Brief Conclusive Report

Exocytosis of azurophil and arginase

1-containing granules by activated
polymorphonuclear neutrophils is required
to inhibit T lymphocyte proliferation
Rita Rotondo,* Maria Bertolotto,† Gaia Barisione,‡ Simonetta Astigiano,‡
Susanna Mandruzzato,§ Luciano Ottonello,† Franco Dallegri,†
Vincenzo Bronte,ⱍⱍ Silvano Ferrini,‡ and Ottavia Barbieri*,‡,1
Departments of *Experimental Medicine and †Internal Medicine, University of Genova, Italy; ‡Department of Translational
Oncology, National Institute for Cancer Research, Genova, Italy; §Department of Oncological and Surgical Sciences, University
of Padova, Italy; and ⱍⱍUniversity Hospital and Department of Pathology, Verona, Italy
RECEIVED NOVEMBER 12, 2009; REVISED JANUARY 26, 2011; ACCEPTED JANUARY 26, 2011; DOI: 10.1189/jlb.1109737

ABSTRACT thine and urea. By depleting the extracellular milieu of l-argi-

ARG1, expressed by human PMNs, inhibits T cell prolif-
eration by depleting extracellular L-arginine. Here, we
report that ARG1, released from gelatinase granules by
PMNs, is inactive at physiological pH unless activated
by factor(s) stored in azurophil granules. Whereas
ARG1 exocytosis was induced by TNF-␣ or ionomycin,
only the latter mediated the release of both granules,
resulting in extracellular ARG enzyme activity at physio-
logical pH. Furthermore, after fractionation of the differ-
ent classes of granules, only the mixture of gelatinase
and azurophil granules resulted in ARG1 activity at
physiological pH. The use of protease inhibitors indi-
cated the involvement of a PMSF- and leupeptin-sus-
ceptible serine protease in ARG1 processing and acti-
vation. Finally, the supernatant of viable PMNs under-
going frustrated phagocytosis, which mediates
gelatinase and azurophil granule release, inhibited T
cell proliferation through ARG-dependent mechanisms.
In vivo, high ARG1 concentrations and increased ARG
enzyme activity, sufficient to inhibit T cell proliferation,
were observed in synovial fluids from RA. These find-
ings suggest that PMNs, recruited at sites of immune
complex deposition, induce ARG1-dependent immune
suppression through concomitant exocytosis of gelati-
nase and azurophil granules. J. Leukoc. Biol. 89:
721–727; 2011.
Introduction
Effective immune suppression can result from the activity of
ARG, the enzyme-catalyzing l-arginine degradation to orni-
Abbreviations: ARG⫽arginase, nor-NOHA⫽N-hydroxy-nor-L-arginine,
RA⫽rheumatoid arthritis
The online version of this paper, found at www.jleukbio.org, includes
supplemental information.
nine, ARG down-regulates the expression of the CD3␨ chain in
T lymphocytes and induces inhibition of T cell proliferation
[1]. In humans, enzyme isoform 1 (ARG1) is stored within the
intracellular granules of PMNs [2], and its exocytosis can be
induced by TNF-␣ or IL-8 [3].
On these bases, it was advanced that immune suppression
could result from exocytosis of ARG1-containing granules by
intact PMNs [4]; however, this hypothesis raises some crucial
issues.
First, inhibition of T cell proliferation was investigated only
in models mimicking PMN massive death typical of purulent
inflammatory reaction [5] but not during regulated exocytosis
by viable PMNs, which is a finely tuned phenomenon. Thus, a
clear demonstration of inhibition of T cell proliferation by
PMNs through ARG1 exocytosis is still missing.
Second, it was described that ARG1 is stored in PMN gelati-
nase granules as an inactive molecule [4], whereas the enzyme
released upon PMN cell lysis effectively catabolizes l-arginine
[5]. Therefore, it is still unclear whether exocytosed ARG1 can
effectively catabolize arginine.
Finally, ARG1 catalytic activity largely depends on pH. In-
deed, the curves of pH dependency for this enzyme show that
ARG1 enzymatic activity is maximal in a strong alkaline envi-
ronment (pH 9.5–10.5) and negligible at neutral pH [6, 7]. It
is therefore unclear how the enzyme, exocytosed or released
upon cell lysis, could be effective under physiological condi-
tions and cause immune suppression.
To elucidate these points, we investigate occurrence and
mechanisms of ARG1-dependent immune suppression follow-
ing regulated exocytosis of granule content by viable human
PMNs. We demonstrate that factor(s) stored in azurophil gran-
1. Correspondence: Embryogenesis and Carcinogenesis in Animal Models,
National Institute for Cancer Research, Largo Rosanna Benzi 10, 16132
Genova, Italy. E-mail: ottavia.barbieri@istge.it

0741-5400/11/0089-721 © Society for Leukocyte Biology Volume 89, May 2011 Journal of Leukocyte Biology 721
ules enable ARG1, released from gelatinase granules, to catab-
olize l-arginine effectively at physiological pH and that activity
of serine-protease(s) is involved in ARG1 activation. We also
show that exocytosis of these two classes of granules induces
ARG1-dependent inhibition of T cell proliferation. Finally, we
provide evidence that immune suppression via ARG1 released
by PMNs might occur in RA joints.
MATERIALS AND METHODS
Cell isolation
Blood was collected from healthy volunteers, after informed consent, in
accordance with the Ethics Committee of the National Institute for Cancer
Research (Genova, Italy) and Declaration of Helsinki. PMNs were isolated
by dextran sedimentation, Ficoll separation, and hypotonic lysis of erythro-
cytes [8]. Nonadherent mononuclear cells (PBLs) were also collected. Cell
viability was measured by the LDHe kit (Aniara, Mason, OH, USA).
Patients
Synovial fluids were obtained after informed consent from affected knees
of six patients with RA. The mean duration of disease was 3 ⫾ 1 years. Pa-
tients—three males and three females, age 51 ⫾ 8 years— had not received
any intra-articular therapy during the 5 weeks before sample collection.
Synovial fluid was also obtained from six patients with gonoarthrosis. Syno-
vial fluids were centrifuged at 1000 g for 15 min, and the supernatants
were removed aseptically and stored at –20°C in aliquots. Prior of ARG ac-
tivity, evaluation samples underwent a further centrifugation at 16,000 g for
10 min at 4°C.
ELISA for ARG1
ARG1 protein was measured in supernatant by using the Arginase I ELISA
kit (BioVendor, Brno, Czeck Republic). Absorbance at 450 nm was evalu-
ated by a spectrophotometer, and ARG1 concentration was calculated on
the basis of a standard curve.
Evaluation of ARG activity
ARG activity was measured by spectrophotometry. To assay cell lysates, 4 ⫻
106 PMNs in 200 ␮l buffer (150 mM NaCl, 10 mM Tris/HCl, at the indi-
cated pH) underwent seven cycles of freezing and thawing. In some tests,
PMSF, pepstatin A, aprotinin, leupeptin bestatin, phosphoramidon, or cal-
pain inhibitor (all from Sigma-Aldrich, St. Louis, MO, USA) was added. To
assay supernatants, in some cases, 4 ⫻ 106 PMNs were resuspended in 200
␮l l-arginine-free DMEM (Sigma-Aldrich) at the indicated pH, incubated
for 30 min at 37°C in the presence or absence of TNF-␣ (PeproTech, Lon-
don, UK) or ionomycin (Serva, Heidelberg, Germany), and then superna-
tant was collected; in other cases, 2 ⫻ 106 PMNs were resuspended in 200
␮l l-arginine-free DMEM and underwent frustrated phagocytosis for 30
min at 37°C, as described below, and afterward, supernatant was collected.
To assay synovial fluids, 100 ␮l of each sample was added with 100 ␮l of
the above-mentioned buffer. All samples were then supplemented with 250
mM l-arginine and 100 ␮M MnCl2 and incubated at 37°C for 2 h, and
urea production was measured as described [9]. As preliminary analyses
showed that synovial fluids contained urea, each of them was measured
against the same sample not supplemented with l-arginine and MnCl2,
used as “blank.” In some cases, results were controlled by mass spectrome-
try, as described [10]. Results from the two techniques coincided; thus,
only results from spectrophotometry are shown herein.
Subcellular fractionation
We followed a technique described previously, with some changes [11].
NaCl (150 mM) was used as medium without protease inhibitors; all proce-
dures were performed at 4°C or in ice bath. After nitrogen cavitation, cell
722 Journal of Leukocyte Biology Volume 89, May 2011
lysates were centrifugated on three-layer Percoll gradient, and granule-con-
taining bands were collected. All fractions underwent four cycles of freeze
and thawing. To assess the effectiveness of fractionation, each band was
assayed for ARG, MPO, and lactoferrin.
MPO and lactoferrin detection
MPO and lactoferrin were measured following techniques described previ-
ously [12].
Briefly, MPO activity was determined using 0.167 mg/ml O-dianisidine
(Sigma-Aldrich) and 0.1 mM hydrogen peroxide in 50 mM phosphate buf-
fer (pH 6.0). One unit of enzyme activity was defined as that oxidizing 1
␮mol O-dianisidine/min/25°C.
Lactoferrin concentration was measured using an ELISA sandwich assay
based on microtiter plates coated with rabbit anti-human lactoferrin IgG
(United States Biochemical, Cleveland, OH, USA) and peroxidase-conju-
gated rabbit anti-human lactoferrin (United States Biochemicals).
Western blot analysis
An aliquot of each subcellular fraction was assayed. A total lysate of PMNs,
obtained by five cycles of freeze and thawing, was used as control. Samples
were then resolved under reducing conditions by 10% SDS-PAGE. After
blotting onto nitrocellulose filters (Hybond C-extra, GE Healthcare, Little
Chalfont, UK), blots were stained using mouse monoclonal IgG1 anti-hu-
man ARG1, clone 1.10, or rabbit polyclonal anti-MPO at a 1:1000 dilution
(DakoCytomation, Glostrup, Denmark). After washings, blots were incu-
bated with the appropriate secondary HRP-conjugated antibodies (Dako-
Cytomation). Bands were visualized by the ECL system (GE Healthcare). In
another set of experiments, PMNs were lysed in the presence or absence of
PMSF 2 mM (Sigma-Aldrich), samples were spun at 13,000 RPM for 10
min, and supernatants were collected and incubated for 30 min at 37°C.
Afterward, a Western blot was carried out using goat polyclonal anti-ARG1
at a 1:1000 dilution (Santa Cruz Biotechnology, Santa Cruz, CA, USA) or
the above-mentioned anti-MPO antiserum. In this case, a total cell lysate
obtained from normal human liver was used as control.
Frustrated phagocytosis
A described previously, technique was followed, with some changes [13]. A
5-mg/ml solution of human IgG (Ig Vena N, Sclavo, Siena, Italy) was incu-
bated for 2 h in 96-well tissue-culture plates. After washing, 2 ⫻ 106 PMNs
in 200 ␮l l-arginine-free DMEM were added to untreated or IgG-coated
plates. After 30 min at 37°C, supernatant was collected.
T cell proliferation
Supernatants of PMNs that underwent frustrated phagocytosis were col-
lected and supplemented with B27 (Gibco Invitrogen, Carlsbad, CA, USA)
at a 1:50 dilution, 100 ␮M l-arginine, 9 ng/ml MnCl2, and in some in-
stances, nor-NOHA (Inalco, Milan, Italy). Afterward, autologous PBLs were
added, incubated for 48 h to starve PBL of intracellular l-arginine, and
then seeded at 5 ⫻ 104 cells/well in anti-CD3-coated wells. Coating was
performed by incubating 96-well tissue-culture plates with 50 ␮l/well 500
␮g/ml UCHT1 mAb for 1 h at 37°C, followed by washing. Five days after,
0.5 ␮Ci 3H-thymidine (GE Healthcare) was added, and incorporation was
measured after 24 h.
RESULTS AND DISCUSSION
Activation of ARG1 at physiological pH requires
concomitant release of gelatinase and azurophil
granules
Previous data indicated that optimal pH for ARG1 from hu-
man liver ranged between 9.5 and 10.5 [6, 7], and our prelim-
inary titration experiments on ARG activity in PMN lysates
www.jleukbio.org

confirmed these findings (Supplemental Fig. 1). As human
PMNs do not express ARG2 [3], ARG1 accounts for all of the
ARG activity. We then addressed the relationship between pH
and enzyme activity in lysates and supernatants of PMNs, at
pH 9.5 and pH 7.5. When lysates of PMNs were assayed for
their ARG activity, significant urea production was detected at
pH 9.5 and pH 7.5 (Fig. 1A), and the differences in kinetic
curves did not reach statistical significance. Indeed, some
amount of ARG activity was detected at all of the pH values
tested (Supplemental Fig. 1).
ARG activity was then assayed in supernatants from viable
PMNs incubated with degranulating stimuli, such as ionomy-
cin, which mobilizes all types of PMN granules [14], or TNF-␣,
which fails to mobilize azurophil granules [15]. At pH 9.5, a
dose-dependent increase in enzymatic activity was found with
both stimuli (Fig. 1B), whereas at pH 7.5, ARG activity was de-
tected only in supernatants of ionomycin-treated PMNs (Fig.
1C). However, when ARG1 protein was measured by ELISA, it
was evident that both stimuli induced similar levels of ARG1
release in the medium (Fig. 1D). These findings indicate that
ARG1 activity at physiological pH is achieved only in the
presence of PMN intracellular factor(s), which are released
after cell lysis or upon treatment with ionomycin but not
TNF-␣.
As ionomycin but not TNF-␣ mobilizes azurophil granules,
we examined further whether these granules contained fac-
tor(s) contributing to ARG1 activity at physiological pH. To
this end, PMNs were lysed, and intracellular granules were sep-
arated by centrifugation on a discontinuous Percoll gradient in
three fractions: ␣, containing azurophil granules; ␤1, contain-
A B
oles/106 PMNs/hourr)
5000
4000
3000
2000
Urea (nmo
1000
0 0
30 60 120
Incubation time (min)
D
Ionomycin
5 μM
TNF-α
10 μg/ml
Medium
0 40 80 120
ARG1 concentration (µg/ml)
www.jleukbio.org
Rotondo et al. PMN exocytosis inhibits T cell proliferation via ARG1
ing specific granules; and ␤2, containing gelatinase granules.
To verify the achievement of an effective separation of the
three classes of granules, marker enzymes for azurophil
(MPO), specific (lactoferrin), and gelatinase (ARG) granules
were assessed for each fraction (Fig. 2A). ARG activity was de-
tected mainly in the ␤2 fraction at pH 9.5 but not at pH 7.5
(Fig. 2B), confirming that the enzyme is stored in gelatinase
granules. To provide further evidence for the localization of
ARG1 in these granules, Western blot analysis was carried out
on the three granule fractions (Fig. 2C). A 41-kDa band, con-
sistent with the ARG1 MW, was detected by anti-ARG1 anti-
body only in the ␤2 fraction. The occurrence of a second
band of lower MW in granulocyte lysates will be discussed be-
low. MPO protein was present in a large amount only in frac-
tion ␣, and the low level of MPO detected in the remaining
fractions is most likely a result of contamination by primary
granules. When fraction ␤2 was coincubated at pH 7.5 with
one of the other fractions, no significant increase in urea pro-
duction was found with fraction ␤1, whereas strong ARG activ-
ity was induced with fraction ␣ (Fig. 2D). These data confirm
that ARG1 stored in gelatinase granules is active only at alka-
line pH and that the concomitant release of azurophil gran-
ules is required to trigger effective ARG1 activity at physiologi-
cal pH.
A proteolytic step is likely involved in this process, as specific
protease inhibitors reduced ARG activity at pH 7.5 in PMN
lysates (Fig. 3A). The pattern of inhibition indicates the in-
volvement of protease(s) resistant to bestatin, calpain inhibi-
tor, phosphoramidon, pepstatin, and aprotinin but susceptible
to PMSF and leupeptin. These findings make the involvment
C
06 PMNs/hour)
1000
800
800
600
600
Urea (nmoles/10
400
400
200 200
0
Untreatted
ate
TNF-α 2 μg//ml
10 μg//ml
50 μg//ml
μM
μM
μM
Lysa
Ionomycin 1 μ
Lysate
5μ
20 μ
Untreated
Figure 1. Depending on the degranulating stimulus, ARG1 exocytosed by PMNs is ac-
tive or inactive at physiological pH. (A) ARG activity at pH 9.5 (e) and pH 7.5 (Œ) was
evaluated in PMN lysates at the indicated times by measuring urea production by spec-
trophotometry. Mean ⫾ 1 sd; n ⫽ 6. (B) PMNs were incubated for 30 min with TNF-␣,
ionomycin, or medium alone, and supernatants were collected for ARG activity mea-
surement at pH 9.5. For comparison, the amount of ARG activity displayed by lysing
the same amount of PMNs is shown. Mean ⫾ 1 sd; n ⫽ 3. (C) The above-mentioned
supernatants were tested for ARG activity at pH 7.5. Mean ⫾ 1 sd; n ⫽ 3. (D) ARG1
concentration in the supernatant of TNF-␣- or ionomycin-stimulated PMNs was mea-
sured by ELISA. Mean ⫾ 1 sd; n ⫽ 3.
Volume 89, May 2011 Journal of Leukocyte Biology 723
 A 90 B

al enzyme activity
2500

a (nmoles/hour)
80
70 2000
60
Figure 2. Factor(s) stored in azurophil 50 1500
granules are responsible for ARG1 activ- 40
1000
ity at physiological pH. (A) PMN lysate 30

% of tota

Urea
was separated on three-layer Percoll gra- 20 500
dient, and ␣, ␤1, and ␤2 fractions were 10
assayed for ARG, lactoferrin, and MPO. 0 0
ARG and MPO were measured by spec-
trophotometry and lactoferrin by ELISA.
One representative experiment of three ARG Lactoferrin MPO
is shown. (B) ARG activity was measured
on subcellular fractions ␣, ␤1, and ␤2 at C D
pH 7.5 (solid bars) and pH 9.5 (open 41kDa 1750
Arg-1

Urea (nmoles//hour)
bars). Mean ⫾ 1 sd; n ⫽ 3. (C) The pres- 1500
ence of ARG1 and MPO was assessed on 36kDa
1250
␣, ␤1, and ␤2 fractions by Western blot.
1000
Total cell lysate of PMNs was used as con-
750
trol (lane 1). (D) One-hundredth of a ␤2
fraction was incubated with one-hun- 500
dredth of each of the other two fractions, 55kDa 250
MPO
and ARG activity was assayed at pH 7.5. 0
Mean ⫾ 1 sd; n ⫽ 3.
β2

β1
PMNs

Fractions α

of metalloproteases, cystine proteases, and aminopeptidases
unlikely and suggest a role for an aprotinin-resistant but
PMSF- and leupeptin-susceptible serine protease in inducing
ARG1 activity at physiological pH. Indeed, the percentage of
PMSF-induced inhibition of ARG activity in PMN lysates de-
creased with the increase in pH (Supplemental Fig. 1). Fur-
thermore, Western blot analysis performed on PMN lysates
showed two bands, the expected one (41 kDa) and one mi-
grating at 36 kDa (Fig. 3B). The latter band could not be de-
tected when samples were processed in the presence of PMSF,
suggesting that it was the result of a proteolityc cleavage of
ARG1 protein by a PMSF-sensitive serine protease. Studies are
currently performed in our laboratory to identify this protease
and elucidate further ARG1 processing. The finding that the
ARG1 released from gelatinase granules of PMNs is inactive at
physiological pH, unless activated by factor(s) stored in azuro-
phil granules, is in agreement with the biological function of
PMNs. Indeed, the enzymes stored within intracellular gran-
ules of PMNs are usually in the form of inactive proenzymes
that are activated by proteolytic cleavage after being released
[16]. Therefore, the behavior of ARG1 is not surprising, as the
native form of this enzyme is active at pH values never at-
tained within our organism. The necessity for ARG1 to be
cleaved to exert its function at physiological pH might under-
lie a control mechanism preventing accidental activation of
the enzyme and consequent immune suppression. As a matter
of fact, gelatinase granules are easily mobilized, and azurophil
ones are exocytosed only in a small amount and under partic-
ular conditions [17], which makes extracellular ARG1 activa-
tion a finely tuned phenomenon.
Supernatant of PMNs activated with nonphagocytable
immune complexes inhibits T cell proliferation
Simultaneous granule release, also involving azurophil gran-
ules, occurs in vitro when PMNs adhere to a substrate and are
then activated or when they attempt to phagocyte large bodies,
such as aggregated bacteria or opsonized cells [18, 19], a phe-
nomenon called frustrated phagocytosis or reverse endocytosis.
To mimic this process in vitro, PMNs were allowed to adhere
to tissue-culture wells coated with IgG, and supernatant was
collected. Under this setting, ARG activity was detectable in
supernatants at pH 7.5 (Fig. 4A), demonstrating that ARG1
was not only released but also enabled to catalyze l-arginine at
physiological pH. Significant amounts of the azurophil granule
marker MPO were detected in the same supernatants, thus
demonstrating that concurrent exocytosis of this class of gran-
ules occurred together with the release of the factor(s) induc-
ing ARG1 activity at pH 7.5 (Fig. 4B). The presence of small
amounts of ARG1 enzyme, active at pH 7.5 in the supernatant
of untreated PMNs (Figs. 1B–D and 4A), can be explained by
the spontaneous granule release occurring upon adherence of
PMNs to a plastic substrate during incubation [20]. As cell via-
bility, detected by LDH release, was unaffected (data not
shown), l-arginine catabolism was likely the result of exocytosis
of gelatinase and azurophil granules by living PMNs with no
involvment of cell lysis.
It is conceivable that PMNs exocytosing gelatinase and
azurophil granules might induce immune suppression by the
released ARG1. Indeed, supernatants of PMNs undergoing
frustrated phagocytosis inhibited T cell proliferation (Fig. 4C).
To confirm the major role of l-arginine catabolism in this

724 Journal of Leukocyte Biology Volume 89, May 2011 www.jleukbio.org

 Rotondo et al. PMN exocytosis inhibits T cell proliferation via ARG1

A PMSF

Leupeptin
A
600

06PMNs/hour)
Phosphoramidon

Calpain inhibitor

Bestatin
400

Urea
Aprotinin

U
(nmoles/10
Pepstatin
200
0 50 100 150
Urea (% of control)

1 2 3 0
B Medium IgG coating
ARG-1 (41 KDa)
B 5
ARG-1 (36 KDa)

U)
MPO (mU
3
+ -

2
Figure 3. Induction of ARG1 activity at physiological pH requires a
proteolytic step. (A) PMNs were lysed in the presence or absence of
100 ␮M pepstatin A, 100 ␮g/ml aprotinin, 40 ␮M bestatin, 20 ␮M cal- 1
pain inhibitor, 10 ␮M phosphoramidon, 50 ␮g/ml leupeptin, or 2 mM
PMSF. After 30 min, ARG activity was tested at pH 7.5 (solid bars) and
pH 9.5 (open bars). Control without protease inhibitors produced 0
728 ⫾ 143 nmoles urea/106 PMN/h. Mean ⫾ 1 sd; n ⫽ 3. (B) PMNs Medium IgG coating
were lysed in the presence (lane 2) or in the absence (lane 3) of 2
mM PMSF, and then Western blot analysis was performed. Total cell
lysate of human liver was used as control (lane 1). C
PBL
phenomenon, T cell inhibition was reduced significantly in the
presence of nor-NOHA, an ARG-specific inhibitor [21]. PBL+ sup resting PMNs
These results provide the first direct demonstration that via-
ble PMNs can inhibit T cell proliferation via the enzymatic PBL + sup PMNs/IgG
activity of exocytosed ARG1.
PBL + sup PMNs/IgG
ARG activity is higher in synovial fluid from RA than + Nor.NOHA
from gonoarthrosis
0 3000 6000 9000 12000
Synchronous release of azurophil and gelatinase granules oc-
curs in PMNs following activation by immune complexes [22, CPM
23]. If the complexes are distributed along a nonphagocytable
surface, exocytosis is observed; when the complexes are phago- Figure 4. Activated PMNs inhibit T cell proliferation via ARG1 activity.
cytosed, some loss of enzymes from the phagocytic vacuoles (A) Ninety-six well tissue-culture plates were incubated with medium
into the extracellular milieu takes place. This release arises or coated with human IgG. PMNs were then added and supernatant
collected after 30 min. ARG activity was measured at pH 7.5. Mean ⫾
from a momentary opening of the vacuole to allow ingestion, 1 sd; n ⫽ 3. (B) The above-mentioned supernatans were assayed for
and it is particularly evident when the phagocytosed particle is MPO activity by spectrophotometry. Mean ⫾ 1 sd; n ⫽ 3. (C) PBLs
relatively large. It is known that the release of proteolytic en- were incubated with supernatant (sup) of PMNs undergoing frustrated
zymes, stored in azurophil granules by PMNs, activated by im- phagocytosis in IgG-coated tissue-culture plates in the presence or ab-
mune complexes, contributes to tissue injury in autoimmune sence of 10 ␮M nor-NOHA, an ARG1 inhibitor. PBLs alone and with
diseases [23]. This is the case of RA, where Igs and immune the supernatant of unstimulated PMNs were tested as control. After
48 h, necessary for depletion of intracellular ARG, PBLs were activated
complexes are tightly bound to the superficial layers of carti-
with immobilized anti-CD3 mAb, and a standard 3H-thymidine incor-
lage [24], thereby providing a solid surface to facilitate PMN poration test was performed. One representative experiment out of
adherence and activation [25]. Furthermore, particulates the three is shown.
formed by Ig aggregates are present in synovial fluid [26].

www.jleukbio.org Volume 89, May 2011 Journal of Leukocyte Biology 725

These conditions pave the way to extracellular release of sec-
ondary, gelatinase, and azurophil granules. As the inflamed
articular cavity is infiltrated predominantly by PMNs [27], it is
not surprising that the proteolytic enzymes stored in azurophil
granules play a major role in causing tissue-injuring in RA [28,
29]. It is conceivable that together with exocytosis of these en-
zymes, release of the factor(s) activating ARG1 might take
place.
On the basis of these considerations, we measured the en-
zyme concentration and activity in knee synovial fluids from
patients with RA or gonoarthrosis, a noninflammatory joint
disease used as negative control. Even if the difference in
ARG1 concentration between the two groups was not statisti-
cally significant, four out of six RA synovial fluids contained
very high concentrations of ARG1 protein (Fig. 5A), whereas
synovial fluids from gonoarthrosis patients contained minimal
amounts of the enzyme. When the amount of ARG activity at
pH 7.5 was measured, the difference between RA and gonoar-
throsis was highly significant (Fig. 5B). Most importantly, in
four RA synovial fluids, the amount of ARG activity detected
was well above the levels detected in the supernatant of PMNs
undergoing frustrated phagocytosis (Fig. 4C).
This finding indicates that in RA, PMN-dependent exocyto-
sis and activation of ARG1 may occur within the articular cav-
ity and that ARG1 activity may reach levels capable of inhibit-
ing T cell proliferation. Other cell types probably contribute
to inhibition of T cell proliferation in RA joints, as the ARG2
isoform is expressed by synoviocytes and macrophages, which
do not exocytose the enzyme [30].
It has been demonstrated that ARG1, freed upon PMN cell
death, also suppresses NK cell proliferation [31]. ARG1-depen-
dent depletion of l-arginine, resulting from release of azuro-
phil and gelatinase granules by activated PMNs, could there-
fore inhibit T and NK cell proliferation. It remains to be eluci-
dated whether in RA patients, this phenomenon might result
in down-regulating inflammatory reaction, deranging immune
response, or inducing both events.
A P = 0,0786
B P = 0,0477
g/ml)
700
ur)
G1 concentration (µg
600
Urea (nmoles/ml/hou
500
400
300
200
100
ARG
0 0
Gonoarthrosis RA
Figure 5. Relevant levels of ARG1 protein and ARG activity can be
detected in synovial fluids from RA but not from gonoarthrosis.
(A) ARG1 concentration was measured by ELISA in six synovial fluids
from RA and in two from gonoarthrosis. (B) ARG activity was mea-
sured at pH 7.5 in the above samples. The horizontal line indicates
the rate of urea production detected in the supernatant of PMNs acti-
vated by frustrated phagocytosis that inhibited T cell proliferation, as
shown in Fig. 3B. Each symbol indicates the same patient in both pan-
els and is the mean of triplicate assay. Statistical analysis was per-
formed by unpaired Student’s t test.
726 Journal of Leukocyte Biology
Besides RA, PMN-associated, ARG1-dependent immune sup-
pression might also play a role in human solid tumors. In this
context, we have previously found reduced intracellular ARG1
content in PMNs infiltrating human lung cancer [10], indicat-
ing a local ARG1 release and immune suppression in the tu-
mor microenvironment, as also demonstrated by others in re-
nal cell carcinoma [32]. If direct evidence of enzyme activity
will be provided, ARG1 might become a potential target for
therapeutic interventions.
AUTHORSHIP
R.R. performed T cell proliferation and subcellular fraction-
ation experiments; M.B. performed frustrated phagocytosis
and MPO and lactoferrin analysis; G.B. performed ELISA and
Western blot experiments; S.A. evaluated ARG activity; S.M.
produced mAb anti-ARG1; and L.O., F.D., V.B., S.F., and O.B.
designed experiments, provided supervision, analyzed data,
and wrote the paper.
ACKNOWLEDGMENTS
This study was supported by research funding from Italian
Ministero della Salute P. Strategico 2006 (V.B. and O.B.) and
P. Ordinario 2006 (S.F.); AICR-UK 2008 grant number 08-0518
(V.B. and O.B.); Fondazione Carige grant number
2008.0812.132 (F.D.); and AIRC (S.F.). We thank Dr. Marina
Fabbi for helpful discussion and expert technical assistance.
REFERENCES
1. Rodriguez, P. C., Zea, A. H., DeSalvo, J., Culotta, K. S., Zabaleta, J., Qui-
ceno, D. G., Ochoa, J. B., Ochoa, A. C. (2003) L-arginine consumption
by macrophages modulates the expression of CD3 ␨ chain in T lympho-
cytes. J. Immunol. 171, 1232–1239.
2. Munder, M., Mollinedo, F., Calafat, J., Canchado, J., Gil-Lamaignere, C.,
Fuentes, J. M., Luckner, C., Doschko, G., Soler, G., Eichmann, K., Muller,
F. M., Ho, A. D., Goerner, M., Modolell, M. (2005) Arginase I is constitu-
tively expressed in human granulocytes and participates in fungicidal ac-
tivity. Blood 105, 2549 –2556.
3. Rotondo, R., Barisione, G., Mastracci, L., Grossi, F., Orengo, A. M., Costa,
R., Truini, M., Fabbi, M., Ferrini, S., Barbieri, O. (2009) IL-8 induces ex-
ocytosis of arginase 1 by neutrophil polymorphonuclears in nonsmall cell
lung cancer. Int. J. Cancer 125, 887– 893.
12500
4. Jacobsen, L. C., Theilgaard-Monch, K., Christensen, E. I., Borregaard, N.
(2007) Arginase 1 is expressed in myelocytes/metamyelocytes and local-
10000
ized in gelatinase granules of human neutrophils. Blood 109, 3084 –3087.
5. Munder, M., Schneider, H., Luckner, C., Giese, T., Langhans, C. D., Fu-
7500
entes, J. M., Kropf, P., Mueller, I., Kolb, A., Modolell, M., Ho, A. D.
(2006) Suppression of T-cell functions by human granulocyte arginase.
5000
Blood 108, 1627–1634.
2500
6. Folley, S. J., Greenbaum, A. L. (1948) Determination of the arginase ac-
tivities of homogenates of liver and mammary gland: effects of pH and
substrate concentration and especially of activation by divalent metal
Gonoarthrosis RA ions. Biochem. J. 43, 537–549.
7. Ikemoto, M., Tabata, M., Miyake, T., Kono, T., Mori, M., Totani, M., Mu-
rachi, T. (1990) Expression of human liver arginase in Escherichia coli. Pu-
rification and properties of the product. Biochem. J. 270, 697–703.
8. Böyum, A. (1968) Separation of leukocytes from blood and bone mar-
row. Introduction. Scand. J. Clin. Lab. Invest. Suppl. 97, 7.
9. Bronte, V., Serafini, P., De Santo, C., Marigo, I., Tosello, V., Mazzoni, A.,
Segal, D. M., Staib, C., Lowel, M., Sutter, G., Colombo, M. P., Zanovello,
P. (2003) IL-4-induced arginase 1 suppresses alloreactive T cells in tu-
mor-bearing mice. J. Immunol. 170, 270 –278.
10. Rotondo, R., Mastracci, L., Piazza, T., Barisione, G., Fabbi, M., Cas-
sanello, M., Costa, R., Morandi, B., Astigiano, S., Cesario, A., Sormani,
M. P., Ferlazzo, G., Grossi, F., Ratto, G. B., Ferrini, S., Frumento, G.
(2008) Arginase 2 is expressed by human lung cancer, but it neither in-
duces immune suppression, nor affects disease progression. Int. J. Cancer
123, 1108 –1116.
Volume 89, May 2011 www.jleukbio.org
 Rotondo et al. PMN exocytosis inhibits T cell proliferation via ARG1
11. Kjeldsen, L., Sengelov, H., Lollike, K., Nielsen, M. H., Borregaard, N.
(1994) Isolation and characterization of gelatinase granules from human
neutrophils. Blood 83, 1640 –1649.
12. Ottonello, L., Tortolina, G., Amelotti, M., Dallegri, F. (1999) Soluble Fas
ligand is chemotactic for human neutrophilic polymorphonuclear leuko-
cytes. J. Immunol. 162, 3601–3606.
13. Chatham, W. W., Turkiewicz, A., Blackburn Jr., W. D. (1994) Determi-
nants of neutrophil HOCl generation: ligand-dependent responses and
the role of surface adhesion. J. Leukoc. Biol. 56, 654 – 660.
14. Faurschou, M., Sorensen, O. E., Johnsen, A. H., Askaa, J., Borregaard, N.
(2002) Defensin-rich granules of human neutrophils: characterization of
secretory properties. Biochim. Biophys. Acta 1591, 29 –35.
15. Hanlon, W. A., Stolk, J., Davies, P., Humes, J. L., Mumford, R., Bonney,
R. J. (1991) rTNF ␣ facilitates human polymorphonuclear leukocyte ad-
herence to fibrinogen matrices with mobilization of specific and tertiary
but not azurophilic granule markers. J. Leukoc. Biol. 50, 43– 48.
16. Gullberg, U., Bengtsson, N., Bulow, E., Garwicz, D., Lindmark, A., Ol-
sson, I. (1999) Processing and targeting of granule proteins in human
neutrophils. J. Immunol. Methods 232, 201–210.
17. Borregaard, N., Cowland, J. B. (1997) Granules of the human neutro-
philic polymorphonuclear leukocyte. Blood 89, 3503–3521.
18. Gardiner, E. E., Mok, S. S., Sriratana, A., Robinson, H. C., Veitch, B. J.,
Lowther, D. A., Handley, C. J. (1994) Polymorphonuclear neutrophils
release 35S-labeled proteoglycans into cartilage during frustrated phago-
cytosis. Eur. J. Biochem. 221, 871– 879.
19. Taichman, N. S., Tsai, C. C., Baehni, P. C., Stoller, N., McArthur, W. P.
(1977) Interaction of inflammatory cells and oral microorganisms. IV. In
vitro release of lysosomal constituents from polymorphonuclear leuko-
cytes exposed to supragingival and subgingival bacterial plaque. Infect.
Immun. 16, 1013–1023.
20. Webster, R. O., Wysolmerski, R. B., Lagunoff, D. (1986) Enhancement of
human polymorphonuclear leukocyte adherence to plastic and endothe-
lium by phorbol myristate acetate. Comparison with human C5a. Am. J.
Pathol. 125, 369 –378.
21. Bronte, V., Kasic, T., Gri, G., Gallana, K., Borsellino, G., Marigo, I., Battis-
tini, L., Iafrate, M., Prayer-Galetti, T., Pagano, F., Viola, A. (2005) Boost-
www.jleukbio.org
ing antitumor responses of T lymphocytes infiltrating human prostate
cancers. J. Exp. Med. 201, 1257–1268.
22. Weissmann, G., Zurier, R. B., Spieler, P. J., Goldstein, I. M. (1971) Mech-
anisms of lysosomal enzyme release from leukocytes exposed to immune
complexes and other particles. J. Exp. Med. 134, 149s–165s.
23. Henson, P. M. (1971) Interaction of cells with immune complexes: ad-
herence, release of constituents, and tissue injury. J. Exp. Med. 134, 114 –
135.
24. Jasin, H. E. (1987) Intra-articular antigen-antibody reactions. Rheum. Dis.
Clin. North Am. 13, 179 –189.
25. Firestein, G. S. (2003) Evolving concepts of rheumatoid arthritis. Nature
423, 356 –361.
26. Hasselbacher, P. (1979) Extracellular aggregates of immunoglobulin in
synovial fluid from rheumatoid arthritis. J. Rheumatol. 6, 374 –380.
27. Zvaifler, N. J. (1971) Immunoreactants in rheumatoid synovial effusions.
J. Exp. Med. 134, 276s–285s.
28. Moore, A. R., Appelboam, A., Kawabata, K., Da Silva, J. A., D'Cruz, D.,
Gowland, G., Willoughby, D. A. (1999) Destruction of articular cartilage
by ␣ 2 macroglobulin elastase complexes: role in rheumatoid arthritis.
Ann. Rheum. Dis. 58, 109 –113.
29. Kitsis, E., Weissmann, G. (1991) The role of the neutrophil in rheuma-
toid arthritis. Clin. Orthop. Relat. Res. 63–72.
30. Corraliza, I., Moncada, S. (2002) Increased expression of arginase II in
patients with different forms of arthritis. Implications of the regulation of
nitric oxide. J. Rheumatol. 29, 2261–2265.
31. Oberlies, J., Watzl, C., Giese, T., Luckner, C., Kropf, P., Muller, I., Ho,
A. D., Munder, M. (2009) Regulation of NK cell function by human
granulocyte arginase. J. Immunol. 182, 5259 –5267.
32. Rodriguez, P. C., Ernstoff, M. S., Hernandez, C., Atkins, M., Zabaleta, J.,
Sierra, R., Ochoa, A. C. (2009) Arginase I-producing myeloid-derived sup-
pressor cells in renal cell carcinoma are a subpopulation of activated
granulocytes. Cancer Res. 69, 1553–1560.
KEY WORDS:
granulocytes 䡠 T cell proliferation 䡠 arginase 䡠 granules
Volume 89, May 2011 Journal of Leukocyte Biology 727