In vivo labeling of neutrophils using a fluorescent cell linker

Kurt H. Albertine and Marlys H. Gee

Departments ofMedicine and Physiology, Division ofPulmonary Medicine and Critical Care, Jefferson Medical
College, Thomas Jefferson University, Philadelphia, Penruylvania

Abstract: To study neutrophil circulation time and of neutrophil trafficking in vivo have relied on labeling the

Downloaded from https://academic.oup.com/jleukbio/article/59/5/631/6979182 by University of Rochester user on 17 January 2024

trafficking in vivo requires labeling the cells so that cells ex vivo, most commonly using radioisotopes [1-3].
their movement can be followed temporally and The drawbacks of such labeling procedures, however, in-
spatially. Labeling procedures used to date, however, dude variable labeling efficiency, altered cell function,
have relied on ex vivo separation and labeling meth- and elution of label from the cells of interest [4, 5]. More-
ods, an undesired consequence of which may be over, the labeled cells, upon reinjection into a host, prefer-
neutrophil activation. Moreover, the labeled cells entially stick in the lung. Two to four hours of recovery
preferentially stick in the pulmonary circulation for time is required before the labeled cells equilibrate in the
several hours upon reinjection into the host. There- circulating pool.
fore, we devised an alternate labeling procedure that The present study evaluates the efficacy of a different
relied on intra-arterial infusion of the fluorescent labeling procedure that is not reliant on isolation and ex
phagocytic cell linker PKH26. We infused PKH26 vivo labeling procedures. Rather, the procedure relies on
(5.5 x 1O- M for 1 h) into six awake sheep and in vivo infusion of a fluorescent phagocytic cell linker
measured systemic and pulmonary hemodynamics system originally developed to follow cell and drug deliv-
and lung lymph dynamics continuously for 3-4 h ery to focal areas of immunological reactions (PKH2#{212}; Zy-
after the infusion. We collected simultaneous arterial naxis Cell Sciences, Malvern, PA; marketed by Sigma
and venous blood samples hourly for 4 h, and again Chemical Co., St. Louis, MO) [6]. This cell linker system
24 h after the infusion ended, for white blood cell uses a fluorescent lipophilic probe that is incorporated into
and neutrophil differential counts and to identify the membranes of phagocytic cells only. We used in vivo
labeled cells in fresh smears by fluorescence and infusion because it avoids the isolation and cx vivo label-
differential interference contrast (DIC) microscopy. ing steps, potentially making the method more efficient.
Infusion of PKH26 had minimal and transient physi- We infused the fluorescent phagocytic cell linker system
ologic effects on systemic and pulmonary artery in awake, instrumented adult sheep to determine an opti-
pressure, lung lymph flow, and leukocyte counts. mum infusion dose to achieve neutrophil labeling without
Labeled cells were present in venous blood after the causing pathophysiological reactions. We subsequently
infusion for at least 24 h. DIC microscopy observa- performed in vivo labeling experiments in which we evalu-
tion of the blood smears indicated that the fluores- ated matched samples of arterial and venous blood for 24
cently labeled cells were exclusively neutrophils. h to determine how long the labeled neutrophils remained
Unstimulated superoxide anion release from neutro- in the circulation. We also examined frozen tissue samples
phils isolated 2 and 24 h after PKH26 infusion was to determine the organ sites of label clearance and whether
not different from baseline release. Phorbol myris- labeled neutrophils migrated to extravascular sites. Our
tate acetate-stimulated release was not affected results indicate that neutrophils can be selectively labeled
either. The labeled neutrophils responded to in vivo, after which their trafficking can be monitored.
chemoattractants by migrathig to extravascular sites.
our results indicate that neutrophils can be selec- Materials and Methods
lively labeled in vivo, after which trafficking of the
labeled neutrophils can be determined. J. Leukoc. Animal preparation
Riot. 59: 631-638; 1996. We used nine adult male sheep of mixed breed weighing 39.6 ± 4.5 kg
(mean ± SD). We screened the sheep by visual inspection and by mess-

Key Words: neutrophil activation ‘ neutrophil migration . infiam-

urement of circulating leukocyte counts and cell differentials. The ac-

motion . lung injury fluorescence microscopy frozen sections

Introduction Abbreviations: AEC, 3-amino-9-ethylcarbazole; DIC, differential interference

contrast; OD, optical density; PMA, phorbol mynstate acetate; WBC, white blood
Neutrophils play an indispensable role in host defense by cell.
accumulating at sites of bacterial invasion. Pivotal steps in Reprint requests: Kurt H. Albertine, Department of Pediatrics, Children’s
Research Center, University of Utah Health Sciences Center, 50 North Medical
this process are neutrophil recruitment (from the bone
Drive, Salt Lake City, UT 84132-1001.
marrow, circulating, and marginated pools) and migration Received August 31, 1995; revised November 30, 1995; accepted December 1,
to extravascular sites [1]. Previous studies of the kinetics 1995.

Albertine and Gee Labeling of neutrophils in vivo 631

 Arrows: Sbod e. for fr#{149}iradhc& roas
protein ratio were within 10% of their respective average values for the

_____ period.
Once
The
the
steady-state
baseline period
(baseline)
was
period
established,
usually
we
required
infused PKH26
2-3 h.
intra-
I cascara I PKI06 L__5k_ch.motax5 b
arterially to label neutrophils (0.25 x 106 M to 1.50 x i0 M in 50 mL
I- I I I I I I...” I of sterile saline for 1 h; 0.764 mljmin). The diluent for the label had an
0 1 3 5 7 24
endotoxin content of <0.25 EU/mL, according to the manufacturer
(Zynaxis). We continuously measured physiological variables for 3-4 h
Hours
after the infusion (Fig. 1). We collected simultaneous arterial and
a PKH2O nfusion rate - 0.704 mt/men in 50 mt salas infused intra.artariaily. venous blood samples hourly for 4 h for white blood cell (WBC) and
bhd sheep stud#{232}es
only.

Fig. 1 . Diagram of the experimental protocol for the sheep experiments.

40
Matched samples of arterial and venous blood were collected at 30-mm

Downloaded from https://academic.oup.com/jleukbio/article/59/5/631/6979182 by University of Rochester user on 17 January 2024

intervals for the first 4 h after the intra-arterial infusion of PKH26 ended.
0
These samples were used for labeled cell counts, using fluorescence and
(1)
differential interference contrast optics. The arrows indicate the times (1) 30
0
when neutrophils were isolated for superoxide anion release assays. Skin 0.
biopsies were taken at the end ofa 2-h chemotaxis period, using anesthe-
#{149}0
tized sheep. t
a
(1 20

ceptable range in leukocyte counts for our laboratory is 5000-12,000 C

0
cells/tL blood, with 30-45% neutrophils and < 5% eosinophils. We E 10
surgically instrumented all of the sheep to provide chronic vascular
0.
access via catheters in a carotid artery and external jugular vein. We
placed a venous introducer (Subclavian/Jugular Infuset; Sorenson, Salt
Lake City, UT) in the external jugular vein so that the indwelling
polyvinyl catheter could be replaced with a thermodilution catheter on
-100 0 100 200 300 24h
the day of an experiment (see below). Many of the sheep had chronic
0
lung lymph fistulas. The surgical procedures to instrument the sheep
a
with vascular and lymph catheters were performed in an operating suite
C
approved for recovery surgery in large animals, using techniques pre-
0
viously described [7]. Food, but not water, was restricted for 24 h before 0
surgery to minimize bloat. For all invasive procedures, anesthesia was 0.

induced by intravenous injection of 0.5-1.0 g of thiamylal sodium and a

0
maintained by pentobarbital sodium drip (30-35 mg/kg). The sheep
a
were given penicillin (600,000 U intramuscularly) before and once after E
‘I)
surgery. Ringer’s lactate was infused intravenously during surgery (1 L) a
0.
and immediately afterward (1 L). We allowed a 5-day recovery before
initiating an experiment in the awake sheep. Each animal was studied 0.

in one protocol only. Approval for these experiments was received from E
-J
the Animal Resources Committee of Thomas Jefferson University.

Experimental protocol -100 0 100 200 300 24h

25
Awake sheep preparation (six sheep)

On the day of an experiment, we replaced the polyvinyl catheter in the 20

external jugular vein with a flow-directed thermodilution catheter which
we floated into a pulmonary artery (8 French; Abbott Laboratories,
North Chicago, IL). We connected the aortic and thermodilution cathe- ; 15
ters to minitransducers (Cobe CD XIII; North Chicago, IL) to continu-
0.
ously record systemic and pulmonary blood pressures, respectively E
(model 7 physiologic recorder; Grass Instruments, Quincy, MA). We .. 10
0)
referenced the pulmonary arterial and arterial wedge pressures to the C
approximate level of the left atrium, estimated by the chest position of -J

the olecranon. We measured cardiac output in triplicate at 30-mm

intervals (model 3500; KMA, Salt Lake City, UT) by injecting 10 mL of
room temperature 0.9% sterile saline. The cardiac output monitor also
provided a measure of blood temperature. We sampled arterial blood -100 0 100 200 300 24h
every 30 mm to measure blood gases and pH (ABL3O Acid/Base Gas
Analyzer; Radiometer, Copenhagen), total leukocyte counts (HA3 He- Minutes
matology Analyzer; Clay Adams, Parsippany, NJ), leukocyte differen-
Fig. 2. Infusion of 5.5 x 10 M PKH26 for 1 h intra-arterially (hatched
tials (Wright-stained slide smears), and total protein concentration (IS
box) in six sheep caused transient, but not statistically significant (NS),
meter refractometer; AO Scientific Instruments, Keene, NH). We col-
rises in pulmonary artery pressure and lung lymph flow (data are shown
lected lung lymph at 30-mm intervals in tared centrifuge tubes treated
as the mean ± SEM). The lymph-to-plasma total protein concentration ratio
with heparin (100 U).
decreased transiently (NS) during the same period, indicating pressure-
Each experiment began by establishing a steady-state period during
induced increased liquid filtration across pulmonary microvessels. All
which systemic and pulmonary artery pressures, cardiac output, circu-
three variables returned to baseline 24 h after the infusion.
lating leukocyte counts, lung lymph flow, and lymph-to-plasma total

632 Journal of Leukocyte Biology Volume 59, May 1996

 1 ). We ended these experiments 2 h later to observe frozen tissue
sections for emigrated neutrophils.

a. In vitro experiments
0
a)
We assessed the effect of PKH26 labeling on neutrophil function by
comparing superoxide anion release and myeloperoxidase activity for
unlabeled versus labeled neutrophils. The unlabeled neutrophils were
obtained from three of the awake sheep at the end of the 2-h baseline
0.0
period, before PKH26 was infused intra-arterially. We divided the un-
0
labeled neutrophils into two aliquots: one aliquot remained unlabeled,
a whereas the other aliquot of neutrophils was labeled with PKH26 in
C
0 vitro, according to the methods provided by the manufacturer (Zynaxis).
E

Downloaded from https://academic.oup.com/jleukbio/article/59/5/631/6979182 by University of Rochester user on 17 January 2024

We isolated neutrophils from peripheral arterial blood and analyzed
0. I
superoxide anion release and myeloperoxidase activity by methods
-100 0 100 200 300 24h standard to our laboratory [8-10]. Leukocytes were separated from ar-
terial blood by centrifugation and hypotonic lysis of erythrocytes. The
buffy coat pellet was resuspended in autologous plasma and layered in
a 1.0-mL volume of discontinuous Percoll-autologous plasma gradient.
100 -
Neutrophils sedimented only at the 80/70% interface. This gradient
fraction of leukocytes was collected, washed in modified HEPES buffer
(in mM: 140 NaCl, 10 HEPES, 10 KC1, 0.1 CaCI2, 11.0 NaHCO3, 5.0
80 glucose, and 14 jiM albumin, pH 7.4), pelleted in the same buffer, and
resuspended in modified HEPES buffer. We determined cell viability
(trypan blue dye exclusion) and total and differential cell counts. Our
Percoll-plasma isolation procedure minimally alters neutrophil ultras-
tructure [1 1]. The isolated cells are ameboid to circular in profile, they

:: __ ,__.__.__.__. I

110

-100 0 100 200 300 24h C

0
Minutes 0
0
Fig. 3. Neither pulmonary capillary wedge pressure nor aortic pressure
0
changed significantly during or after the 1-h infusion of the fluorescent 0

cell linker PKH26 (hatched box; data are shown as the mean ± SEM; six 0
sheep).
a
0
I-

neutrophil differential cell counts and to identify the labeled cells in

fresh smears by fluorescence and differential interference contrast (DIC) -100 0 100 200 300 24h
microscopy. For fluorescence microscopy, the excitation wavelength was
488 nm (or 551 nm) and the emission wavelength was 567 nm, allowing
the use of standard rhodamine filter sets. We monitored the sheep for
24 h, repeating the physiological measurements during the last 2 h. We 100
then killed the sheep by pentobarbital sodium overdose to collect sam-
ples of lung, liver, and spleen for fluorescence/DIC observation of frozen
tissue sections. : 80
a
C
Anesthetized sheep preparation (three sheep) a) 60

We anesthetized the sheep as described for the awake sheep surgical

procedures. Anesthesia was maintained by intravenous drip administra-
Lion of pentobarbital sodium (35 mg/kg). We also used the same surgical
procedures, except that lung lymph fistulas were not made. We placed
a cuffed endotracheal tube in the trachea to mechanically ventilate the
sheep with room air at a rate of 12-14 breaths/mm, a tidal volume of 15
mL/kg, and positive end-expiratory pressure of 5 cmH2O. We adjusted
ventilation to maintain the partial pressure of carbon dioxide in arterial -100 0 100 200 300 24h
blood (P;iaCO2) between 35 and 45 mmHg. We placed the sheep prone
before beginning the baseline period. We infused PKH26 intra-arteri- Minutes
ally for 1 h.
Fig. 4. Total leukocyte and neutrophil differential cell counts in arterial
We used the anesthetized sheep to evaluate neutrophil migration into
blood were not significantly affected by the 1-h infusion of PKH26
the skin in response to intradermally injected leukotriene B4 (LTB4; 50
(hatched box; data are shown as the mean ± SEM; six sheep).
nmol in 0.2 mL of sterile saline) 2 h after the PKH26 was infused (Fig.

Albertine and Gee Labeling of neutrophils in vivo 633

Downloaded from https://academic.oup.com/jleukbio/article/59/5/631/6979182 by University of Rochester user on 17 January 2024

1996
May
59,
Volume
Biology
Leukocyte
of
Journal
634
 P51125 were incubated for 30 mm at room temperature in the dark before being
25
rinsed with distilled water and counterstained with Mayer’s hematoxylin.
One hundred neutrophils were evaluated for myeloperoxidase staining
20 by a three-point rating system: 0 (no stain; degranulated), + 1 (light
.C
0.
stain; partially degranulated), and +2 (intense stain; normal primary
15 granule content) [9]. The cytochemical distribution of mycloperoxidase
#{149}0
C (as well as other granule enzymes, such as alkaline phosphatase and
C,
acid phosphatase) in the isolated cells was comparable to that (letected
10
in neutrophils in situ in blood vessels [9, 10].

ArtsrIsl blood
11. 5
Vsnous blood
Statistical analysis
Data are presented as the mean ± standard error of the mean. Statistical
0-

Downloaded from https://academic.oup.com/jleukbio/article/59/5/631/6979182 by University of Rochester user on 17 January 2024

analysis of the physiologic data was (lone by analysis of variance and
-100 0 100 300 500 24h
Dunnett’s multiple comparison test to compare baseline steady-state
Minutes
data with the data that were collected after the infusion of PKll26 [15].
Fig. 6. Time course of appearance of fluorescently labeled neutrophils Nonparametric comparisons were made between the median ratings of
in matched smears of arterial and venous blood (six sheep). Labeled neutrophil myeloperoxidase cytochemical activity, using the Wil-
neutrophils were equally distributed in arterial and venous blood (data coxon paired-sample test [15]. Statistical significance was accepted
are shown as the mean ± SEM). Other leukocytes (lymphocytes and as P 0.05.
eosinophils) were not labeled. Blood monocytes and basophils were not
seen in the smears. The number of labeled neutrophils decreased over RESULTS
time, reaching an apparent steady state 4 h after the infusion ended.
Labeled neutrophils were still present in the circulation 24 h later.
Physiological responses to PKH26 in awake sheep
In preliminary studies we optimized the concentration of
PKH26 for intra-arterial infusion in awake sheep. The low-
have short microvillus extensions, their cytoplasm contains numerous
rod-shaped and circular granules, and their multilobated nucleus has est concentration we tested (0.25 x j#{216}6 M) failed to label
both euchromatin and heterochromatin bounded by intact nuclear enve- sufficient numbers of neutrophils (data not shown). The
lope. These ultrastructural features are characteristic of neutrophils highest concentration we tested (1.50 x iO M) caused
fixed in situ in blood vessels. The function of the isolated cells is also
significant rises in systemic arterial, pulmonary arterial,
minimally affected, at least by the parameters we assess. For example,
unstimulated, spontaneous release of superoxide anion is very low (< and pulmonary capillary wedge pressures and lung lymph
0.4 nmol superoxide anion from 106 sheep neutrophils) for at least 2 h flow and significant decreases in cardiac output and oxy-
L12]. genation (data also not shown). A compromise between
The superoxide anion release assay relied on changes in optical labeling adequate numbers of neutrophils without causing
density (OD) detected by a microtiter plate reader (Dynatech MR 600
physiological changes was reached by using a concentra-
[8, 13]). Unlabeled neutrophils and those labeled with PKH26 in vivo
or in vitro were allowed to settle for 15 mm in microtiter wells (Nunc,
tion of 5.5 x 106 M PKH26.
high affinity; 250,000 neutrophils in 50 pit of HEPES buffer per well) Intra-arterial infusion of 5.5 x 1O M PKH2#{212} for 1 h
before the measurement of superoxide anion release. Test solutions caused transient, but not statistically significant, rises in
containing from 10.11 to i0 M phorbol myristate acetate (PMA; Sigma) pulmonary artery pressure and lung lymph flow (Fig. 2).
and 0.32 mM femcytochrome c (type III from horse heart; Sigma) were
Both transient rises coincided with a transient decrease in
applied in a 50-pL volume to test wells. Three wells received ferricyto-
chrome c and HEPES buffer, without PMA, to measure basal release of
the lymph-to-plasma total protein concentration ratio, mdi-
superoxide anion. One assay blank well that contained femcytochrome cating a hydrostatic pressure-induced increase in liquid
c, HEPES buffer, and 700 U superoxide dismutase (bovine erythrocytes, filtration across the pulmonary microvessels [16]. Pulmo-
Calbiochem) to confirm that ferricytochrome c reduction was inhibitable nary capillary wedge and aortic pressures also did not
by superoxide dismutase. Changes in OD were measured intermittently
change significantly during or after the 1 h infusion of
for 60 mm, using a test wavelength of 550 nm and a reference wave-
length of 670 nm. Superoxide anion release was calculated using the
PKH26 (Fig. 3). Likewise, total leukocyte and neutrophil
conversion proposed in [14]: nmol O2 [(test OD - reference OD) x differential cell counts in arterial blood were not sigitifi-
100]/6.3. cantly altered by the 1-h infusion of PKH26 (Fig. 4).
Myeloperoxidase, a constituent of neutrophil primary granules, was Differential leukocyte counts during the 24-h study period
localized cytochemically by placing smears of unlabeled neutrophils
for the six awake sheep averaged 44 ± 1 1% for neutrophils,
and those labeled with PKH26 in vivo or in vitro in a solution of
3-amino-9-ethylcarbazole (AEC) dissolved in N,N-dimethylformamide
2 ± 2% for juvenile neutrophils, 53 ± 12% for lympho-
and sodium acetate (vol/vol 2.5 mL and 5.5 mL, respectively), to cytes, and 1 ± 1% for eosinophils. We did not see mono-
which was added 25 j.tL of3O% hydrogen peroxide [9]. The preparations cytes in any of the blood smears. The 1-h infusion of

Fig.5. Appearance of peripheral blood smears and frozen tissue sections following infusion of PKIl26 in sheep. (A) Simultaneous fluorescence and
DIC micrograph of a pair of PKH26-labeled neutrophils (arrows) in a venous blood smear 30 ruin after the infusion en(led. The arrowhead identifies a
lymphocyte that is not labeled. (B) Fluorescence micrograph ofa labeled neutrophil (arrow) in a venous blood smear prepared from blood that was taken
8 h afterthe infusion ended. Red blood cells are notlabeled. (C-E) Fluorescence micrographs ofPKH26 uptake in the lung, liver, and spleen, respectively.
The fluorescent staining pattern is consistent with uptake by reticuloendothelial cells (RE) in each organ, such as intravascular macrophages in lung
microvessels, Kupffer cells in the liver, and RE cells in the spleen. (F) Simultaneous fluorescence and DIC micrograph of a labeled neutrophil (arrow)
in the dermis of a frozen section of skin that was taken 2 h after an intradermal injection of LTB4 (50 nmol; 0.2 mL; H designates a hair follicle).

Albertine and Gee Labeling of neutrophils in who 635

 TABLE 1. Effect of Ititra-arterial Infusion of PKH26 on Superoxide Anion by routine histology because the fluorescent signal was
Release by Peripheral Blood Neutrophils
quenched during the processing steps. Fluorescence and
DIC microscopy also were used together to identify the
Period Superoxide anion release (nmol/45min)
labeled neutrophils in frozen tissue sections of the dermis
- Unstimulated iO M
taken from the anesthetized sheep 2 h after the intradermal
Baseline 0.44 ± 0.27 5.37 ± 0.48 injection of LTB4 (0.2 mL of a SO nmol solution). Figure
2 Ii after infusion 0.16 ± 0.15 4.54 ± 0.57 5F shows that the labeled neutrophils (arrow) migrated to
extravascular sites in the dermis in response to LTB4.
24 h after infusion 0.41 ± 0.17 5.75 ± 0.58
The superoxide anion release characteristics of periph-
Data are the iiwan ± SEM for four sheep.
eral blood neutrophils that we isolated from four of the six
awake sheep are summarized in Table 1 (cells were not

Downloaded from https://academic.oup.com/jleukbio/article/59/5/631/6979182 by University of Rochester user on 17 January 2024

PKH26 also did not change blood temperature (baseline isolated from the first two awake sheep). Both unstimulated
40.1 ± 0.3#{176}C;3-4 h after the infusion of PKH26 40.0 ± and PMA-stimulated superoxide anion release by the iso-
0.3#{176}C;not significant). Each of the physiological variables lated, adherent neutrophils were unaffected by the infusion
that we measured 24 h after the infusion of PKH26 was the of PKH26. However, based on our estimate of the labeling
same as during the previous day’s baseline period. efficiency (0.1 to 0.5% of the circulating neutrophils were
The appearance of peripheral blood smears and tissue fluorescently labeled), the number of fluorescently labeled
sections following the infusion of PKH26 is shown in Fig- neutrophils may have been too small to accurately detect
ure 5. We used a combination of fluorescence and DIC PKH26-induced changes in superoxide anion release.
microscopy to quantify the number of fluorescent cells and Therefore, we performed in vitro experiments to determine
identify their cell type, respectively. For example, Figure first whether PKH26 labeling caused superoxide anion and
5, A and B show fluorescently labeled neutrophils (arrows) myeloperoxidase release, and second whether the release
in blood smears at 30 mm and 8 h, respectively, after the characteristics of labeled neutrophils were different from
infusion of PKH2#{212}. Simultaneous observation of the speci- those of unlabeled cells.
mens by fluorescence and DIC microscopy (Fig. 5A) al- Labeling sheep neutrophils in vitro with PKH26 (100%
lowed identification of the specific cell type labeled with labeling) did not affect maximal superoxide anion release
PKH26. Neutrophils were identified by their multilobed (1O M PMA) compared with that of unlabeled neutrophils
nucleus, which was visible in negative relief. Nonphago- (Fig. 7). Furthermore, the neutrophils that were labeled
cytic cells, such as lymphocytes (arrowhead in Fig. 5A) and stimulated in vitro had the same maximal release of
and red blood cells (Fig. SB), were not fluorescent. We did superoxide anion as neutrophils that were isolated 2 h after
not see blood monocytes or eosinophils in the smears at the PKH26 was infused in vivo (0.1-0.5% labeling; Fig. 7).
time points examined.
The persistence of the fluorescently labeled neutrophils . Reagents + PKH26

in matched smears of arterial and venous blood of the six 0 Reagents

C
awake sheep is summarized graphically in Figure 6. The 0 PKH26-Iabeled neutrophils
.

main points of Figure 6 are that the labeled neutrophils to 4 #{149} 0 Unlabeled neutrophils
were equally distributed in arterial and venous blood and (n=3)
z
the number of labeled neutrophils decreased over time,
reaching an apparent steady state 4 h after the infusion of
PKH26 ended. Labeled neutrophils remained detectable
in the
of PKH26
neutrophils,
fluorescent
circulation
in vivo
based
neutrophils
24 h later.
labeled
on the ratio of fluorescent
in the slide smears.
0.1
We estimate
to 0.5%
that
of the
the infusion
circulating
versus non-
a.
0
,
3
2
Ii
Figure 5, C, D, and E show that fluorescence was dem-
:2
onstrated in the lung, liver, and spleen, respectively. The x
0
fluorescent staining pattern was consistent with uptake by I-

reticuloendothelial cells in each organ, such as intravascu-

lar macrophages in lung microvessels, Kupffer cells in the a.
0
(I)
o.ilo oilioo oJiIo
liver, and reticuloendothelial cells in the spleen. The fluo- .2! .

rescent cells generally were elongated and their nucleus 0 Unstim -11 -9 -7
E
was not lobated when observed by DIC microscopy. These C Phorbol myristate acetate (M)
morphologic characteristics suggested that the labeled
Fig. 7. Comparison of superoxide anion release between adherent neu-
cells were not neutrophils. However, these characteristics
trophils that were labeled in vitro and assayed for 1 h (solid bar) or were
were not always clear in the frozen tissue sections, leaving labeled in vivo and isolated 2 h later(stippled bar). Responses to different
open the possibility that some of the fluorescent cells in concentrations of phorbol myristate acetate are shown. PKH26 did not
the frozen tissue sections were neutrophils that seques- affect superoxide anion release (data are shown as the mean ± SEM for
three experiments).
tered in the vasculature. We did not verify the cell types

636 Journal of Leukocyte Biology Volume 59, May 1996

 A. EffeCt of adding PKH26 to assay wells
cells are not labeled. We did not see labeled eosinophils
or monocytes in the peripheral blood smears so we cannot
z
0.
determine whether their labeling efficiency is different
from that of neutrophils. We did not expect to see eosino-
phils because we screen our sheep for eosinophils; if they

j
have 5% or more eosinophils we reject them on the as-
-0-- Reagents. unstimuisted
-..- Reagents + P10126. unetimulstsd sumption that they have a parasitic infection. Eosinophils
-0-- Reagents, S PSA in the sheep we used had a differential count 1%. Mono-
-.- Reagents + PKH2S. I0
Q253)
cytes normally represent < 2% of the leukocyte differen-
tial in sheep [17], so it was not unexpected that we did not
see them in the slide smears of peripheral arterial blood

Downloaded from https://academic.oup.com/jleukbio/article/59/5/631/6979182 by University of Rochester user on 17 January 2024

0 20 40 60 80 from the sheep used for this study.
We tested PKH26 as a potential label because this fluo-
rescent probe is selectively taken up by phagocytic cells.
Therefore, PKH26 is potentially well suited for labeling
neutrophils in vivo. This cell linking compound is highly
0
B. Effect of adding iabeled PMNs to assay welts aliphatic and therefore incorporated into the lipid bilayer
z of cytoplasmic membranes. Its incorporation into the lipid
0.
moiety of membranes does not interfere with the protein
S
a
molecules that are in the membrane, according to the
manufacturer. Selective uptake by phagocytic cells is fa-
0
C
-0-- Unlabeled PMN. unstimuisted cilitated by the lipophilic property of PKH26: it forms
C
-.-- PKH261ebeled PMN unstimuietsd
. micelles in an aqueous medium. The micelles are engulfed
-0-#{149} Unisbeled PMNs, 10 N PMA
11
0 -a-. PKH26.iabeled PMN5, 10 U PMA by phagocytic cells. Once incorporated into cellular mem-
a 1 (n=3)
branes of phagocytic cells, PKH26 does not escape from
0
0
0 the labeled cells or transfer to nonlabeled cells. The stabil-
0
E 0 ity of this and related fluorescent cell linkers is consider-
C 0 40 60 80
able. For example, peritoneal macrophages remained
Minutes
irreversibly labeled for at least 49 days in vivo [18] and
Comparison of the time course of unstimulated versus i0 M
erythrocytes remained labeled for at least 60 days in vivo
phorbol myristate acetate-stimulated superoxide anion release between
[6]. In our study, peripheral blood neutrophils are labeled
adherent neutrophils that were labeled in vitro for 1 h (A) or were labeled
in vivo and isolated 2 h later (B). PKH26 did not affect the time course for at least 24 h. One might expect few labeled neutrophils
of superoxide anion release (data are shown as the mean ± SEM for three to persist in the circulation beyond 24 h, given that their
experiments). circulation half-time in humans is estimated to be 6-8 h
[1].
Myeloperoxidase cytochemical activity of neutrophils was
An advantage of the in vivo infusion labeling procedure
also unaffected by PKH26 in vitro. The median rating of
is that it eliminates the need to isolate neutrophils prior to
myeloperoxidase cytochemical activity for PKH26-labeled
labeling [19]. The in vivo infusion labeling procedure does
and unlabeled neutrophils was 1.76 and 1.79, respectively
not affect cell morphology, function, lifetime, or migration.
(n 4). Finally, labeling the neutrophils with PKH26 in
In these respects, the infusion labeling procedure using
vitro also did not affect the time course of superoxide anion
PKH26 is consistent with ex vivo labeling procedures that
release by neutrophils that were unstimulated or stimu-
use related fluorescent lipophilic probes [6, 18-22]. We
lated with 10 M PMA for 1 h (Fig. 8). The same time
also attempted to label sheep neutrophils ex vivo by with-
course of superoxide anion release was observed for neu-
drawing blood into sterile syringes that contained PKH26.
trophils that were labeled in vivo and then isolated for in
However, very few fluorescently labeled neutrophils were
vitro study (Fig. 8).
detected in the circulation (data not shown).
We infused PKH26 intra-arterially, rather than intrave-
DISCUSSION nously, to avoid the rapid clearance of the phagocytic cell
label by pulmonary intravascular macrophages before the
We developed a novel method of labeling circulating neu- neutrophils are labeled. These phagocytic cells reside in
trophils in vivo. This method allows selective fluorescent the pulmonary circulation of sheep and several other spe-
labeling of neutrophils with minimal perturbation to the cies (reviewed by Staub [23] and the references therein).
cells or experimental animal, and it avoids radioisotopes. Upon phagocytosis of blood-borne particulates, the pulmo-
The slow, continuous infusion appears to label only neutro- nary intravascular macrophages release mediators of in-
phils in the circulating blood. The labeled neutrophils per- flammation (e.g., thromboxane) that cause pulmonary
sist in the circulation for up to 24 h after the infusion. We vascular and airway constriction [24, 25]. Both physiologi-
estimate that about 0.1-0.5% of the circulating neutrophils cal responses can be minimized, or even avoided, by intra-
are labeled by this method. Lymphocytes and red blood arterial infusion, which enables the foreign particulates to

Albertine and Gee Labeling of neutrophils in vivo 637

pass through at least one systemic capillary bed before REFERENCES
reaching the pulmonary circulation. Therefore, a lower
concentration of the blood-borne particulates reaches the
1. Cartwright, G. E., Athens, J. W., Wintrobe, M. M. (1964) The kinetics of
pulmonary intravascular macrophages than when an intra- granulocytopoiesis in normal man. Blood 24, 780-803.
venous infusion protocol is used. As our results show, the 2. Athens, J. W., Rabb, S. 0., Hash, 0. P., Mauer, A. M., Ashenbrucker, H.,
physiological responses to the intra-arterial infusion of Cartwright, G. E., Wintrobe, M. M. (1961a) Leukokinetic studies. HI.
Distribution of granulocytes in the blood of normal subjects. J. Gun. invest.
PKH26 are minimal and transient, suggesting that the fluo- 40, 159-164.
rescent cell linker had little inflammatory activity in either 3. Athens, J. W., Haab, 0. P., Rabb, S. 0., Mauer, A. M., Ashenbrucker, H.,
Cartwright, G. E., Wintrobe, M. M. (1961b) Leukokinetic studies. IV. The
awake or anesthetized sheep. This attribute is important for total blood, circulating and marginal granulocyte pools and the granulocyte
studies that are designed to track neutrophils in vivo. turnover rate in normal subjects. J. Clin. invest. 40, 989-995.
4. Butcher, E. C., Weissman, I. L. (1980) Direct fluorescent labeling of cells
Some of the intra-arterially infused PKH26 is nonethe- with fluorescein or rhodamine isothiocyanate. 1. Technical aspects. J. immu-

Downloaded from https://academic.oup.com/jleukbio/article/59/5/631/6979182 by University of Rochester user on 17 January 2024

less cleared by components of the reticuloendothelial sys- aol. Meth. 37, 97-108.
5. Jacobs, R. M., Boyce, J. T., Kociba, G. J. (1986) Flow cytometric and
tem, which in sheep includes the pulmonary intravascular radioisotopic determinations of platelet survival time in normal cats and
macrophages in the lung, Kupffer cells in the liver, and feline leukemia virus-infected. Cytomeuy 7, 64-69.
6. Slezak, S. E., Horan, P. K. (1989) Fluorescent in vivo tracking of hematopoie-
reticuloendothelial cells in the spleen. All three organs
tic cells. Part I. Technicalconsiderations. Blood 74, 2172-2177.
demonstrated fluorescence in fresh frozen tissue sections. 7. Albertine, K. H., Gee, M. H. (1989) Sheep lung lymph fistula preparation
The greatest uptake appeared to be the lung, as has been applied to measurements ofcomplement-mediated pulmonary microvascular
injury. In Handbook ofAnimal MOdels ofPulmonary Disease (J. 0. Cantor,
shown for other tracer molecules [24-26]. ed) CRC Press, Boca Raton, FL, 209-229.
Release of superoxide anion and proteolytic enzymes 8. McKenna, P. J., Rosolia, D. L., Ishihara, Y., Albertine, K. H., Staub, N. C.,
Gee, M. H. (1992) Down regulation of blood and hone marrow neutrophils
(such as myeloperoxidase) by activated neutrophils is nec- decreases expression of acute lung injury in sheep. Am. J. Physiol. 263,
essary for bacterial killing and contributes to the incidental H1492-H1498.
9. Rosolia, D. L., McKenna, P. J., Gee, M. H., Albertine, K. H. (1992) Infusion
tissue injury that is associated with robust acute inflamma-
ofzymosan-activated plasma affects neutrophils in peripheral blood and bone
tion (reviewed by Gee and Albertine [27]). Therefore, we marrow in sheep. J. Leukoc. Biol. 52, 501-515.
10. Albertine, K. H., Rosolia, D. L., Schuhl, R. A., Peters, S. P., Gee, M. H.
assessed the release of superoxide anion and myeloperoxi-
(1993) Physical and cytochemical properties of neutrophils activated in situ
dase by sheep neutrophils as indices of neutrophil killing in the lung during ZAP infusion in sheep. J. Appi. Physiol. 74, 1361-1373.
activity. The assays we used are ones that are reliable and 11. Albertine, K. H., Cerasoli, F. J., Gee, M. H., Ishihara, Y., Tahamont, M. V.,
Peters, S. P. (1988) Morphological analysis of the activation of adherent
reproducible for sheep and human neutrophils in situ in neutrophils in vitro. Tissue Cell 20, 519-530.
the lung and in vitro [8-12, 28]. Our results show that 12. Cerasoli, F., Jr. Gee, M., Ishihara, Y., Albertine, K., Tahamont, M., Peters,
S. (1988) Biochemical analysis of the activation of adherent neutrophils in
PKH26 does not stimulate release of either inflammatory vitro. Tissue Cell 20, 505-517.
mediator from sheep neutrophils. Whether earlier re- 13. Albertine, K. H., Cerasoli, F., Tahamont, M. V., Ishihara, Y., Flynn, J. T.,
Peters, S. P., Gee, G. H. (1989) Zymosan activated plasma causes prolonged
sponses to stimulation, such as up-regulation of 2-inte-
decreases in PMN superoxide release in sheep. J. Appl. Physiol. 67,
grins, may be triggered by PKH26 was not determined in 2481-2490.
14. Pick, E., Mizel,
D. (1981) Rapid microassays for the measurement of
our study.
superoxide hydrogen
and peroxide production by macrophages in culture
In summary, we devised an in vivo infusion method to using an automatic enzyme immunoassay reader. J. immunol. Meth. 46,
label neutrophils. The disadvantages of this method in- 211-226.
15. Zar, J. H. (1984) Biosta.tis:ical Analysis, 2nd ed Prentice-Hall, Englewood
dude (1) the uptake of unbound label by reticuloen- Cliffs, NJ, 122-205.
dothelial cells in the lung, liver, and spleen; (2) the small 16. Staub, N. C. (1974) Pulmonary edema. Physiol. Rev. 54,678-811.

number of neutrophils that become labeled; and (3) the 17. Jam, N. C. (1986) Schwalm’s Veterinary Hematology. Lea & Febiger, Phila-
delphia, 208-224.
requirement for fluorescence and DIC observation to iden- 18. Melnicoff, M. J., Horan, P. K., Breslin, E. W., Morahan, P. S. (1988b)
tify the labeled cells as neutrophils, especially in frozen Maintenance of peritoneal macrophages in the steady state. J. Leukoc. Biol.
44, 367-375.
tissue sections. The advantages of this method, on the other 19. Yuan, Y., Fleming, B. P. (1990) A method for isolation and fluorescent
hand, are (1) the elimination of the isolation and ex vivo labeling of rat neutrophils for intravital microvascular studies. Microva.sc.
Res. 40, 218-229.
labeling steps that require subsequent injection of the la- 20. Horan, P. K., Slezak, S. (1989) Stable cell membrane labeling. Nature 340,
beled cells, (2) selective labeling of neutrophils among the 167-168.
21. Horan, P. K., Slezak, S. (1989) Fluorescent in vivo tracking of hematopoietic
circulating white blood cells, (3) mild and transient physi-
cells. FASEBJ. 3, A274.
ological responses, and (4) avoidance of radioisotopes. We 22. Melnicoff, M. J., Morahan, P. S., Jensen, B. D., Breslin, E. W., Horan, P. K.
believe that these advantages offer a new approach to (1988a) In vivo labeling ofresident peritoneal macrophages. J. Leukoc. Biol.
43, 387-397.
studying the circulatory and migratory kinetics of neutro- 23. Staub, N. C. (1989) The Pulmonary intravascular Macrophage. Futura
phils normally and during inflammation. Publishing, Mt. Kisco, NY.
24. Albertine, K. H., Staub, N. C. (1986) Vascular tracers alter hemodynamics
and airway pressure in anesthetized sheep. Microvasc. Res. 32, 279-288.
ACKNOWLEDGMENTS 25. Miyamoto, K., Schultz, E., Heath,T., Mitchell, M. D., Albertine, K. H., Staub,
N. C. (1988) Pulmonary intravascular macrophages and hemodynamic ef-
fects of liposomes in sheep. J. Appl. Physiol. 64, 1143-1152.
The authors extend their gratitude to David Rosolia, Rob 26. Warner, A. E., Brain, J. D. (1990) The cell biology and pathogenic role of
pulmonary intravascular macrophages. Am. J. Physiol. 258, L1-L12.
Schuhl, and Jill Fisher for technical assistance. We also
27. Gee, M. H., Albertine, K. H. (1993) Neutrophil-endothelial cell interaction
thank Stacey E. Yarnall, Sue E. Slezak, and Paul K. Horan in the lung. Annu. Rev. Physiol. 55, 227-248.
of Zynaxis Cell Science for supplying PKH26. Supported 28. Peters, S. P., Cerasoli, F., Albertine, K. H., Gee, M. H., Berd, D., Ishihara,
Y. (1990) “Autoregulation” ofhuman neutrophil activation in vitro: regulation
by National Institutes of Health grants HL38075, of phorbol myristate acetate-induced neutrophil activity by cell density. J.
HL36237, and Zynaxis Cell Science. Leukoc. Biol. 47, 457-474.

638 Journal of Leukocyte Biology Volume 59, May 1996