Published online before print October 16, 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Cardiovascular Research Unit, School of Medicine and Biomedical Sciences, University of Sheffield, Sheffield, United Kingdom
2 Correspondence: Cardiovascular Research Unit, School of Medicine and Biomedical Sciences, University of Sheffield, Sheffield, S10 2JF, UK. E-mail: p.g.hellewell{at}sheffield.ac.uk
ABSTRACT
Ly-6G is a member of the Ly-6 family of GPI-linked proteins, which is expressed on murine neutrophils. Antibodies against Ly-6G cause neutropenia, and fatal reactions also develop if mice are primed with TNF-
prior to antibody treatment. We have investigated the mechanisms behind these responses to Ly-6G ligation in the belief that similar mechanisms may be involved in neutropenia and respiratory disorders associated with alloantibody ligation of the related Ly-6 family member, NB1, in humans. Neutrophil adhesion, microvascular obstruction, breathing difficulties, and death initiated by anti-Ly-6G antibodies in TNF--primed mice were shown to be highly complement-dependent, partly mediated by CD11b, CD18, and Fc
R and associated with clustering of Ly-6G. Neutrophil depletion, on the other hand, was only partly complement-dependent and was not altered by blockade of CD11b, CD18, or FcR. Unlike other neutrophil-activating agents, Ly-6G ligation did not induce neutropenia via sequestration in the lungs. Cross-linking Ly-6G mimicked the responses seen with whole antibody in vivo and also activated murine neutrophils in vitro. Although this suggests that the responses are, in part, mediated by nonspecific properties of antibody ligation, neutrophil depletion requires an additional mechanism possibly specific to the natural function of Ly-6G.
Key Words: neutrophils • cell activation • adhesion
INTRODUCTION
Ly-6 proteins are a family of GPI-anchored molecules [1
], many of which have been implicated in the regulation of human and murine leukocyte activation and adhesion [2
3
4
5
6
]. Ly-6G is expressed on murine neutrophils [7
]. Although its physiological ligand and biological function are unknown, antibody ligation of the protein causes significant systemic responses. RB6-8C5, an antibody that recognizes Ly-6G, has been used regularly as an experimental tool to investigate the role of neutrophils in various inflammatory responses as a result of its ability to deplete these cells from the systemic circulation [8
, 9
]. We reported previously that ligation of Ly-6G in animals with an underlying background inflammation (induced by TNF-) also led to profound cardiovascular and respiratory responses, including severe breathing difficulties, leukocyte and platelet aggregation, and coagulation, leading to microvessel obstruction and subsequent death [8
]. RB6-8C5 also cross-reacts with Ly-6C, which is expressed on subpopulations of monocytes and T cells and dendritic cells as well as neutrophils. However, the antibody 1A8 specifically recognizes Ly-6G and causes similar responses to RB6-8C5 [8
].
Investigations into the mechanisms underlying the rapid neutropenia and microvessel complications have been limited as a result of the inability of RB6-8C5 to activate neutrophils in vitro [8
]. This is in spite of the ability of RB6-8C5 to cause rapid activation and arrest of neutrophils in venules of TNF--treated mice. It is likely therefore that a number of factors and processes are involved in the in vivo responses. We were interested to determine how these two distinctly characterizable responses to RB6-8C5 occur and specifically, whether they were mediated by discrete or similar mechanisms. Two possibilities were considered: first that the presence of TNF- simply primes neutrophils for an increased response to RB6-8C5 and second that TNF- initiates a distinct series of events to induce the massive systemic response compared with those mediating neutropenia. In addition, we wished to determine whether these mechanisms are specific to the ligation of Ly-6G and its biological function or rather to nonspecific actions such as complement activation or FcR engagement, which are secondary to antibody binding to its ligand.
Our previous work showed that a combination of antithrombotic (heparin) and anti-inflammatory (selectin blockade) strategies was required to protect against the severe cardiovascular and respiratory reactions [8
], although these treatments did not prevent disappearance of leukocytes from the circulation. These interventions prevented the leukocyte adhesion and microvascular occlusion in TNF--primed mice but not, however, the leukocyte activation that initiated these responses.
The microcirculatory disturbances induced by RB6-8C5 in TNF--treated mice are consistent with activation of the complement system, and mediators released from activated neutrophils are among the many factors that can lead to this activation [10
11
12
]. In addition, although in vivo microcirculatory responses were not dependent on FcRII or FcRIII, the capacity of Fab and F(ab')2 fragments of RB6-8C5 to induce reactions was attenuated compared with whole antibody. These data suggested that although the effects of RB6-8C5 were not FcR-mediated, the possibility of a contribution of cross-linking Ly-6G remained.
Based on these observations, we therefore investigated mechanisms underlying neutropenia and neutrophil activation associated with Ly-6G ligation to gain an insight into its natural function or its role in autoimmunity. We hypothesized that ligation of Ly-6G activates cellular responses associated with its biological function and that the widespread cardiovascular and respiratory disturbances in TNF--treated animals are a result of an exaggerated response by already-primed neutrophils. We found that cross-linking Ly-6G [with F(ab')2 and secondary antibody] in vitro activated murine neutrophils, and cross-linking Ly-6G in vivo essentially mimicked the neutropenia and lethal intravascular reaction seen with the whole antibody. The microvascular response to Ly-6G ligation in TNF--primed mice was strongly complement-dependent with some influence of FcRs, whereas neutrophil depletion was delayed but not prevented in the absence of complement or FcR. This suggests that although there may be common pathways involved in the two responses to Ly-6G ligation, they are regulated independently of each other via as-yet unknown mechanisms.
MATERIALS AND METHODS
Materials
mAb were purchased from PharMingen (Oxford, UK) with the exception of the anti-rat 
light-chain antibody RT-39 (IgG1; Sigma Chemical Co., Poole, UK), anti-mouse C1q antibody (7H8; IgG2a; Cell Sciences Inc., Canton, MA, USA), and the antibody recognizing CD18 (2E6; IgG2b). Hybridoma for the latter was purchased from the American Type Culture Collection (Manassas, VA, USA), and the Antibody Resource Center (Sheffield, UK) produced the antibody. RB6-8C5 (IgG2b,) was obtained from two sources: purchased from PharMingen and produced from hybridoma cell lines by the Antibody Resource Center. Responses to RB6-8C5 produced by the Antibody Resource Center mirrored those using RB6-8C5 from PharMingen (<0.01 ng endotoxin/µg protein). F(ab')2 fragments of RB6-8C5 were generated by Cymbus Biotechnology (Hampshire, UK). Alexa-fluor 555 and FITC-labeling kits were purchased from Molecular Probes (Leiden, The Netherlands) and were used according to the manufacturer’s instructions. Mouse anti-rabbit RBC antiserum was a kind gift from Prof. B. Paul Morgan (Complement Biology Group, Cardiff University, UK). Recombinant murine TNF- (270,000 U/µg) was purchased from R&D Systems (Oxon, UK). Cobra venom factor (CVF) was purchased from Technoclone Ltd. (Dorking, Surrey, UK). BSA (Fraction V), gelatin veronal buffer (GVB), myeloperoxidase (MPO), and FITC-phalloidin were purchased from Sigma Chemical Co. PBS containing 1 mM Ca2+ and 0.5 mM Mg2+ (PBS) was purchased from Gibco (Paisley, UK).
Animals
C57BL/6 mice were purchased from Harlan (Oxon, UK). Dr. Daniel Bullard (University of Alabama at Birmingham, AL, USA) provided CD11b–/– mice. Dr. Neena Kalia (University of Birmingham, UK) provided the FcR knockout mice (lacking the common -chain). For some experiments, mice with circulating cells deficient in FcR but endothelium-expressing FcR were produced by transplantation of FcR knockout bone marrow into wild-type mice as described. Using this method, >95% of circulating cells were of the desired phenotype [8
, 13
]. Male mice weighing between 22 and 30 g were used in these experiments. All procedures were approved by The University of Sheffield Ethics Committee (Sheffield, UK) and performed in accordance with the Home Office Animals (Scientific Procedures) Act 1986 of the UK.
Pretreatments
In some experiments, mice were treated overnight with CVF (4.5 µg in 150 µl saline, i.p.; confirmed phospholipase-free on Certificate of Analysis) to deplete complement (see below). For experiments investigating microvascular arrest and death, inflammation was induced by treatment with TNF- (500 ng, intrascrotal) 4 h before antibody treatment. On some occasions, mice received an i.p. injection of anti-CD11b antibody (M1/70; IgG2b,; 30 µg/mouse) or anti-CD18 antibody (2E6; 50 µg/mouse), in the absence or in combination with heparin (50 U/mouse), 15 min before TNF-.
Measurement of complement activity
Complement activity in CVF-treated and nontreated mice was determined using an assay based on the extent of complement-mediated lysis of antibody-coated rabbit RBC [14
], which were washed, diluted to a 2% suspension in PBS, and sensitized by addition of an equal volume of mouse anti-rabbit RBC antiserum for 30 min at room temperature. Sensitized RBC (100 µl) were added to each of a 96-well plate and washed in gelatin veronal buffer. Serum from CVF-treated and nontreated mice was subjected to a series of doubling dilutions and incubated with the sensitized rabbit RBC at 37°C for 30 min. Positive (distilled water; 100% lysis) and negative (GVB; 0% lysis) controls were also included. Remaining RBC were removed by centrifugation, the absorbance of the resulting serum read at 414 nm, and the degree of lysis calculated as a percentage of the maximum achieved.
Intravital microscopy and imaging
Approximately 3.5 h after TNF- injection, mice were anesthetized with an i.p. injection of a mixture of ketamine hydrochloride (125 mg/kg), xylazine hydrochloride (12.5 mg/kg), and atropine sulfate (0.025 mg/kg). Cannulation of the trachea and carotid artery was performed to facilitate breathing and to permit blood sampling and injection of antibodies. The cremaster muscle was then exposed and spread over a specialized viewing platform for intravital microscopy. Temperature was controlled by placing the mouse on a thermistor-regulated heating pad (PDTronics, Sheffield, UK), and the cremaster was superfused continually with thermo-controlled (36°C) bicarbonate-buffered saline.
Microscopic observations were made using an upright microscope (Nikon Eclipse E600-FN, Nikon UK, Kingston upon Thames, UK) with a water-immersion objective (40x/0.80 W) and equipped for brightfield, widefield fluorescence, and confocal microscopy. Brightfield images were viewed on a video monitor via a charge-coupled device camera (DC-330, Dage MTI, Michigan City, IN, USA) and recorded onto miniDV videocassettes. Single, unbranched venules (20–50 µm diameter) were typically selected and observed for the entire experimental period. Center-line blood-flow velocity was measured in vessels of interest using a commercially available velocimeter (Circusoft Instrumentation, Hockessin, DE, USA).
RB6-8C5 (10 µg in 200 µl saline) was injected via the carotid artery. For cross-linking experiments, F(ab')2 of RB6-8C5 (10 µg in 100 µl saline) was injected, followed 5 min later by RT-39 (50 µg in 100 µl saline) to cross-link Ly-6G. Blood-flow velocity was monitored before and for up to 1 h after treatment. In some experiments, animals were treated with fluorescent antibodies (Alexa-555-RB6-8C5, FITC-CD41, FITC-anti-C1q) to selectively label neutrophils and platelets and to investigate the distribution of Ly-6G and C1q on activated neutrophils.
Widefield fluorescent images were acquired using hardware controlled by Slidebook software (Intelligent Imaging Innovations, Denver, CO, USA). Excitation was provided by a Sutter 
DG4 Xenon source equipped with 480 nm and 560 nm excitation filters. Fluorescence was collected through a triple-band pass filter for 4',6-diamidino-2-phenylindole, FITC, and tetramethyl rhodamine isothiocyanate and signals amplified using a Videoscope (Sterling, VA, USA) intensifier prior to capture by a Sensicam (Cooke, Auburn Hills, MI, USA). Confocal microscopy of venules was performed using a Yokogawa scanning head coupled to a three-line (488/568/647) Argon/Krypton laser. Confocal images were captured using a Roper Scientific Coolsnap HQ camera under the control of Slidebook software. Intelligent Imaging Innovations provided equipment for widefield fluorescence and confocal microscopy.
In other experiments, intravital microscopy was not performed, and mice were simply observed for 1 h after RB6-8C5 injection or Ly-6G cross-linking. Time to death or survival at 1 h was recorded.
Blood counts
Mice were anesthetized, and RB6-8C5, 1A8 (IgG2a,), or A95-1 (IgG2b,) was injected (10 µg in 200 µl saline), or Ly-6G cross-linking was performed as above. Blood samples were taken via the carotid cannula into 3% acetic acid at various time-points before and after treatment (up to 1 h) to determine the effects of antibody treatment on circulating leukocyte number. Samples were analyzed for total and differential leukocyte concentrations.
MPO assay
Mice treated with A95-1, RB6-8C5, or zymosan (200 µg) for up to 30 min were killed, and lungs and spleens were collected. Tissues were washed clean of surface blood in ice-cold saline and kept on ice or snap-frozen in liquid nitrogen. Samples were weighed and homogenized in 0.5% hexa-decyl-trimethyl-ammonium bromide (1 ml/50 mg tissue) in 10 mM MOPS. Samples were subjected to sonication, 4x freeze/thaw cycles, and further sonication, before tissue homogenates were centrifuged at 14,000 g for 30 min and supernatant collected. Samples and supernatants were kept on ice at all times.
Tissue supernatants were diluted 1:10 in sodium phosphate buffer (NaH2PO4.H2O; pH 5.5), and 20 µl was added to a 96-well plate. Known concentrations of MPO were also added to separate wells to provide a standard curve. H2O2 (20 µl 1 mM) was added to each well, followed by 2 mM tetramethyl-benzidine in sodium phosphate buffer, and incubated in the dark for 20 min. The OD of each standard or sample was measured at 650 nm, and the concentration of MPO in the samples was calculated by cross-referencing to the standard curve. The concentration of MPO was proportional to the number of neutrophils present in each sample.
Flow cytometry
Blood was drawn from anesthetized mice and anticoagulated with EDTA. RBC were removed by hypotonic lysis and leukocytes suspended in PBS containing 0.15% BSA. Leukocytes were incubated with F(ab')2 fragments of RB6-8C5 (4 µg/ml, 30 min, 4ºC) followed by RT-39 (20 µg/ml, 30 min, 4ºC) to cross-link Ly-6G. Negative control samples were left untreated or treated with F(ab')2 or RT-39 alone, and positive control samples were incubated with 10–7 M PMA for 15 min at 37ºC. For experiments determining F-actin content of neutrophils, cells were subsequently fixed with 4% methanol-free formaldehyde (20 min, room temperature) and permeabilized with 0.1% Triton X-100 (10 min, room temperature). FITC-conjugated antibodies against CD11b (M1/70) or isotype control (A95-1; both 30 min, 4ºC) or FITC-phalloidin (0.25 µg/ml; 20 min, room temperature) were added before CD11b expression, or F-actin content of neutrophils was analyzed by flow cytometry of appropriately gated cell populations.
To investigate neutrophil release from bone marrow, mice treated with A95-1, RB6-8C5, or zymosan for up to 30 min were killed and bone marrow collected. Cells were dispersed using a 25G needle and RBC removed by hypotonic lysis. Cell samples were incubated with PE-1A8 (30 min, 4ºC) before the proportion of bone marrow cells that were neutrophils (1A8-positive) was determined using flow cytometry.
Statistics
All data shown represent mean ± SEM. Statistical analyses were performed using GraphPad Prism software (San Diego, CA, USA) according to the manufacturer’s instructions. Analyses used were Student’s t-test, two-way ANOVA, or one-way ANOVA followed by Bonferroni’s multiple comparison test where appropriate.
RESULTS
Neutrophil depletion in response to Ly-6G ligation consists of complement-dependent and independent responses
Anti-Ly-6G and Ly-6C antibody, RB6-8C5, causes depletion of neutrophils from the peripheral circulation [9
, 15
16
17
18
]. This property of RB6-8C5 has been widely exploited to investigate the contribution of neutrophils to inflammatory disease models. Data presented in Figure 1A
demonstrate rapid removal of polymorphonuclear neutrophils (PMN) from the peripheral circulation of mice, and the majority of neutrophils was removed within 1 min. This effect was mirrored by antibody 1A8, which also recognizes Ly-6G (data not shown), but not by isotype control antibody (A95-1; see Fig. 1D
).
 View larger version (30K): [in a new window] |
Figure 1. Mechanisms of rapid neutrophil depletion induced by anti-Ly-6G antibodies. (A) Time course of neutrophil depletion induced by RB6-8C5. Neutrophils were depleted from the circulation within 1 min of injection. (B) Release of neutrophils from bone marrow (BM) in response to RB6-8C5 and zymosan. (C) MPO content, indicative of neutrophil content, in the lungs in RB6-8C5 and zymosan-treated mice. (D) Neutrophil depletion in response to anti-Ly-6G antibody RB6-8C5 (compared with isotype control, A95-1) was delayed but not prevented by CVF pretreatment. (E) Antiadhesion regimes had no significant effect on RB6-8C5-induced neutrophil depletion (data shown 3 min after RB6-8C5 administration). All experiments were repeated in three to five mice. *, P < 0.05; ***, P < 0.001, compared with baseline/isotype control;  , P < 0.01
|
A well-established cause of rapid neutropenia is via complement factors C3a and C5a [22
23
24
]. Therefore, we decided to investigate the role of complement in PMN depletion induced by RB6-8C5 antibody. Mice were pretreated with CVF (overnight treatment), which reduced the ability of serum from treated mice to lyse antibody-coated rabbit RBC to less than 2.5% of that in untreated mice, indicating near total exhaustion of complement. CVF depletes complement through activation of the complement system, which therefore induces a rapid neutropenia followed by a later neutrophilia [25
]. However, baseline neutrophil counts, immediately prior to beginning the experiment, demonstrated no significant difference between nontreated and CVF-treated mice, indicating recovery had occurred (data not shown). Following overnight CVF treatment, mice received systemic injections of RB6-8C5. Blood samples were drawn from the carotid artery at serial time-points and analyzed for PMN concentration. RB6-8C5-induced PMN depletion was delayed but not prevented by CVF pretreatment and was unaffected by the additional administration of Fc-block (anti-CD16 and anti-CD32 antibodies; Fig. 1D
). RB6-8C5-induced PMN depletion was also unaffected in FcR-knockout mice, compared with wild-type controls (e.g., neutrophil counts were 8.4±6.2% and 0±0% of baseline, respectively, 10 min after RB6-8C5 injection).
Antiadhesion regimes delivered no protection from the neutropenic effects of RB6-8C5: Antibodies against CD11b and CD18 failed to prevent RB6-8C5-induced PMN depletion, and CD11b knockout mice were also unprotected against this effect (Fig. 1E) .
Microvascular disturbance and lethality induced by TNF- and Ly-6G ligation are complement- and CD11b/CD18-dependent
In addition to rapid neutropenia, systemic administration of RB6-8C5 causes adhesion of large leukocyte and platelet aggregates to vessel walls and a precipitous drop in blood flow through cremaster venules if mice are primed 4 h prior to antibody with a local injection (500 ng, intrascrotal) of TNF- [8
]. Evidence of microvascular injury and a severe systemic reaction is also evident, and mice die rapidly [8
]. In the present study, six of six TNF--primed mice lost all flow-through, previously perfused venules within 10 min of RB6-8C5 application (Fig. 2A
), and 14/14 mice died within 25 min (Fig. 2B)
. Antibody 1A8 also induced a similar responses (ref. [8
], and data not shown). These reactions are specific to Ly-6G ligation, as they are not initiated in TNF--primed animals subsequently treated with Fc-block (anti-CD16 and anti-CD32 antibodies, PharMingen) or Mel-14 [8
], which both bind to neutrophil surface receptors and are of the same isotype as RB6-8C5 and 1A8, respectively. CVF-treated mice were fully protected from loss of blood flow and death induced by TNF- and RB6-8C5 (Fig. 2 A and B)
, suggesting that these reactions are strongly complement-dependent.
 View larger version (23K): [in a new window] |
Figure 2. Complement depletion by CVF protects against microvascular arrest and mortality induced by TNF- and RB6-8C5. Blood-flow velocities (A; n=3–6) and mortality (B; n=7–14) were measured for 1 h after application of RB6-8C5 to TNF--primed mice. Some mice were pretreated overnight with CVF. Two-way ANOVA showed a significant effect (P<0.0001) of CVF compared with untreated control on blood-flow velocity and mortality.
|
-primed mice with Alexa 555-labeled RB6-8C5. Venules were observed by confocal microscopy, and nonuniform distribution of fluorescently labeled RB6-8C5 on the cell surface indeed suggested that clustering of Ly-6G had occurred (Fig. 3A
). Observation of neutrophils in the vasculature of TNF--primed animals was possibly a result of rapid formation of stable leukocyte platelet aggregates adherent to the inflamed vessel wall. In contrast, we were unable to observe Ly-6G distribution on non-TNF--primed neutrophils in vivo because of the rapid depletion of the majority of the neutrophils from the peripheral circulation upon injection of RB6-8C5 (see above) and the remaining cells flowing in the bloodstream at a velocity too great to capture confocal images. In separate experiments, we treated TNF--primed mice with FITC-labeled anti-C1q antibody and Alexa 555-labeled RB6-8C5. The antibodies colocalized in postcapillary venules (Fig. 3B)
, suggesting that RB6-8C5 clustered on the surface of neutrophils activates the complement cascade.
 View larger version (59K): [in a new window] |
Figure 3. Clustering of Ly-6G on activated neutrophils and colocalization with C1q. TNF--primed mice were treated with Alexa-555-labeled RB6-8C5, which both labeled neutrophils and stimulated microvascular arrest and mortality. Some mice also received FITC-C1q antibody. (A) Confocal image of RB6-8C5 clustered on the surface of adherent neutrophils in a venule (x60 objective). Arrows indicate areas of more intense antibody staining. (B) Complement component C1q (green) colocalized with leukocytes (red; x40 objective). Main image shows an overlay of the two inset images.
|
plus RB6-8C5. We discovered previously that antiselectin molecule antibodies only protect against these responses when coadministered with heparin [8
]. Results shown in Figure 4 A and B
, therefore indicate the effects of anti-CD11b and anti-CD18 plus heparin administered 15 min before TNF-. Interestingly, anti-CD11b and anti-CD18 provided partial protection from loss of blood flow induced by RB6-8C5 treatment in TNF--primed mice (Fig. 4A)
. Blood-flow velocity in these mice fell and remained below baseline for the duration of experiments, but most animals maintained some flow. In addition, anti-CD11b or anti-CD18 treatment protected most mice from the lethal effects of TNF- plus RB6-8C5 (Fig. 4B)
. This was also true in a small cohort of mice that was administered anti-CD18 antibody in the absence of heparin; three of four mice survived 60 min following TNF- plus RB6-8C5. The protection given by these integrin-blocking antibodies was less than seen previously in E- or P-selectin knockout mice or mice treated with a combination of E- and P-selectin antibodies plus heparin [8
]. A role for CD11b in the lethal response to RB6-8C5 in combination with TNF-, in the absence of heparin, was confirmed by the survival of 100% of CD11b-knockout mice treated in this manner (Fig. 4B)
.
 View larger version (17K): [in a new window] |
Figure 4. Mechanisms of microvascular arrest and mortality induced by TNF- and RB6-8C5. Blood-flow velocities (A and C, n=3–6) and mortality (B; n=5–18) were measured for 1 h after application of RB6-8C5 to TNF--primed mice. Interventions investigated were anti-CD11b mAb, anti-CD18 mAb, or the use of CD11b knockout mice (CD11b–/–). Blood-flow velocities were also investigated in wild-type mice transplanted with bone marrow from wild-type (Control) or FcR knockout (FcR–/–) mice. Two-way ANOVA showed a significant effect (P<0.0001) of anti-CD11b, anti-CD18, and CD11b, and FcR knockout mice compared with control on blood-flow velocity and of all antiadhesion interventions on mortality.
|
Rs is involved in the inflammatory response [26
]. Our earlier study [8
] showed that antibody block of FcRII and FcRIII failed to protect mice from RB6-8C5-induced microvascular arrest and death. FcRs are present on neutrophils and endothelial cells. In the present study, we sought to further investigate the role of neutrophil FcR by using wild-type mice transplanted with bone marrow from FcR knockout mice such that circulating neutrophils were deficient in all forms of FcR, and endothelial cells still expressed their usual levels of receptors. In contrast to our previous study, deficiency in FcR delayed but did not prevent the reduction in blood flow (Fig. 4C)
. This indicates some interaction between complement and FcR ligation, although the majority of the response was FcR-independent.
Cross-linking Ly-6G activates neutrophils in vitro and mimics ligation by whole RB6-8C5 antibody in vivo
Our results so far highlight attributes shared by RB6-8C5 and 1A8 that may relate to the predominance of functions of Ly-6G rather than nonspecific effects of antibodies against it. Antibodies against Ly-6G cause PMN depletion that is accelerated by, but not dependent on, complement. Responses to such antibodies are also worsened in the presence of existing inflammation. Closer investigation of the functions of Ly-6G by systematic in vitro experimentation is precluded by the observations that anti-Ly-6G antibodies do not appear to activate neutrophils in vitro [8
], and no natural ligands for Ly-6G are known. To circumvent these limitations, we investigated the effects of Ly-6G cross-linking on mouse neutrophil activation in vitro.
Mouse blood was collected and erythrocytes removed by hypotonic lysis. Samples were then incubated with different combinations of F(ab')2 fragments of RB6-8C5 and secondary antibody (RT-39) against these F(ab')2 as controls or to cross-linking Ly-6G, which caused a significant up-regulation of CD11b on the surface of neutrophils, whereas F(ab')2 or anti-F(ab')2 alone did not (Fig. 5 A and B ). Further evidence for neutrophil activation was demonstrated by a 22% increase in F-actin content and therefore, increased cell rigidity upon Ly-6G cross-linking. However, this response was small compared with a 94% increase in F-actin content upon PMA stimulation.
 View larger version (29K): [in a new window] |
Figure 5. Cross-linking (X-link) Ly-6G induces CD11b up-regulation in vitro and mimics neutrophil responses to whole RB6-8C5 antibodies in vivo. (A) CD11b expression on control and Ly-6G-cross-linked neutrophils in vitro. **, P < 0.01. (B) Representative histogram plot showing fluorescence of FITC-CD11b mAb bound to resting and Ly-6G-cross-linked neutrophils. FL1-H, Fluorescence 1-height. (C) Neutrophil depletion in response to Ly-6G cross-linking in vivo. Data are shown 5 min after antibody treatment or cross-linking. (D) Time course of neutrophil depletion induced by Ly-6G cross-linking in comparison with RB6-8C5. (E) Blood-flow velocities of mice subjected to TNF- priming followed by Ly-6G cross-linking in the presence or absence of overnight CVF pretreatment; P < 0.001. Baseline readings were taken at –5 min. Ly-6G was cross-linked at 0 min. (F) Mortality induced by RB6-8C5 or Ly-6G cross-linking in the presence or absence of overnight CVF pretreatment. Two-way ANOVA showed a significant difference (P<0.001) among the three treatments. All experiments were repeated in three to 16 mice. ***, P < 0.001, compared with control.
|
-primed mice, although there were significantly fewer deaths compared with whole RB6-8C5. Loss of microvascular perfusion and death induced by Ly-6G cross-linking in TNF--primed mice had a strong requirement for complement (flow was maintained above 90%, and all CVF-treated animals survived, Fig. 5 E and F
). On the other hand, PMN depletion induced by cross-linking was complement-independent in TNF--primed or nonprimed mice (Table 1
). |
View this table: [in a new window] |
Table 1. PMN Depletion Induced by Cross-Linking Ly-6G
|
Ly-6G, a GPI-linked protein expressed on the surface of murine neutrophils, is a member of the Ly-6 superfamily of proteins. Although some members of this family {e.g., urokinase-type plasminogen activator receptor (uPAR) [2 , 27 , 28 ] and CD59 [29 30 31 32 33 ]} have been studied extensively, functions of the majority, including Ly-6G, are unknown. Several members of the Ly-6 superfamily are expressed on the surface of hematopoietic stem cells, where their expression levels can reflect different levels of differentiation [7 , 34 35 36 ], suggesting that Ly-6 family members might be involved in regulating interaction of these cells with their environment. Support for such a possibility is gained from observations that cross-linking various Ly-6 family members induces activation of the cells upon which they are expressed [3 , 4 , 6 , 37 ]. Our investigations suggest a similar function for Ly-6G.
Ligation of Ly-6G causes neutrophil depletion in healthy mice. In addition, a range of more severe cardiovascular and respiratory reactions, leading to death, also occurs in mice with an underlying background of inflammation. Antibodies against NB1, a human neutrophil Ly-6 family member, also cause strikingly similar responses in man. In the absence of inflammation, alloantibodies against NB1 cause a relatively mild reaction, e.g., neonatal neutropenia in NB1-positive children born to NB1-negative mothers [38 ]. In contrast, in the case of pre-existing inflammation (e.g., in patients undergoing blood transfusion), the presence of antibodies against NB1 in donor serum may cause transfusion-related acute lung injury. Given these comparable physical manifestations of Ly-6 ligation in mice and humans, it is possible that equivalent cellular and chemical processes could mediate these responses in both species.
We predicted that similar mechanisms mediate the mild and severe reactions to anti-Ly-6G antibodies, and responses are simply amplified by priming of the vasculature or leukocytes, where inflammation is already present. As far as complement is concerned, however, this turns out not to be the case. Depleting complement prevented the severe effects of Ly-6G ligation in TNF--primed mice but only delayed anti-Ly-6G-induced neutropenia. Activation of the complement cascade provides an explanation for many of the features of the reaction to TNF-/Ly-6G ligation, which we described previously [8
]. These include leukocyte arrest mediated by CD11b, aggregation, vascular leakage, vasculitis, and arterial constriction [39
40
41
]. Exactly how complement activation is initiated in the reaction to TNF- and Ly-6G ligation is not clear. As direct activation by TNF- is unlikely, injection of RB6-8C5 antibody is the obvious initiating event. Colocalization of C1q with neutrophils implies activation of the classical complement pathway, and this could be initiated by binding of C1q to RB6-8C5 ligated to Ly-6G. C1q also binds exposed phospholipids on activated cells [42
43
44
], and this remains an alternative explanation for its association with neutrophils. Activated neutrophils rapidly release factors (e.g., properdin and proteases) that activate complement [10
11
12
], and this could represent an additional route by which complement activation occurs via the alternative pathway. Further studies are required to investigate these possibilities.
The delay in anti-Ly-6G-induced neutropenia (in the absence of TNF- pretreatment) by complement depletion suggested that ligation of Ly-6G in vivo rapidly activates the complement cascade, promoting removal of neutrophils from the circulation. This may involve the anaphylatoxins C3a and C5a, opsonization of neutrophils by C3b, or the terminal complement product membrane attack complex (MAC). Although MAC is known to cause cell lysis, we do not consider this a likely mechanism for the initial rapid neutrophil depletion, as phagocytic cells are resistant to MAC lysis [45
], and we did not observe lysis of adherent leukocytes in venules. C3b opsonization of neutrophils has been implicated recently in removal of Ly-6G-ligated cells from the systemic circulation [46
] over a time period of 3 days. This is a possibility in our studies, although it is unclear whether this process can occur so rapidly as to remove approximately 90% of the circulating neutrophils within 1 min. In contrast, C5a and C3a are well known to induce a rapid but transient neutropenia in rodents, with maximal depletion occurring within 1–5 min, returning to baseline 5–10 min later [22
23
24
]. Although neutrophil depletion seen in our model is induced similarly and rapidly in the presence of complement, depletion is maintained for up to 60 min, suggesting an additional complement-independent, slower mechanism, keeping neutrophils out of the circulation.
The role of FcR ligation in the initiation of the responses to RB6-8C5 appears, at first, to be conflicting. Mice deficient in the -chain common to all FcR are partially protected from the rapid reduced blood flow initiated by RB6-8C5, and this response in wild-type mice treated with Fc-block (anti-FcRII and -FcRIII antibodies) was unaffected [8
]. This discrepancy can most likely be explained by ligation of the activating receptor FcRIV by the Fc portion of RB6-8C5. FcRIV has been described to preferentially bind IgG2a and IgG2b subclasses [47
], the latter category being that into which antibody RB6-8C5 falls. As knockout of leukocyte FcR provides some protection against the severe actions of Ly-6G antibodies but has no effect on RB6-8C5-induced neutropenia, it is likely that a combination of mechanisms contributes to these responses.
GPI-linked proteins such as Ly-6G are relatively mobile in the plasma membrane and can accumulate preferentially in cholesterol-rich microdomains known as lipid rafts or glycosphingolipid-enriched microdomains. Although they lack a cytoplasmic domain, clustering of GPI-linked proteins and lateral interactions with other cell-surface proteins enable them to recruit other membrane molecules to lipid rafts and have a major influence on cellular responses. Antibody cross-linking of GPI-linked proteins induces similar cellular responses as natural ligand binding [33
]. We find that cross-linking Ly-6G in vitro causes up-regulation of a marker of neutrophil activation, CD11b, presumably via release from intracellular stores. This may be suggestive of a physical association between Ly-6G and integrins on the cell surface, as has been demonstrated for three other GPI-linked proteins: uPAR, CD14, and FcRIIIB [48
]. Although this up-regulation could plausibly explain the neutrophil depletion, microvascular disturbances, and death detected in mice subjected to Ly-6G cross-linking, experiments, where the effect of CD11b is removed by antibodies or the use of CD11b knockout mice, suggest that other mechanisms prevail. At first glance, it seems counterintuitive that antiadhesion regimes protect against the lethal effects of RB6-8C5 as a result of rapid removal of neutrophils from the circulation. However, it should be remembered that neutrophil depletion is not complete; 5–10% of neutrophils remain in the peripheral circulation. This appears to be a sufficient proportion to induce microvascular disturbances via activation and adhesion to the vessel wall.
Although we have clearly demonstrated rapid neutrophil depletion in response to Ly-6G ligation, the fate of these neutrophils is unclear. We observed a slight increase in MPO content in the lungs upon Ly-6G ligation and a corresponding increase in F-actin content of neutrophils of similar magnitude upon Ly-6G cross-linking. This would support the possibility of a small proportion of neutrophils being mechanically trapped in smaller pulmonary vessels as a result of their increased rigidity. In comparison, zymosan and PMA induced significantly greater neutrophil sequestration in the lungs and neutrophil rigidity, respectively. Taken together, these findings suggest that pulmonary sequestration is unlikely to be the major site of neutrophil depletion in response to Ly-6G ligation, especially when removal of additional neutrophils subsequently released from bone marrow is also considered and that the observed depletion occurs via a mechanism different from that induced by more traditional activating agents. Uptake of neutrophils into the bone marrow and subsequent phagocytosis by resident macrophages have been described recently as a mechanism of neutrophil clearance [49 ]. However, as this process is reported to occur over a time period of 24 h, compared with our observation of a reduction of neutrophils in the bone marrow 30 min after RB6-8C5 injection, it is unlikely that this mechanism is a major contributor to neutrophil removal in this study. Given that complement appears to play a role in the early stages of neutrophil depletion induced by Ly-6G, it is possible that opsonization by C3b initiates neutrophil clearance via the reticuloendothelial system [46 ]. Rapid removal of neutrophils implies a role for tissues of this system, which are in direct contact with blood, such as the spleen (which we have eliminated as a site of sequestration) or the liver. As Kupffer cells in the liver are also a site of phagocytosis of apoptotic neutrophils [50 ], this organ should be considered a contender for removal of Ly-6G-ligated neutrophils, although further studies would be required to definitively elucidate the site of neutrophil sequestration or destruction.
We have provided evidence that antibodies against Ly-6G elicit neutrophil depletion and cause a lethal, intravascular response via a number of mechanisms. Treatments that prevent or reduce the lethal response to anti-Ly-6G antibodies include complement depletion, genetic deficiency of FcR, E- or P-selectin [8
] or CD11b, and antiadhesion molecule (E- and/or P-selectin [8
], CD11b, or CD18) antibodies given in combination with heparin. None of these treatments prevents anti-Ly-6G-induced neutrophil depletion. It appears therefore that although the severe cardiovascular and respiratory responses to RB6-8C5 may be mediated by nonspecific actions of ligand ligation by antibody, neutrophil depletion is mediated by other responses relevant to the functions of Ly-6G. Further investigations into the mechanisms of neutrophil depletion by ligation of Ly-6G may provide clues to its natural function in vivo.
ACKNOWLEDGEMENTS
This work was supported by a Career Establishment Grant from the Medical Research Council and a studentship (FS99040) from the British Heart Foundation (M. J. C.). We thank Jayne Chamberlain and Professor Paul Morgan (Complement Biology Group, Cardiff University, UK) for the kind gift of mouse anti-rabbit RBC antiserum and for their guidance with the complement-mediated hemolysis assay [14 ]. Thanks also go to Majid Ali and Paul Dodd for their assistance with mouse surgery and Mark Ariaans for his experimental support.
FOOTNOTES
Received May 17, 2007; revised August 7, 2008; accepted August 17, 2008.
REFERENCES
-bungarotoxin and other neurotoxins J. Immunol. 150,5379-5390[Abstract] RI/III and C5aR in the reverse Arthus reaction J. Immunol. 164,1065-1070}-dependent iNOS expression in the vasculature of healthy mice Blood 111,5187-5194RIV: a novel FcR with distinct IgG subclass specificity Immunity 23,41-51[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  M. P. Kuligowski, R. Y. Q. Kwan, C. Lo, C. Wong, W. G. James, D. Bourges, J. D. Ooi, L. D. Abeynaike, P. Hall, A. R. Kitching, et al. Antimyeloperoxidase antibodies rapidly induce {alpha}4-integrin-dependent glomerular neutrophil adhesion Blood, June 18, 2009; 113(25): 6485 - 6494. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
