survived longer when erythrocytes were lysed by water than by a m m o n i u m chloride. At 37°C, eosinophils died more quickly than at 4°C, but again cells from which erythrocytes had been removed by lysis with water survived longer than those exposed to

191

M. Ide et al. /Journal of lmmunological Methods 168 (1994) 187-196 IL-3

100"

GM-CSF

100A

A

80"

80"

~ 60"

....

"~

60"

~ 40

~

40"

"~ 2 O" i11

"~ 20 W

0

0.1

1

10

100

1000

100

1000

IL-3 (poJml)

0

0.01

0.1 1 10 GM-CSF (pg/ml)

100

IL-5

100 80 60

i

40

20 0

0.1

1 10 IL-S (pg/ml)

Fig. 2. Eosinophil survival response (mean + SEM) to exogenous IL-3, IL-5 and GM-CSF after 4 days of culture at 37°C (n = 3). Eosinophils were purified either by Percoll (o) or MACS (11) method, and treated either by ammonium chloride (dotted line) or water (solid line) to eliminate erythrocytes.

ammonium chloride. Fig. 2 shows eosinophil survival response to exogenous IL-3, IL-5 and GMCSF after four days of culture at 37°C. Eosinophils were purified either by Percoll or MACS and treated either by ammonium chloride or water to eliminate erythrocytes. Eosinophils purified by Percoll and by MACS showed similar dose responses to IL-3, IL-5 and GM-CSF (Fig. 2) provided that erythrocytes were removed by hypotonic lysis. However, lysis of erythrocytes with ammonium chloride significantly suppressed eosinophil survival (p--0.02 for IL-3, p = 0.005 for IL-5, and p = 0.005 for GM-CSF at 100 p g / ml) regardless of whether eosinophils were purified by Percoll or MACS. These observations (Figs. 1 and 2) suggest that survival of eosinophils exposed to ammonium chloride is inhibited both in the presence and absence of exogenous cytokines. As shown in Fig. 3, eosinophil survival induced by IL-5 was significantly suppressed (p

< 0.05) by using either 1 mM or 10 mM ammonium chloride in the culture medium.

3.3. Morphological analysis by electron microscopy Eosinophil preparations from four normal donors were selected for examination by electron microscopy. Three independent observers scored and counted the lucency of the granules in eosinophils present in electron photomicrographs taken at × 1800. Two fields, each containing approximately 20 cells, were investigated for each patient by each method. MACS purified eosinophils, as shown in Table 2 and Fig. 4, demonstrated significantly more eosinophils with lucent granules when treated with ammonium chloride than when treated with hypotonic lysis (p < 0.01). No significant morphological difference was apparent between eosinophils purified by the Per-

M. Ide et al. /Journal of lmrnunological Methods 168 (1994) 187-196

192

Table 2 Morphological comparison of granule lucency in eosinophils purified by MACS or Percoll Donor

MACS (water)

MACS ( a m m o n i u m chloride)

Percoll (water)

None (0)%

Moderate (1)%

Marked (2)%

None (0)%

Moderate (1)%

Marked (2)%

None (0)%

Moderate (1)%

Marked (2)%

2 3 4

22 17 2 12

63 61 51 49

16 22 48 39

5 4 0 4

42 53 19 30

53 43 81 67

28 15

57 66

16 18

Mean%

13.3

56.0

31.3

3.3 b

36.0 b

61.0 c

21.5 d

61.5 d

17.0 d

1

_

a

-

Not determined. b Significantly lower than eosinophils treated by water in MACS method, p < 0.05. c Significantly higher than eosinophils treated by water in MACS method, p < 0.01. d No significant differences than eosinophils treated by MACS (water). a

coll and MACS methods when cells were treated with hypertonic lysis.

3.4. Eosinophil degranulation To test the effects of purification procedures on eosinophil degranulation, eosinophils purified by MACS with erythrocytes lysed by either hypotonic lysis or ammonium chloride were compared to Percoll purified eosinophils. As shown in Fig. 5, eosinophils purified by any of these procedures released EDN when stimulated by IgG or slgA coated beads. Within the same purification

100t 80t

ioo 20

0

0

0.1

1

10

100

IL-5 (poj~ Fig. 3. Eosinophil survival response ( m e a n + SEM) to exogenous IL-5 after 4 days of culture at 37°C (n = 4) with different concentrations of a m m o n i u m chloride. Eosinophils were purified by M A C S with hypotonic lysis to remove erythrocytes. T h e • indicates 0 m M a m m o n i u m chloride, © indicates 0.1 mM, A indicates 1 m M and [] indicates 10 mM.

method, the response to slgA was greater than to IgG in agreement with previous reports (AbuGhazaleh et al., 1989; Kita et al., 1991). Spontaneous EDN release, as shown by medium or OVA, was slightly greater from MACS purified eosinophils than from Percoll purified eosinophils. Release of EDN with IgG beads was significantly greater from MACS purified eosinophils than from Percoll purified cells (p < 0.05); no difference was observed between erythrocyte lysis by water and by ammonium chloride in MACS purified eosinophils. When slgA beads were used as stimuli, the amounts of EDN release were roughly comparable between MACS and Percoll purified eosinophils even though MACS purified eosinophils exposed to ammonium chloride released slightly less EDN than cells exposed to hypotonic lysis. These findings suggest that eosinophils purified by MACS have qualitatively similar, but slightly greater, responses to stimuli than do Percoll purified cells. In addition, exposure of eosinophils to ammonium chloride during the purification procedure has little effect on eosinophil degranulation. We also tested the direct effects of ammonium chloride on eosinophil degranulation. Eosinophils purified by MACS with hypotonic lysis of red blood ceils were incubated with IgG or slgA coated beads in the presence of various concentrations (0, 0.01, 0.1, 1.0, 10.0 mM) of ammonium chloride. Under these conditions, ammonium chloride up to 1 mM had no effect on degranula-

M. lde et al. /Journal of lmmunological Methods 168 (1994) 187-196

193

2

l

Y

J

t

Fig. 4. Electron photomicrographs ( × 1800) of eosinophils showing degree of granule lucency. N u m b e r e d arrows indicate scores from 0 to 2, where 0 = none, 1 = moderate, and 2 = marked granule lucency.

z 0

30-

U.I m

O

I--

20

C

Medium

OVA

IgG Beads

slgA

Fig. 5. Eosinophil degranulation ( m e a n + S E M ) induced by stimulation with OVA, IgG and slgA coated Sepharose beads (n = 5). Solid bars indicate purification by Percoll and lysis by water, hatched bars purification by M A C S and lysis by water, and open bars purification by M A C S and lysis by 155 m M a m m o n i u m chloride. * p < 0.05.

tion although 10 mM ammonium chloride slightly inhibited degranulation (data not shown). IL-5 is known to enhance eosinophil degranulation induced by immunoglobulin-coated beads (Fujisawa et al., 1990). Because exposure of eosinophils to ammonium chloride inhibited IL-5induced survival in vitro, we tested whether ammonium chloride exposure inhibited IL-5 enhanced eosinophil degranulation. Purified eosinophils were first incubated with 1000 p g / m l of IL-5 for 1 h and then stimulated for 2 h with IgG or slgA coated beads. As shown in Fig. 6, IL-5 enhanced eosinophil degranulation in every case even though the increases were not always significant. For example, with IgG beads, eosinophils purified by MACS with erythrocyte lysis by ammonium chloride released significantly higher amounts of EDN in the presence of IL-5 than in

194

M. 1de et al. /Journal of lmmunological Methods 168 (1994) 187-196

] Medium

i OVA

10

]

*

,°

.

•

o

o

PW

MW Method

MA

IgG

PW

~-

0

O' PW

MW Method

MA

el~A

PW

MW Method

MA

*

MW Method

MA

Fig. 6. Eosinophil degranulation (mean + SEM) in presence (hatched bars) or absence (solid bars) of 1000 p g / m l IL-5 (n = 3). PW indicates purification by Percoll method and lysis by water, MW indicates purification by MACS method and lysis by water, MA indicates purification by MACS method and lysis by 155 mM ammonium chloride. * p < 0.05.

the absence of IL-5 (p < 0.05); Percoll purified eosinophils behaved similarly. No obvious differences were noted in degranulation responses to IL-5 when eosinophils were purified by Percoll or MACS and erythrocytes were lysed by water or ammonium chloride. These findings suggest that the effect of IL-5 on eosinophil degranulation is not inhibited by ammonium chloride.

4.Discussion We have compared a new isolation technique for eosinophils, which uses immunomagnetic beads to obtain high yields and highly purified eosinophils from human peripheral blood, to the prior method, which uses Percoll density gradient centrifugation (Gartner, 1980). MACS proved to be very efficient, yielding increased numbers of highly purified eosinophils even from normal in-

dividuals (Table 1). To use MACS purified eosinophils for investigations of cell function, we compared the responses of eosinophils purified by this procedure to those purified on Percoll. During these studies, we discovered that the ammonium chloride used to lyse the erythrocytes altered the eosinophils. Initially, we compared the eosinophil survival response to IL-5 using cells obtained by the standard discontinuous Percoll gradient and by MACS (in which erythrocytes were lysed by exposure to ammonium chloride). Interestingly, preliminary results indicated that the two eosinophil populations differed. For example, Percoll purified eosinophils showed a typical dose-response survival to IL-5 that increased from 1 p g / m l and plateaued at 100 p g / m l (Fig. 2). In contrast, the MACS purified eosinophils demonstrated a weak response to IL-5; for example, IL-5, 100 pg/ml, showed about 80% eosinophil survival after hypo-

M. lde et al. /Journal of Immunological Methods 168 (1994) 187-196

tonic lysis of erythrocytes, whereas eosinophils treated with ammonium chloride showed less than 40% survival by day 4. Therefore, we hypothesized that the MACS method could alter the eosinophils, rendering them less sensitive to exogenous IL-5. We tested the hypothesis that the ammonium chloride used to lyse erythrocytes might affect eosinophil viability. An initial study of eosinophil viability after exposure to either water or ammonium chloride indicated that the ammonium chloride reduced viability (Fig. 1), presumably due to damage or deactivation of the cells. To determine whether the CD16 negative selection step might affect the eosinophils, we purified eosinophils and neutrophils by discontinuous Percoll density gradient and then used negative selection with anti-CD16 to eliminate neutrophils. Lysis of erythrocytes was not needed in this procedure. The response to IL-5 of these eosinophils was similar to that of Percoll purified cells, suggesting that the decreased viability of MACS purified eosinophils was not due to the CD16 negative selection step (data not shown). This result supported the observation that the erythrocyte lysis procedure employed in MACS affected eosinophil viability. Therefore, we compared eosinophils after treatment with each of the two lysis methods, ammonium chloride and hypotonic lysis with water, in both the Percoll and MACS purification procedures. The resulting dose-response curves to cytokines showed that the erythrocyte lysis method affected the eosinophil's response to IL-3, IL-5 and GM-CSF (Fig. 2), and that exposure to ammonium chloride markedly reduced these responses. No significant differences were observed in the eosinophil responses to cytokines when erythrocytes were lysed by water during either Percoll or MACS purification. To elucidate the differences observed in the eosinophil survival assay, we examined eosinophils by electron microscopy. There was a significant difference in the degree of granule lucency between those cells exposed to ammonium chloride and those exposed only to water, independent of the purification method used (Table 2). This difference may explain why eosinophils treated with ammonium chloride did not respond to cytokines.

195

Possible differences in the eosinophil response to secretagogues may have been related to the purification method and, therefore, were studied by stimulating cells with protein-coated beads and measuring EDN release. Regardless of purification methods, the response to OVA, IgG and slgA were similar to those published previously (Abu-Ghazaleh et al., 1989). We found that the EDN release in response to IgG beads was greater from eosinophils purified by the MACS method than that from Percoll purified cells. Furthermore, MACS purified eosinophils treated with ammonium chloride released less EDN than MACS purified eosinophils with hypotonic lysis when stimulated by slgA beads (Fig. 5). Although eosinophil survival responses to IL-5, GM-CSF and IL-3 were decreased in cells purified by MACS with ammonium chloride erythrocyte lysis (Fig. 2), the degranulation responses of those cells to IL-5 did not differ from those purified by the other procedures. These observations argue against the impairment of the IL-5 receptor by ammonium chloride treatment. Our work has confirmed that the MACS purification method is a very efficient procedure to obtain eosinophils with high purity even from normal individuals. Furthermore, lysis with ammonium chloride had little effect on eosinophil degranulation, but large effects on survival and morphology. Hypotonic lysis with water was found to be the better method for erythrocyte lysis when eosinophil biology is to be studied.

5. Acknowledgements We thank Dr. T. Hansel for numerous suggestions aiding the implementation of the MACS procedure, Mrs. Cheryl R. Adolphson for editorial assistance and Mrs. Linda H. Arneson for secretarial help.

6. References Abu-Ghazaleh, R.I., Fujisawa, T., Mestecky,J., Kyle, R.A. and Gleich,G.J. (1989) IgA-inducedeosinophildegranulation. J. Immunol.142, 2393.

196

M. Ide et al. /Journal of lmmunological Methods 168 (1994) 187-196

Ackerman, S.J., Loegering, D.A., Venge, P., Olsson, I., Harley, J.B., Fauci, A.S. and Gleich, G.J. (1983) Distinctive cationic proteins of the human eosinophil granule: major basic protein, eosinophil cationic protein, and eosinophilderived neurotoxin. J. Immunol. 131, 2977. Fujisawa, T., Abu-Ghazaleh, R., Kita, H., Sanderson, C.J. and Gleich, G.J. (1990) Regulatory effect of cytokines on eosinophil degranulation. J. Immunol. 144, 642. Gartner, I. (1980) Separation of human eosinophils in density gradients of polyvinylpyrrolidone-coated silica gel (Percoll). Immunology 40, 133. Gleich, G.J., Adolphson, C.R. and Leiferman, K.M. (1992) The eosinophil. In: J.I. Gallin, I.M. Goldstein and R. Snyderman (Eds.), Inflammation: Basic Principles and Clinical Correlates, 2nd edn. Raven Press, New York, p. 663. Hansel, T.T., Ound, J.D. and Thompson, R.A. (1990) Isolation of eosinophils from human blood. J. Immunol. Methods 127, 153. Hansel, T.T., De Vries, I.J.M., Iff, M.T., Rihs, S., Wandzilak,

M., Betz, S., Blaser, K. and Walker, C. (1991) An improved immunomagnetic procedure for the isolation of highly purified human blood eosinophils. J. Immunol. Methods 145, 105. Kita, H., Abu-Ghazaleh, R.I., Gleich, G.J. and Abraham, R.T. (1991) Role of pertussis toxin-sensitive G proteins in stimulus-dependent human eosinophit degranulation. J. Immunol. 147, 3466. Peters, M.S., Gleich, G.J., Dunnette, S.L. and Fukuda, T. (1988) Ultrastructural study of eosinophils from patients with the hypereosinophilic syndrome: A morphological basis of hypodense eosinophils. Blood 71, 780. Van Haelst Pisani, C., Kovach, J.S., Kita, H., Leiferman, K.M., Gleich, G.J., Silver, J.E., Dennin, R. and Abrams, J.S. (1991) Administration of interleukin-2 (IL-2) results in increased plasma concentrations of IL-5 and eosinophilia in patients with cancer. Blood 78, 1538. Wallen, N., Kita, H., Weiler, D. and Gleich, G.J. (1991) Glucocorticoids inhibit cytokine-mediated eosinophil survival. J. Immunol. 147, 3490.