|
|||||||||
|
|
|
|
|
|
|
|
|
|
|
PHAGOCYTES
From the Departments of Pediatrics, Internal Medicine,
and Experimental Pathology, the Huntsman Cancer Institute, and the
Program in Human Molecular Biology and Genetics, the University of Utah
School of Medicine, Salt Lake City, UT.
Leukocyte adhesion deficiency type I (LAD-1) is a disorder
associated with severe and recurrent bacterial infections, impaired extravascular targeting and accumulation of myeloid leukocytes, altered
wound healing, and significant morbidity that is caused by absent or
greatly diminished surface expression of integrins of the
To eliminate invasive microrganisms, leukocytes
must adhere to endothelial cells and migrate from the blood through the
endothelial barrier and the extracellular matrix to sites of infection
and tissue injury. The ability of neutrophils (polymorphonuclear
leukocytes [PMNs]) and other classes of leukocytes to reach the
extravascular milieu is dependent on the coordinate actions of adhesion
molecules on their surfaces and on endothelial cells.1-4
These critical tethering molecules include integrins.5 The
principal class of integrins on neutrophils and other leukocytes is
composed of a common Neutrophils and other leukocytes circulate in a quiescent nonadhesive
state but can rapidly become adhesive when activated by signaling
molecules that include bacterial and complement peptides, chemokines,
lipid factors, and other inflammatory mediators.7,10-12 Cellular activation causes heterodimers to become competent to recognize their ligands on other cells and in matrix structures, a
process that is termed "inside-out signaling" or "integrin
activation"1,13,14 and that involves altered conformation
and/or clustering of the integrins.6,7,15,16 Cytoskeletal
proteins and intracellular transduction molecules interact directly or
indirectly with the cytoplasmic domains of integrins of different
classes and are involved in their conversion from the quiescent to the
adhesive state.13,14,17 The specific biochemical
mechanisms that mediate inside-out signaling of integrins are, however,
not completely defined.14,17 In addition to their adhesive
functions, integrins can modulate cellular behavior by transmitting
signals from the exterior to the inside of the cell, a process referred
to as "outside-in signaling."13,18,19
Human syndromes resulting from deficiency or impaired function of
integrins and other adhesion molecules have contributed critical
insights into their biologic roles.7 They are also important, albeit rare, clinical problems. Leukocyte adhesion deficiency type I (LAD-1) is an autosomal recessive disorder
characterized by recurrent infections that are frequently
life-threatening, impaired extravascular accumulation of PMNs and
monocytes that is manifested clinically as lack of pus formation, and
dysregulated wound healing that causes late separation of the umbilical
cord and other abnormalities of tissue repair in many
subjects.6,20,21 LAD-1 is distinguished from LAD-2, which
is caused by a defect in synthesis of fucosylated glycoconjugates that
are required for the functions of selectin ligands,22-24
by impaired Recently 2 subjects with "variant" forms of LAD-1 have been
reported.29,30 The surface levels of Antibodies
Assay of surface expression of cell adhesion molecules
Assays of neutrophil  2 function on stimulated neutrophils with assays of adhesion, aggregation, and chemotaxis. The 2
integrin-dependent adhesion to human umbilical vein endothelial cells
(HUVECs) and to immobilized matrix proteins was assayed as previously
described.33-35 Matrix proteins were added to wells at
concentrations from 30 to 60 µg/mL, allowed to adhere at 4°C
overnight, blocked with 1% human serum albumin (HSA) for 2 hours at
37°C, and washed with Hanks Balanced Salt Solution (HBSS;
Biowhittaker; Walkersville, MD) before use. HUVECs were isolated with
collagenase and grown in a primary culture on gelatin-treated
wells.36 Neutrophils were isolated and labeled with 0.0185 MBq (0.5 µCi) 111Indium per 106
cells.36 Adhesion was induced with activating agonists
that trigger inside-out signaling of 2
integrins34 or with divalent cation
substitution.37-39 We have previously shown that adhesion under these conditions is specifically mediated by 2
integrins.34,35 For divalent cation substitution,
neutrophils were suspended in HBSS, free of Ca++ and
Mg2+, containing 1 mM EDTA, washed one time, and
then resuspended in 1 mM Mn2+.38 The mAb KIM
185 (10 µg/mL) was also used to stimulate 2-dependent adhesion of PMNs.40,41 Aggregation assays were performed
with 5.5 × 106 neutrophils per milliliter and were
measured by the increase in light transmission in an aggregometer
(Payton Scientific, Buffalo, NY).36 Chemotaxis was
measured as previously described.42
Analysis of other leukocyte functions Oxygen radical generation.
Superoxide anion generation was measured as the fraction of
cytochrome c reduction inhibited by superoxide
dismutase.33 Briefly, 2.2 × 106 neutrophils
were added to the well of a 96-well plate in a final volume of 20 mL
HBSS containing 1% HSA and cytochrome c (0.5 mg/mL) with or
without superoxide dismutase (0.1 mg/mL). Neutrophils were stimulated
with phorbol myristate acetate (PMA, 10 Intracellular calcium.
Cells were loaded with the calcium-sensitive fluorescent dyes Indo-1
(neutrophils) or Fura-2 AM (EBV-transformed lymphoblasts) by incubation with 10 Actin polymerization. Neutrophils at 5 × 106 cells/mL were combined with 8% formaldehyde, 150 µg/mL lysophosphatidylcholine, and 50 µg/mL FITC-conjugated phalloidin (2.5 µg/mL, Sigma Chemical) for 4 hours on ice and then washed with phosphate-buffered saline (PBS). Filamentous actin was detected by flow cytometry as previously described.45 Preparation of transformed lymphoblasts Lymphocytes from the patient or from control subjects were suspended at 2 × 106 cells/m in RPMI 1640 (Biowhittaker, Walkersville, MD) with 2 mM L-glutamine, 20% fetal calf serum (Hyclone, Logan, UT), and 0.4 µg/mL cyclosporin A,46 and plated at 2 × 105 cells per well into a 96-well flat-bottomed plate. The 100-µL filtered (0.45 µm) supernatant from a culture of the EBV-shedding line B958 (American Tissue Culture Collection, Rockville, MD) was added to each well.47 The cells were fed as needed with media containing 10% fetal calf serum and 0.2 µg/mL cyclosporin A. Wells were examined for growth, and cells were transferred first to 24-well plates and later to flasks. Cyclosporin A was discontinued after 3 weeks. The transformed lymphoblasts were then maintained in RPMI-1640 with 5% fetal calf serum, 5% fetal clone 1 (Hyclone), and 2 mM glutamine with antibiotic supplementation (penicillin-streptomycin and amphotericin B). The patient-derived EBV-transformed lymphoblasts (LADV cells) and age-matched control EBV-transformed lymphoblasts (control lymphoblasts) were used in subsequent experiments.Transformed lymphoblast adhesion assay EBV-transformed lymphoblasts were labeled with 0.185 MBq (5 µCi) 111Indium per 106 cells using the same protocol as previously described for neutrophils.36 The labeled lymphoblasts were then added to wells coated with immobilized ligands (details in figure legends) and allowed to settle for 60 minutes at 4°C, and the adhesion assay was performed as described for neutrophils (above). In other experiments, lymphoblast adhesion was measured by using fluorescent labeling as previously described.48,49 Briefly, 96-well microtiter plates (Costar, Cambridge, MA) were incubated with the indicated concentrations of fibronectin or a recombinant ICAM-1 IgG chimera (a gift from Joel Hayflick and Pat Hoffman, Icos, Bothell, WA) overnight at 4°C and then blocked with HBSS/2.5% bovine serum albumin (BSA). Transformed lymphoblasts were labeled with 2 µg/mL Calcein-AM (Molecular Probes), washed according to the manufacturer's protocol, added to wells (5 × 104 cells per well with appropriate stimuli in a final volume of 100 µL), and stimulated with agonists as indicated. The lymphoblasts were then allowed to settle for 60 minutes at 4°C, warmed to 37°C for the indicated times, and nonadherent cells were removed by washing. Adherent cells were quantified by using a fluorescence plate reader (Cytofluor 2300, Millipore, Bedford, MA). The fraction of adherent cells was assessed as the level of fluorescence in the well after washing, divided by total input fluorescence, and multiplied by 100. In experiments examining divalent cation substitution, the incubations were performed as described previously for primary neutrophils, except that EDTA was not added.Lymphocyte aggregation assay EBV-transformed lymphoblasts were resuspended at a concentration of 0.85 × 106 cells/mL with PMA at final 10 7 M concentration in the aggregometer, using the same
protocol used for PMNs (previously described). PMN aggregation was
assayed in parallel as a positive control.
Western immunoblotting EBV-transformed lymphoblasts at 20 × 106 cells/mL were lysed in RIPA buffer with 1% Nonidet P-40 (Sigma Chemical) according to the manufacturer's instructions (Santa Cruz Biotechnology, Santa Cruz, CA) and centrifuged at 15 000g for 20 minutes at 4°C. For2 integrin detection, the
lysates (1 mL) were collected and 10 µg anti-2 mAb,
60.3, was added for 1 hour at 4°C. Protein A-Agarose (20 µL, Santa
Cruz Biotechnology) was added, and the lysates were kept at 4°C, with
mixing overnight. The agarose pellet was washed 4 times with RIPA
buffer and then boiled with 40 µL electrophoresis sample buffer for
90 seconds. Western blot analysis was performed as previously
described50 using goat antimouse IgG conjugated to
horseradish peroxidase as the secondary antibody (BioSource
International, Camarillo, CA) and ECL Western blot detection reagents
(Amersham Life Science, Arlington Heights, IL). For 1
subunit detection, 6 × 106 of the patient-derived
EBV-transformed lymphoblasts (LADV cells), control EBV-transformed
lymphoblasts, and Jurkat T cells were lysed in 100 µL RIPA buffer,
passed through an 18-gauge needle 3 times, placed in boiling water for
2 minutes, and stored at  20°C. Lysates in nonreducing buffer were
then run on a 7.5% SDS-PAGE gel, and Western blot analysis was
performed as described previously using the anti-1 mAb
MAB 13 as the primary antibody.
Sequence analysis of 1 integrin in 3 overlapping fragments. The 5' end
fragment was amplified by using the following primers: adapter primer 1 (supplied from Marathon cDNA Amplification Kit), 5'-CCA TCC TAA TAC GAC
TCA CTA TAG GGC and 5'-GTA GCT AAA TGG GGT GGT GCA GTT CTG. The second
fragment was amplified with the following primers: 5'-CAG CTA AGC TCA
GGA ACC CTT GCA C and 5'-CTG CAC GCG CCA CAC TCA AAT GTC. The 3' end
fragment was amplified using the following primers: 5'-CCA AAG CGA AGG
CAT CCC TGA AAG and 5'-CAG TGT TGT GGG ATT TGC ACG GGC AG. For
2 sequence, 3 pairs of primers were designed to cover
the entire coding region in overlapping fragments. The 5' end fragment
was amplified from single strand cDNA using the primers: 5'-CCA GCA CAC
CGA GGG ACA TGC TG and 5'-CGT GAC GTT GCG CCA GCC GAT TTC. The second
fragment and the 3' end fragment were amplified from double-strand cDNA
by using the following 2 pairs of primers respectively: 5'-CTG GAC GCC
ATG ATG CAG GTC GC and 5'-CTG GCC GTT GTA GCG CTC ACA GTT G, and
5'-CAA CTC CAT CAT CTG CTC AGG GCT G and 5'-CTG ACG GCC TTG TCT
TCA CCA AGT G. A "touchdown" polymerase chain reaction (PCR) was
carried out in PE GeneAmp Systems 2400, according to the instructions
from the Marathon cDNA Amplification Kit. The PCR products were ligated into the PCR2.1 vector (Invitrogen) using the manufacturer's
instructions. Individual white colonies were screened, and the plasmid
DNAs were analyzed in the University of Utah core sequencing facility. Multiple clones were sequenced for each fragment.
Clinical features of the affected infant indicated a unique syndrome of leukocyte adhesion deficiency The patient, a male infant, first presented at 3 weeks of age with anemia, thrombocytopenia, leukocytosis (35-45 × 109 cells/L), and a "blueberry muffin" rash. He was delivered after a full-term pregnancy, and his size was appropriate for gestational age. There was no family history of recurrent infection, immune deficiency, or consanguinity. Maternal screening was negative for infections, including human immunodeficiency virus and hepatitis B. At the time of presentation, assay of the infant's serum for IgM antibodies directed against cytomegalovirus, rubella, toxoplasmosis, EBV, parvovirus, human herpesvirus 6, and herpes simplex 1 and 2 were all negative. Viral cultures were also negative. In the fourth through seventh months of life, he had several bacterial infections: an infected urachal duct cyst (cultures grew Bacteroides vulgaris, Enterococcus faecalis, Enterococcus avium, and hemolytic streptococcus), pneumonia (B vulgaris and Candida
albicans were cultured from the bronchoalveolar lavage fluid),
nonhemolytic streptococcal septicemia, and a perirectal ulceration and
cellulitis without pus collection that grew Escherichia coli
on culture. He also had episodes of pneumonia caused by respiratory
syncitial and parainfluenza viruses, oral candidiasis, and a C
albicans urinary tract infection, and pneumonia caused by
Pneumocystis carinii. The leukocytosis found at the time of
presentation persisted until bone marrow transplant, and leukocyte
counts as high as 96 × 109 cells/L were recorded.
Because the persistent leukocytosis and recurrent severe infections
suggested the possibility of LAD-1, analysis of cell surface
2 integrins on the infant's circulating neutrophils was
performed. Normal levels were detected (Table 1 and data not shown). Studies of
adhesive function of the child's circulating neutrophils were then
performed as described in detail below. A biopsy of the "blueberry
muffin" rash showed dermal erythropoiesis. At 6 weeks of age, the
infant had normal quantitative IgG, IgA, and IgM levels. CD4 and CD8 T
lymphocyte numbers were normal. At 6 months of age, the T-lymphocyte
function was assessed, and there were positive responses in vitro to
the mitogens phytohemagglutinin, concanavalin A, and pokeweed and to
specific antigens (tetanus and candida) after immunization and
infection, respectively.
In addition to the leukocyte abnormalities, the child was anemic and thrombocytopenic from the time of presentation. Later, the patient had hepatosplenomegaly and mucosal bleeding develop coincident with omphalitis caused by Bacteroides and an enterococcal species. In vitro analysis of the patient's platelet function demonstrated that platelet suspensions did not aggregate in response to collagen, adenosine diphosphate, arachidonic acid, or ristocetin, and there was absent clot retraction. The template bleeding time was more than 15 minutes. A bone marrow biopsy at 1 month of age revealed that the marrow was markedly hypocellular, but there were no dysplastic changes. A bone marrow biopsy at 7 months of age revealed trilineage dysplasia most pronounced in the granulocytic elements. Cytogenetic analysis was normal; specifically, monosomy 7 and deletion of 5q, 8q, or 17q alterations that are associated with myeloid malignancies,51,52 were not detected. At 8 months, the patient underwent successful bone marrow transplantation using an HLA-matched sibling donor. Three years after engraftment, he has no clinical or laboratory evidence for an immune deficit. A recent circulating leukocyte count was 10.7 × 109 cells/L with an absolute neutrophil count of 3.9 × 109 cells/L. There have been no further severe bacterial or viral infections. Primary neutrophils isolated from the subject expressed
2 integrin marker, M2
(Table 1). In addition, the surface levels of M and
2 increased 1.5-fold in response to stimulation of the
PMNs with PMA. Therefore, we assessed the adhesive functions of the
patient's neutrophils by using in vitro assays. There was dramatic
impairment of adhesion to gelatin, which depends on inside-out signaling of 2 integrins,34,35,53 when
freshly isolated PMNs from the subject were stimulated with
N-formyl-met-leu-phe (fMLP) and compared with neutrophils from a
control subject (Figure 1). The
impairment in adhesion was similar to that seen previously in studies
of neutrophils with absent surface expression of 2 integrins from a patient with LAD-1, and when neutrophils from control
subjects were treated with function-blocking anti-2
integrin antibodies.33-35 A similar impairment in adhesion
was seen when the patient's PMNs were stimulated with
platelet-activating factor (PAF), which is recognized by a
G-protein-linked receptor on the leukocyte surface that is different
from the receptor for fMLP,54 and when the neutrophils
were treated with the receptor-independent agonists PMA and calcium
ionophore A23187 (not shown). Stimulated adhesion of the patient's
neutrophils to fibrinogen, a ligand for
M2 and
X2,6 and to monolayers of
cultured human endothelial cells was also dramatically impaired
compared with that of control PMNs (Figure 1 and data not shown),
although there was a small increase in adhesion in some
experiments.
The patient's neutrophils were found to express PSGL-1, the fucosylated leukocyte counterligand that recognizes the endothelial tethering molecules P-selectin and E-selectin,55 with cell surface levels that were greater than those on neutrophils from a control subject studied in parallel (Table 1). We assessed the function of PSGL-1 on the patient's PMNs and found that they adhered normally to purified, immobilized P-selectin in adhesion assays conducted as previously described33 (not shown). The adhesion was inhibited by a specific monoclonal antibody against P-selectin. This indicated that the PSGL-1 detected on the surfaces of the PMNs was functional and that the infant did not have LAD-2.22,25 In a second experiment that confirmed this conclusion, the patient's neutrophils adhered to purified immobilized E-selectin, and this was inhibited by an anti-E-selectin antibody (not shown). We then examined the performance of the patient's neutrophils in 2 additional assays that measure
In parallel studies, we found that intracellular calcium transients in response to fMLP or calcium ionophore A23187,33 polymerization of actin in response to PMA as measured with FITC-conjugated phalloidin,45 superoxide generation in response to PMA,33 and shedding of L-selectin in response to PMA57 were each intact in the patient's neutrophils when compared with the PMNs from control subjects (data not shown). In addition, the patient's PMNs polarized in solution in response to fMLP (not shown). These experiments excluded a generalized defect in surface receptor function, intracellular signaling cascades, and signal-dependent effector functions of the PMNs from the subject. We then asked whether the extracellular domains of integrins on the
infant's PMNs could recognize ligands in the presence of the exogenous
cation Mn2+, a manipulation that eliminates the requirement
for cellular activation and inside-out signaling via cytoplasmic
pathways,37,38,58 or in response to an antibody that can
induce ligand recognition. By using the protocols described
previously,38 we found that, when Mn2+ was
substituted for Ca++ and Mg2+ in a suspension
of the patient's neutrophils, there was increased adhesiveness in 2 of
3 experiments. The increase in adhesion of the infant's PMNs to
immobilized ligand was 2- to 6-fold compared with a 4- to 8-fold
increase in adhesion of control neutrophils (data not shown). In a
second experiment, we used KIM 185, a function-perturbing mAb against
Transformed lymphoblasts from the affected subject have defective
adhesive functions mediated by 2 and L, the dominant
subunit paired with 2 on lymphocytes and
lymphoblasts,60 were expressed at levels equal to or
greater than those found on the surfaces of control lymphoblastoid
cells from age-matched subjects (Figure 5). In addition, the integrin
1 chain and the 4 and 5
subunits were expressed on LADV lymphoblasts at levels equivalent to
those on control lymphoblastoid cells (Figure 5). LADV cells and
age-matched control lymphoblastoid cells each shed 50% or more of
their surface L-selectin when treated with 107 M
PMA,61 and the intracellular calcium transients triggered by treatment with calcium ionophore were similar in both LADV cells and
control lymphoblasts (not shown). In addition, LADV cells
demonstrated a calcium transient in response to concanavalin A (not
shown). These features indicated that the LADV cells display surface
integrins at the expected density and perform signal-dependent functions and therefore could be used as surrogate cells in additional studies of the adhesive defects in the patient's leukocytes.
LADV cells did not aggregate when treated with PMA, whereas control
lymphoblastoid cells aggregated as expected (not shown). In a second
assay of their adhesive function, we examined the binding of LADV cells
to an immobilized ICAM-1/IgG chimera. ICAM-1 is a ligand for
We then asked whether the defect was specific for
Inside-out signaling triggers clustering of integrins and consequent
increased avidity for ligands, in addition to triggering increased
affinity as a result of altered conformation.15,64-66 Release of integrins from cytoskeletal constraints by treating cells
with cytochalasins also induces increased avidity and consequent adhesion under some conditions.65,66 We found that
treatment of LADV cells with low concentrations of cytochalasin D
enhanced their adhesion to fibronectin in response to PMA, but did not correct the defect in adhesion in comparison to control lymphoblastoid cells (Figure 8). This result indicates
that cytoskeletal interactions of
The adhesion defects in LADV cells are not a result of altered
structures of 2 protein from LADV cells demonstrated a band of the
expected molecular size, 95 kd (Figure
9A). Similarly, analysis of the
1 protein from LADV cells revealed a band at the
expected size, 110 kd, on a nonreduced gel (Figure 9B). Further
analysis of 2 cloned from LADV cells revealed that the
cDNA sequence was normal in all clones examined (2 each) for fragments
2 and 3, and 4 of 5 clones for the 5' proximal fragment (fragment 1, details in "Patient, materials, and methods"). The sequences of the
1 cDNA from LADV and control lymphoblastoid cells (2 clones each) were identical, but in each case, there was a histidine at
position 112 rather than threonine as in the original published
sequence.67-69 This polymorphism has previously been
reported in human cDNA clones (Hillier et al, GenBank accession
AA112739, 1995). Position 112 is outside of the known 1
ligand-binding regions.70 In addition, the control
EBV-transformed lymphoblasts exhibited adhesion to fibronectin (Figure
7), indicating that this polymorphism generated a functional protein.
Thus, structural alterations in 1 and 2 do not account for the adhesion defects in the LADV cells and appear
not to account for the abnormalities of integrin-dependent adhesion in
primary leukocytes from the patient.
|
1 and
Top
Abstract
Introduction
and anti-IgG1a,
) or from a control subject
(
) studied in parallel. This figure represents a single experiment
(n = 2 for each condition) and is representative of 9 studies, using
one or more of the indicated surfaces, each of which demonstrated
decreased adhesion of the patient's PMNs compared with control
neutrophils. The error bars in this figure represent the
range.
) were incubated in wells coated with immobilized ICAM-1 for
15 minutes at 37°C. In parallel, the cells were treated with buffer,
PMA (10
