Journal of Lenkocyte Biology �‘olttme 59, Februtt) 1996 189

Gadolinium induces macrophage apoptosisJoseph P. Mizgerd,* Ramon M. Molina,*t Rebecca C. Stearns, *Joseph D. Brain,* and AngelineE. Warner*.t*Physiology Program, Harvard School ofPublic Health, and tCenterfor Animal Resources and Comparative

Medicine, Harvard Medical School, Boston, Massachusetts

Abstract: Gadolinium (Gd) suppresses reticuloen-

dothelial functions in vivo by unknown mechanisms.

In vitro exposure of rat alveolar macrophages to

GdC136H2O caused cell death, as measured by try-

pan blue permeability, in both dose- and time-de-

pendent fashions. Even a 10-mm exposure to Gd

caused significant cell death by 24 h. The morphol-

ogy of Gd-treated cells, pyknosis and karyorrhexis

prior to loss of membrane integrity, suggested apop-

tosis. Upon flow cytometric examination, Gd-treated

propidium iodide-excluding cells demonstrated light

scatter changes characteristic of apoptotic cells (de-

creased forward and increased right angle scatter).

Gel electrophoresis of DNA from Gd-treated macro-

phages clearly showed the ladder pattern unique to

apoptotic cells. Electron-dense structures containing

Gd were observed via electron spectroscopic imaging

within phagosomes and also within nuclei (associated

with condensed chromatin). Gadolinium, endocyto-

sed by macrophages and distributed to nuclei, causes

apoptosis of macrophages in vitro. J. Leukoc. Biol.59: 189-195; 1996.

Key Words: cellular suicide . electron spectroscopic imaging

lanthanides . alt rast ructure

INTRODUCTION

Gadolinium (Gd), a lanthanide rare earth element, is com-

monly used to study the physiology of the reticuloen-

dothelial system. It “inactivates” macrophages (M�s),

particularly as measured by reduced clearance of test par-

tides from the blood [1] and by decreased localization of

circulating particles to resident M�s [2-4]. Macrophage-

mediated immune and inflammatory responses are also

suppressed in Gd-treated animals, such as the induction of

tolerance to portal venous antigen [5] and the development

of lethal endotoxin shock [6].

The extent of toxicity of Gd for M�s has not been re-

ported. Gadolinium is commonly accepted as a benign in-

hibitor of phagocytosis, but evidence suggests that the

number of M4s is lowered by Gd exposure. Gadolinium

injected intravenously reduces the M�-specific iinmuno-

histochemical staining of sections from rat liver or spleen

[7], and Gd treatment of M4s in vitro decreases the pres-

ence of adherent cells after 24 h of culture [6]. The present

studies were undertaken to investigate potential lethal ef-

fects of Gd for rat M�s and explore the type of cell death

induced by Gd (apoptosis or necrosis).

MATERIALS AND METHODS

Materials

The following items were purchased : gadolinium chloride hexahydrate

(GdCl361I20, from Aldrich Chemical Co., Milwaukee, WI), lipopolysac-

charide (LPS) from Escherichia coli serotype Olil:B4 (Sigma Chemical

Co., St. Louis, MO), recombinant rat interferon-gamma (IFN-y from

Gibco BRL, Grand Island, NY), 100 base pair ladder (Pharmacia, Pis-

cataway, NJ), fetal calf serum (HyClone Laboratories, Logan, UT), RPMI

1640 medium (Sigma Chemical Co.), and Dulbecco’s phosphate-buff-

ered saline (PBS, from Gibco BRL). To prevent cell loss by adhesion to

tissue culture plates, 24-well plates were coated prior to use by allowing

evaporation of 0.3% poly(2-hydroxyethyl methacry’late) (polyHEMA,

from Aldrich Chemnical Co.) in 95% ethyl alcohol under sterile condi-

tions [8]. Complete RPMI (cRPMI) was RPM1 1640 supplemented with

heat-inactivated fetal calf serum (10%), m-glutamine (2 mM), and anti-

biotics (iOO U/ml penicillin and 100 �.tg/mnl stre�)tomnycin).

Cell culture

Alveolar M�s were harvested fmomn 250 g of virus antigen-free male

Sprague-Dawley rats (Taconic, Germantown, NY) by repeated p�stmnor-

tem bronchoalveolar lavage with Dtmlbecco’s PBS supplemented with 1.2

mM ethylenediaminetetraacetic acid (EDTA). The cells were washed

and suspended in cRPMI to 5 x 10’ living M4/mnl in po1yHEMA-coated

24-well plates. Viability was determined at the beginning of the experi-

ment and at designated time points by exclusion of the �ital (lye trypan

blue (0.32%) as assessed by light microscopy. Gd preparations or vehi-

dc control (sterile saline) was added to all wells to reach final concen-

trations of 0, 0.027, 0.27, 2.7, 27, or 270 j.tf�1 Gd. Plates were incubated

for the specified timne (0, 8, 16, or 24 h) at 37#{176}Cin 5� CO�. For some

plates (transient Gd exposures), cells and mnediumn were transferred to

Eppendorf tubes at the noted timne (0 Is, iO mnin, 1 h, 4 h, or 24 h); cells

were spun out of Gd-containing mediumn (270 j.tf%1). washed, resus-

pended in Gd-free cRPMI, and incubated in poly HEMA-coated plates

at 37#{176}Cin 5% CO2 for the remainder of 24 h.

Abbreviations: BCG, bacillus Calmette-Gu#{233}ii n; cRPMI, coniplete

RPMI; ESI, electron spectroscopic imnagitig; Gd, gadol i niurn; IFN-y,interferon-y; LPS, lipopolysaccliaride; f%’1�, macrophage; MRI. magnetic

resonance imnaging; PBS, phosphate-buffered saline; polyHEMA, poly(2-

hydmoxyethyl methacrylate).

Reprint requests: Joseph P. Mizgerd, Physiology Pmograns, Harvard

School ofPuhlic Health, 665 Huntington Avenue, Boston. MA 02115.

Received July 10, 1995; revised October 2, 1995; accepted October 6,

1995.

A

B�. �

A

E

aa

V

a0

‘a

5�

V

UV

‘a

270

6

Time (hours)

IGdI (�tg/mI)

0. 0.03. 0.27

2.7

27

*

T

I � *

T -j.

� U1106 4h

Duration of Exposure

*

a. 1.�

25�

0

B

lh 24h

Microscopy

190 Journal of Leukocyte Biology Volumne 59, February 1996

For light microscopy, samples from the cell culture plates were diluted

with cRPMI and spun (800 rpm, 5 mm) with a cytocentrifuge (Shandon

Cytospin 2, Astmoor, England) onto glass slides that were stained with

Diff-Quik (Baxter Scientific, McGraw Park, IL). For electron micros-

copy, samples were fixed in 2.5% glutaraldehyde and then 1% OsO4

(each in 100 mM sodium cacodylate buffer, pH 7.4), dehydrated in a

series of graded ethanols and then propylene oxide, infiltrated with a

mixture of propylene oxide and Epon (Ernest F. Fullam, Latham, NY),

and finally embedded in Epon. Sections (100 nm) were poststained with

uranyl acetate and lead citrate and were examined with a Philips 300

electron microscope.

Flow cytometry

Flow cytometry was performed with an Ortho Cytofluorograf 50H flow

cytometer equipped with a Cicero computer interface system (Cytoma-

Fig 1 . Cytotoxicity of Gd for rat M4s in vitro. Each point represents the

mean and SEM from n 3 experiments. Groups were compared by analysis

ofvariance, and differences were considered significant (*) ifP < .05 as

determined by Scheffe’s F test. (A) Dose-response and time course of

toxicity. Alveolar M4s (5 x 105/ml) were incubated for the specified time

in the specified concentration of GdC1s6H2O at 37#{176}Cin po1yHEMA-

coated plates, and viability was assessed by the ability of cells to exclude

the vital dye trypan blue under light microscopic examination. (B) Loss

of viability induced by transient exposure to Gd. Alveolar M4s (5 x105/ml), incubated for the specified time in 270 �M Gd at 37#{176}Cin

polyHEMA-coated plates, were washed and resuspended in fresh Gd-free

medium and incubated further. Viability was assessed as above after a

total incubation time of 24 h.

10pm

Fig. 2. Microscopic morpholog�r of Gd-treated rat M4s. (A) Light mi-

crograph. Alveolar M�s (5 x 10 /ml) were incubated for 8 h in 270 j.�M

Gd at 37#{176}Cin po1yHEMA-coated plates and then cytocentrifuged onto

glass slides and stained with Diff-Quik. Note the clear pyknosis, karyor-

rhexis, and free apoptotic bodies. (B) Transmission electron micrograph.

Alveolar M4s were incubated as in (A). Cell pellets were fixed and

embedded as described in Materials and Methods. Shown is a profile from

a 100-nm section stained with uranyl acetate and lead citrate. Note the

pyknotic nucleus, cytoplasmic condensation, and particulate matter

within phagosomes (probably Gd).

tion, Fort Collins, CO). Mean forward and right angle light scatter data

were collected for 6000 M�s per treatment group. Data were collected

only for cells with intact membranes, which excluded the fluorescent

vital dye propidium iodide.

DNA analysis

Cells were suspended in lysis buffer (100 mM NaCl, 10 mM Tris-HC1

at pH 8.0, 5 mM EDTA, 0.5% sodium dodecyl sulfate, and 500 �igJml

pronase) and incubated overnight at 37#{176}C.The DNA was extracted with

an equal volume of25:24:1 phenol, chloroform, isoamyl alcohol mixture

and then precipitated in 400 mM sodium acetate and 70% ethyl alcohol

(EtOH) for 20 h at -80#{176}C.The DNA was washed twice (70% EtOH), air

dried, and then resuspended in RNase buffer (1 mM EDTA and 10 mM

Tris-HC1 at pH 8.0 with DNase-free RNase added to 10 U/mI final) for

1 h at 37#{176}C.After RNase activity was quenched (56#{176}Cfor 5 mm), lO-j.tg

samples of DNA (determined spectrophotometrically) were electropho-

resed on a 1% agarose gel in TBE buffer (50 V. 2.5 h). Electrophoresed

DNA was stained with ethidium bromide (1 �xg/ml) and visualized with

ultraviolet irradiation.

Electron spectroscopic imaging

Electron spectroscopic imaging (ESI) allows specific and sensitive iden-

tification of Gd atoms. Gadolinium content in unstained 30-nm sections,

fixed as above but infiltrated and embedded in Araldite 502 (Ernest F.

Fullam), was determined with the Zeiss CEM9O2 elemental analysis

program (CEM9O2 distribution auto program, IBAS User’s Manual,

1989). To determine background noise, proximate resin was examined

immediately prior to analysis of the cellular unit of interest. The back-

ground images (at 1 167, 1 157, and 1 147 eV) were subtracted from the

elemental edge of Gd (at 1 187, 1 167, and 1 147 eV) on the resin. Using

the highest gray level of this background as a cutoff, Gd within the

subcellular profile was determined (without subsequent changes in elec-

tron optics or video camera settings). Images for spectroscopy and map-

ping were collected with a Dage 511’ 66 video camera (Dage-MTI,

Michigan City, IN). ES! images at 250 eV were also collected on photo-

graphic sheet emulsion film (SO 163, Kodak).

RESULT

Cell death and Gd

Rat alveolar M4s were killed by Gd as measured by trypan

blue staining of cells with leaky plasma membranes (Fig.

ma

C.)

ma

ma

0

A Right Angle Scatter

1A). The effect ofGd was both dose and time dependent;

higher concentrations of Gd caused more rapid and wide-

spread cell death than lower concentrations. After 24 h in

270 �iM Gd, viability was lowered from 95 ± 1.5% in

control cells to 12 ± 2.1%. Loss of membrane integrity was

seen only after a considerable exposure time. Even at the

highest dose (270 riM), however, there was little or no

trypan blue staining after 8 h of culture.

The loss of viability observed at 24 h after transient Gd

treatment correlated with the length of time for which M4s

were exposed to 270 jiM Gd (Fig. 1 B). However, even a

brief (10-mm) exposure resulted in significant cell death

after 24 h; half of the cells (50 ± 3.2%) were killed. This

suggests either that Gd quickly and stably associates with

M�s or that Gd quickly initiates irreversible cell injury

that results in membranolysis hours later.

Morph ologic evidence of Gd-induced M4 apoptosis

Morphological changes associated with Gd treatment sug-

gested an apoptotic mode of cell death. Prior to loss of

membrane integrity (e.g.. at 8 h), light microscopy of cy-

tospin preparations clearly demonstrated pyknosis and

karyorrhexis typical of apoptotic cells (Fig. 2A). Cell bud-

600

. * *

400

200

0

Forward Right Angle

B Direction of Light Scatter

Mizgerd et al. Gadolinium induces macrophage apoptosis 191

,�

C- �015�=15

E

control M#{248}s� gadolinium-treated M#{248}s�

Fig. 3. Light scatter of Gd-treated rat M4s. Alveolar M4s (5 x 105/ml) were incubated at 37#{176}Cin polyHEMA-coated plates for 24 h with either 27 j.�M

Gd or vehicle control. Mean forward and right angle light scatter data were collected for 6000 propidium iodide-excluding cells per treatment group for

each n of 1. (A) Scattergram from a single experiment, demonstrating that the light scatter of most of the cell population incubated with Gd (in red)

differs from that of cells incubated with saline vehicle (in blue). The box represents the window from which data were collected, ranging from

approximately 450 to 1000 for forward scatter and from approximately 100 to 1000 for right angle scatter (arbitrary units). (B) Quantitative effects of

Gd on light scatter properties of M4s. Shown are the mean and SEM for n 4 experiments. The effects of Gd were statistically significant at P < .05(*), as assessed by the paired t-test.

192 Journal of Leukocyte Biology Volume 59, February 1996

ding and free apoptotic bodies .were also observed. Trans-

mission electron microscopy (TEM) demonstrated M�s

with nuclear pyknosis and karyorrhexis, as well as cyto-

plasmic condensation (Fig. 2B). In addition, some Gd-

treated cells appeared highly vacularized, with “lacy”

regions of cytosol, suggesti ng macropi nocytosis.

The proportion of cells exhibiting morphologic evidence

of apoptosis, as determined in stained cytocentrifuged

preparations examined under light microscopy by a

“blinded” investigator, increased with the duration of expo-

sure to Gd. Prior to Gd exposure, there were few dead cells

as determined by trypan blue exclusion (5.9 ± 2.2%; mean

± SEM, n = 3 experiments) and almost no apoptotic cells

(0.9 ± 0.6%). After 8 or 16 h of exposure to 270 j.tM Gd,

the percentage of cells appearing apoptotic had increased

(to 15 ± 1.8% and 46 ± 2.2%, respectively) and was equal

to or greater than the percentage dead (at 9.9 ± 2.5% and

123456

Fig. 4 Gel electrophoresis of DNA collected from Gd-treated rat M4s.

DNA were collected (as described in Materials and Methods) from alveolar�, . 0

M4s (5 X 10 /ml) mncul)ated at 37 C in polyHEMA-coated plates and

resolved on a 1% agarose gel. Approximately 10 j.�g of DNA was loaded

to the well of each lane as follows: (1) normal M�s, cultured 8 h; (2)

necrotic M�s, cultured 8 h after brief exposure to OCF (5 mM, �5 mm);(3) apoptotic M4s (positive control), cultured 48 h with 10 U/mI IFN-y

and 1 �g/nsl LPS; (4) Gd-treated M�s, cultured 8 h with 270 �M Gd; (5)

Gd-treated M4s, as in lane 4 but with DNA from a separate experiment;

(6) 100-hp ladder (Pharmacia). The 180-200-hp ladder pattern unique

to apoptosis is readily observed in lanes 3, 4, and 5.

42 ± 5.1%, respectively). This suggests that apoptosis con-

tributed importantly to cell death among M�s exposed to

Gd. After 24 h, the percentage of cells dead (88.5 ± 2.1%)

exceeded the percentage recognizably apoptotic (49.1 ±

2.2%). Nearly all of the cells cultured 24 h with Gd exhib-

ited abnormal morphology; it was difficult to differentiate

cells in very late stages of apoptosis amid necrosis.

Flow cytometric evidence of Gd-induced M4apoptosis

Apoptotic cells can be distinguished with a flow cytometer

by their pattern of light scatter [9, 10]. Apoptotic cells

show decreased forward light scatter (associated with cell

shrinkage) and increased right angle light scatter (associ-

ated with increased granularity). This pattern was observed

with rat alveolar M4s exposed to Gd (Fig. 3). Mean for-

ward scatter of Gd-treated cells (27 �.tM Gd, 24 h) de-

creased to 81.5 ± 2.1% of that of untreated controls,

whereas mean right angle scatter of Gd-treated cells in-

creased to 134 ± 7.1% of the control value. No changes in

forward or right angle light scatter were observed in M4s

exposed to Gd for 1 h (data not shown), demonstrating that

light scatter changes were not an artifactual result of Gd

interaction with cells. The changes in light scatter by cells

incubated with Gd are strongly suggestive of apoptosis.

DNA evidence of Gd-induced M4 apoptosis

Gel electrophoresis of DNA from Gd-treated M4s demon-

strated the ladder pattern unique to apoptotic cells (Fig.

4). This ladder pattern reflects apoptosis-specific DNA

endonuclease activity; preferential cleavage of internu-

cleosomal stretches of DNA yields DNA fragments in in-

crements of 180-200 bp. The ladder pattern observed in

gel electrophoresis of DNA from Gd-treated M4s is similar

to that from a positive control for M4 apoptosis, induced by

culture in IFN-y plus LPS [11]. Figure 4 shows that the

ladder was not observed in M�s that were normal (no treat-

ment) or necrotic (OCl treatment). The presence of the

DNA ladder pattern, combined with the morphological and

flow cytometric observations noted above, is conclusive

evidence of apoptosis.

Intracellular localization of Gd

Gadolinium was readily apparent within M4s exposed to

270 �.tM Gd for 8 h. Electron spectroscopic imaging clearly

demonstrated electron-dense structures containing Gd

within phagolysosomes (Fig. 5). Electron-dense material

observed within nuclei also contained Gd when examined

by ESI (Fig. 6); these Gd-containing structures were typi-

cally associated with condensed chromatin. Chromatin as-

sociated with Gd may reflect apoptotic DNA

condensations, as the electron density and shape of the

chromatin condensations appeared distinct from those of

Mizgerd et a!. Gadolinium induces macropliage apoptosis 193

normal heterochromatin or nucleolar condensations. Elec-

tron-dense regions that were associated with mitochondria,

endoplasmic reticulum, Golgi apparatus, or plasma mem-

brane were examined spectroscopically, but none demon-

strated spectra indicating Gd (data not shown).

DISCUSSION

. .. S

Fig. 5. Gd within M4 phagolysosomes. Rat alveolar M4s (5 X 10 /ml)

were incubated for8 h in 270DM Gd at 37#{176}Cin po1yHEMA-coated plates

and then fixed and embedded as described in Materials and Methods. (A)

Ultrastructure of M4 cytoplasm, demonstrating electron-dense structures

within phagosomes. Shown is a sheet film photographic image at 250 eV

from a 30-nm unstained section. (B) ESI binary image of Gd (in yellow)

from the region shown in (A), collected with a Zeiss CEM9O2. (C) Overlay

of ES! binary image and digitally recovered ultrastructural image, dem-

onstrating the ultrastructural localization of Gd (in yellow) within

phagosomes

The present studies demonstrate that Gd induces apoptosis

of M4s. The effects of Gd are both dose and time depend-

ent. The association of Gd with M4s occurs within minutes

and probably reflects stable binding and internalization

(phagocytosis or pinocytosis), because Gd is observed

within endosomes. Morphological and biochemical

changes characteristic of apoptosis are apparent by 8 h

after exposure, and the loss of plasma membrane integrity

appears 16 to 24 h after exposure.

Previously, Gd was considered a benign inactivator of

M� functions, inhibiting phagocytosis of test particles and

limiting the production of immune and inflammatory me-

diators. However, indirect evidence suggested that Gd

might kill M�s [6, 7]. The data reported here demonstrate

that Gd kills M4s by inducing apoptosis.

Apoptosis may contribute to Gd-induced M� dysfunc-

tions observed in vivo. The M4s affected most by intrave-

nous Gd injections in vivo are those resident within the

vasculature (Kupffer cells, splenic M4s, and pulmonary

intravascular M4s); it remains to be determined whether

the kinetics or dose-response curves of Gd-induced apop-

tosis in vascular bed M4s differ in any significant fashion

from those of alveolar M4s. Alveolar M4 apoptosis is oh-

served in vitro within 8 h, and even a 10-mm pulse of Gd

causes significant cytotoxicity 24 h later. It is impossible

to determine precisely the dose of Gd to which vascular

bed M�s are exposed in vivo, but the doses effective in

vitro (27 �.tM is equivalent to 10 ;gmg/ml) are plausible

considering the total amounts (5 to 20 mg/kg) administered

in vivo [3, 4, 6, 7]. Thus, apoptosis of M4s in vivo may

contribute to the effects noted in Gd-treated animals.

Inhibition of M4 functions may further contribute to the

overall suppression of the mononuclear phagocyte system

caused by intravenous administration of Gd. Rat M4s

treated with Gd in vitro are not phagocytic (R.M. Molina

and J.P. Mizgerd, unpublished observations). Other func-

tions may also be suppressed. Thus, the Gd-induced sup-

pression of mononuclear phagocytes observed in vivo

probably reflects a combination of reduced numbers of

viable M4s and suppressed activity of the remaining M�s.

These studies corroborate a previous study identifying

Gd within phagosomes of Kupffer cells after intravenous

administration [7]. The present investigation adds to the

previous study by demonstrating intracellular Gd outside

phagosomes (inside nuclei). It is tempting to hypothesize

that intranuclear Gd is responsible for inducing apoptosis

in M4s-for example, by activating Ca2tdependent en-

donucleases or by directly inducing DNA damage. This

hypothesis suggests several avenues for research.

194 Journal of Leukocyte Biology Volume 59, February 1996

These results confirm previous reports establishing flow

cytometric methods for characterizing apoptotic cells by

light scatter [9, 10]. In addition to the data shown, prelimi-

nary experiments examined the light scatter patterns from

rat alveolar M4s made necrotic by either reactive oxygen

species damage by OCl or membrane attack by the toxic

plant metabolite saponin. Both models of necrosis caused

changes in light scatter from normal cells, neither of which

was similar in pattern to the changes attributed to apop-

tosis (data not shown). The former increased forward scat-

ter without affecting right angle scatter, whereas the latter

decreased right angle scatter without affecting forward

scatter. Apoptotic cells (Gd-treated M�s) displayed de-

creased forward scatter and increased right angle scatter

when compared with healthy cells, as previously reported

to be characteristic of apoptosis [9, 10].

A

1.7 urn

Gd is administered clinically to facilitate scanning by

magnetic resonance imaging (MRI). Does the Gd in MRI

analysis induce apoptosis in patients’ M4s? This seems

unlikely, because for MRI analyses Gd is bound to large

carner molecules, significantly enhancing the urinary ex-

cretion of Gd and virtually eliminating the association of

intravenously injected Gd with liver and spleen M4s [12].

Apoptosis of M4s may be a critical mechanism of host

defense, in particular against microbes that survive within

M�s. Apoptosis of M4s is inhibited by the obligate intra-

cellular parasite Leishmania donovani [13]. Induction of

apoptosis in monocytes infected with the intracellular bac-

teria Mycobacterium bovis bacillus Calmette-Gu#{233}rin (BCG)

reduces the number of viable bacteria, whereas no such

effect is noted with induction of necrosis [14]. Improved

understanding of the cell biology and regulation of M4

Fig. 6. Gd within M4 nucleus. Rat alveolar M�s (5 x 105/ml) were incubated for 8 h in 270 �M Gd at 37#{176}Cin po1yHEMA-coated plates and then fixed

and embedded as described in Materials and Methods. (A) Ultrastructure ofan M4 exposed to Gd in vitro. Shown is a sheet film photographic image at

250 eV from a 30-nm unstained section. The arrow indicates the region containing intranuclear electron-dense structures, which was magnified for ES!

analysis in (B)-(D) (note that orientation is shifted slightly in panels B-D). (B) Ultrastructure of M� nuclear matrix, demonstrating electron-dense

structures associated with condensed chromatin. Shown is a sheet film photographic image at 250 eV, as in (A). (C) ESI binaiy image ofGd (in yellow)

fromthe region shown in (B), collected with aZeiss CEM9O2. (D) Overlay ofESI binary image and digitally recovered ultrastructural image, demonstrating

the nuclear localization of Gd (in yellow) associated with condensed chromatin.

Mizgerd et al. Gadoliniwn induces macrophage apoptosis 195

apoptosis and the ability to manipulate M4 apoptosis phar-

macologically may facilitate future medical advances

based on enhancing or inhibiting M� functions.

ACKNOWLEDGMENTS

We thank the following individuals for excellent technical

support: Bonnie Meek and Caroline Snowman with TEM,

and Amy Imrich with flow cytometry.

REFERENCES

1 . Lazar, G. (1973) The reticuloendothelial-blocking effect of rare earth metalsin rats. J. Reticutoendothel. Soc. 13, 231-237.

2. Husztik, E., Lazar, C., Parducz, A. (1980) Electron microscopic study ofKupffer-cell phagocytosis blockade induced by gadolinium chloride. Br. J.Exp. Pat/wi. 61, 624-630.

3. Roland, C.R., Mangino, MJ., Flye, M.W. (1993) Lanthanide “blockade” ofantigen-presenting cells suppresses lymphocyte proliferation by inducingnitric oxide synthesis. J. Surg. Res. 54, 401-410.

4. Molina, R.M., Warner, A.E., Brain, J.D. (1995) Uptake ofcirculating path-

des by sheep pulmonary intravascular macrophages (P!Ms) is significantlydecreased by intravenous gadolinium chloride. FASEB J. 9, A570.

5. Squiers, E.C., Salomon, D.R., Pickard, LL, Howard, R.R., Pfaff, W.W.(1990) Abrogation of the induction of portal venous tolerance in a cardiactransplantmodelresultingfrom Kupifercell inhibition bygadolinium. Trans-

plantation 50, 171-173.

6. limuro, Y., Yamamoto, M., Kohno, H., Itakura, J., Fujii, H., Matsumoto, Y.(1994) Blockade of liver macrophages by gadolinium chloride reduceslethality in endotoxemic rats-analysis of mechanisms of lethality in endo-toxemia. J. Leukoc. Biol. 55, 723-728.

7. Hardonk, M.J., Dijkhuis, F.W.J., Hulstaert, C.E., Koudstaal, J. (1992)Heterogeneity of rat liver and spleen macrophages in gadolinium chlo-ride-induced elimination and repopulation. J. Leukoc. Biol. 52, 296-302.

8. Folkman, J., Moscona, A. (1978) Role ofcell shape in growth control. Nature273, 345-349.

9. Swat, W., Ignatowicz, L., Kisielow, P. (1991) Detection of apoptosis ofimmature CD4+8+ thymocytes by flow cytometry. J. immunol. Methods

137, 79-87.

10. Dive, C., Gregory, C.D., Phipps, DJ., Evans, D.L., Milner, A.E., Wyllie,A.H.(1992) Analysis and discrimination of necrosis and apoptosis (programmedcelideath) by multiparameterfiow cytometry. Biochim. Biophys. Acta 1133,275-285.

1 1. Albina, i.E., Cui, S., Mateo, RB., Reichner, J.S. (1993) Nitric oxide-medi-ated apoptosis in murine peritoneal macrophages. J. Immunol. 150,5080-5085.

12. Weinmann, HJ., Brasch, R.C., Press, W.R., Wesbey, G.E. (1984) Charac-teristics ofgadolinium-DTPA complex: a potential NMR contrast agent. AiR142, 619-624.

13. Moore, K.J., Matlashewski, G. (1994) Intracellular infection by Lei.shnmania

donovani inhibits macrophage apoptosis. J. immunol. 152, 2930-2937.

14. Molloy, A., Laochumroonvorapong, P., Kaplan, C. (1994) Apoptosis, but not

necrosis, of infected monocytes is coupled with killing of intracellular

bacillus Calmette-Gu#{233}rin. J. Exp. Med. 180, 1499-1509.