| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Institut National de la Santé et de la Recherche Médicale, Unité 332, Institut Cochin de Genetique Moleculaire, Paris, France
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We previously showed that HIV-1 can cross the epithelium in vitro by transcytosis across epithelial cells, the most abundant cell type at mucosal surfaces (10, 11). In these studies, we reconstituted the first step of HIV-1 infection in an in vitro model using a monostratified epithelial cell line barrier. PBMCs infected with HIV-1 primary isolates (HIV-1+ PBMCs) adhered to the apical surface of the epithelial cells (12), establishing a microenvironment similar to the "immunological synapse" at the dendritic cell-CD4+ T lymphocyte interface during Ag presentation (13). Contact between HIV-1-infected PBMCs and epithelial cells induces massive polarized budding of HIV-1 in the microenvironment of the synapse. The newly formed virus is rapidly internalized by the epithelial cells, transcytosed to the basolateral pole of the cell, and released as infectious virions into the basolateral (serosal) environment where the virions could infect lamina propria mononuclear cells. The relevance of this in vitro model using human transformed cell lines as target cells to in vivo early mechanism of HIV-1 entry was confirmed using specimens of human mucosal tissue (11), as infection in human is absolutely nonethical.
Two molecules are required for the transcytosis of HIV-1 across CD4-negative epithelial cells (10): the HIV-1 surface envelope glycoprotein subunit gp120 and glycosphingolipid galactosyl ceramide (Gal Cer),3 which is thought to act as gp120 epithelial cell receptor (14, 15). Gal Cer is markedly enriched at the apical surface of epithelial cells (16) and participates in the establishment of microdomains referred to as rafts, which act as platforms for endocytosis (17, 18) and transcytosis (19, 20).
Here we show that a conserved epitope on gp41, the transmembrane subunit of the HIV envelope glycoprotein, mediates gp41 binding to epithelial cell Gal Cer and that S-IgA neutralizes this binding, thereby preventing HIV-1 transcytosis across epithelial cell monolayers. In addition, we identify a conserved epitope on gp41 to which the neutralizing S-IgA is directed and that this conserved epitope is immunodominant for the humoral mucosal response in HIV-seropositive patients. Finally, we demonstrate that both the binding of HIV-1 envelope glycoproteins to Gal Cer and the organization of Gal Cer in raft microdomains is required for effective HIV-1 transcytosis.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Cervicovaginal secretions were collected from eight HIV-1(clade E)-seropositive women at the Institut Pasteur de Pnom Phen, Cambodia. Secretions were collected after 2 days of sexual abstinence using 3 ml of sterile PBS. Samples were centrifuged and frozen and stored at -80°C. Sample contamination by sperm was measured using the seminal fluid detection assay (SEMA; Humangen Fertility Diagnostics, Charlottesville, VA) according to the manufacturer, but optimized by increasing the incubation times and using o-phenylenediamine dihydrochloride as substrate. Contaminated samples were discarded. Colostrum samples were collected from five HIV-1(clade A)-infected women aged 25–40 years from the Institut Pasteur de Bangui, Cambodia (21). All subjects were at non-AIDS stages of HIV infection. Control specimens consisted of two pools of 10 colostrums from healthy seronegative volunteers. S-IgA and IgG were purified as we described earlier (21). 2F5 is a human monoclonal IgG (26). Monoclonal human IgGs C31 anti-gp120 and D1 anti-gp41 used in ligand blotting experiments were obtained from immortalized PBMCs of an HIV-1-seropositive patient (62, 63).
Antigens
Recombinant HIV-1 envelope glycoprotein gp160 HIV-1 MN/LAI (VV.TG.9150) was obtained from Pasteur-Mérieux Sérum et Vaccins (Lyon, France) or from Advanced BioScience Laboratories (Kensington, MD). Recombinant HIV-1-gp 41 and gp120 were obtained from Intracell (Cambridge, MA).
ELISA
ELISAs were performed as described previously (21). Briefly, plates were coated overnight at 4°C with 100 µl/well of either recombinant HIV-1 envelope glycoprotein gp160 (2.5 µg/ml) or HIV-1 peptides (0.2 µg/ml) diluted in PBS. After blocking the plates with 3% skimmed milk and washing, purified S-IgA or IgG were added and incubated for 1 h at 37°C. After washing, rabbit HRP-labeled Abs to human F(ab')2 (1/2000) were added for 1 h at 37°C. HRP activity was detected with o-phenylenediamine (Sigma, St. Louis, MO) and read at 492 nm with a Titertek Multiskan spectrophotometer (Flow Laboratories, Glasgow, U.K.). The results were compared with a standard curve of a pool of whey from human milk containing 460 µg/ml of S-IgA. A large pool of 250 normal human sera containing 14 mg/ml of IgG was used as serum standard. The specific activities of S-IgA and of IgG against gp160 were expressed as arbitrary units per mg of Igs.
The ratio of S-IgA/IgG anti-gp160 = S-IgA specific activity for gp160/IgG specific activity for gp160 represents the relative amount of Abs in the two Ig subclasses. For the peptide specific activity analysis, the results for each colostrum sample and cervicovaginal secretion (#) are given by the ratio of S-IgA/IgG anti-peptide = Qi/ratio of S-IgA/IgG anti-gp160, where Qi = S-IgA specific activity for peptide i/IgG specific activity for peptide i is the ratio of the specific activity of S-IgA to that of IgG for the various HIV-1 envelope glycoprotein peptide i. Therefore, the ratio of S-IgA/IgG anti-peptide is Qi normalized to the total amount of anti-HIV-1 envelope glycoprotein Abs contained in S-IgA subclass relative to IgG.
HIV-1+ PBMC
HIV-1+ PBMC were prepared as described previously (11). Primary isolates were obtained by culture of PBMCs from HIV-1-infected Cameroonese women (subtype A, nonsyncytia-inducing phenotype) and from HIV-1-infected Italian mothers (subtype B, syncytia-inducing and nonsyncytia-inducing phenotypes) (64).
Epithelial cell transcytosis assay and neutralization of
transcytosis by S-IgA and IgG or by methyl-
-cyclodextrin (mCD)
HIV-1 transcytosis across epithelial cells and neutralization of
transcytosis by Abs was performed as previously described
(21). Briefly, the endometrial cell line HEC-1 or
intestinal cell line HT-29 clone 19 (65) cells expressing
apical Gal Cer (data not shown) were each grown as a tight, polarized
monolayer for 7 days on a permeable filter support (0.45-µm pore
size) forming the interface between two independent chambers, the upper
one bathing the apical (luminal) surface of the epithelial monolayer,
and the lower one bathing the basolateral (serosal) surface. Purified
colostrum S-IgA and IgG (20 µg), or the 2F5 IgG (.1 µg) were added
to the apical chamber and incubated for 10 min at 37°C. To initiate
virus transcytosis, 2 x 106
HIV-1+ PBMCs were added to the apical chamber.
Contact between HIV-1+ PBMCs and the epithelial
cell monolayer resulted in rapid budding of the HIV-1-virions, followed
by their transcytosis from the apical to the basolateral pole of the
epithelial cells. After 2 h, inhibition of transcytosis by Ab was
determined by detection of p24 in the basolateral medium by ELISA
(Pasteur Sanofi, France). After 2 h, the level of p24 in the
absence of Ab or in the presence of control nonimmune clostral S-IgA or
IgG was measured and found to be similar with a value of 150 pg/ml.
This value was taken as 100% of transcytosis and used to express the
results. When indicated, Ig was preincubated with the ELDKWA peptide to
block the recognition site on the Ig molecule as described
(11). As control, an irrelevant ELDKWA scrambled peptide
was used. After overnight binding at 4°C, remaining unbound peptide
was dialyzed out before addition of the Ig fraction as described above.
Assays were performed in duplicate for each of the variables: HIV
subtype (PBMCs infected with HIV-1 subtype A or B subtypes) and cell
lines (HT29 or HEC-1), which is between four and ten different
conditions. Interassay variation was <10%. Fig. 2
shows the mean
values of these combined experiments.
|
CD (Sigma), optimal drug concentration and incubation time
were determined using previously described conditions (29, 66, 67). Tight HT29 cell monolayers were preincubated first for 30
min at 37°C in RPMI 1640 supplemented with 0.25% fatty acid-free BSA
(Boehringer Manheim, Mannheim, Germany). mCD was then added at 1, 5,
and 10 mM, and the cell monolayers were incubated up to 4 h at
37°C. Transepithelial resistance (TER) was measured every hour
(10). Whereas no significant differences between the
control (no mCD) and drug-treated monolayer could be detected after
1-h treatment with 10 mM mCD (110 ± 5 ohm x
cm2), TER started gradually to decrease, reaching
10 ± 4 ohm x cm2 after 4-h incubation
(data not shown). mCD (5 mM) gave only a 10–15% reduction in TER.
Full reversal of these effects were observed for all conditions after
washing out the drug and reculturing for 12 h in RPMI 1640
supplemented with 10% FCS, except for incubation for 4 h with 10
mM mCD when TER did not recover. For these studies, epithelial
monolayers were pretreated for 1 h at 1–10 mM mCD, cells then
were washed free of drug, and transcytosis was assayed. TER was
measured at the end of the assay and was not significantly different
from the values under standard conditions (110 ± 5 ohm x
cm2). The results are expressed as the mean
percentage of the transcytosis inhibition obtained by at least three
independent experiments performed in duplicate as compared with
standard conditions in the absence of drug. To test the reversibility
of the cholesterol depletion, after 1 h of drug treatment, the
cells were washed free of drug and recultured for 12 h before
transcytosis was measured. Binding of recombinant envelope glycoprotein to epithelial cells
Recombinant envelope glycoproteins, gp41 or gp120 (4 nM) were added to the apical pole of confluent, tight monolayers of HT29 or HEC-1 epithelial cells pre-equilibrated in RPMI 1640, 10% FCS at 12°C. The interaction of proteins with epithelial cells was performed first at 12°C for 30 min to maximize the binding without incurring endocytosis. Then, the epithelial monolayer system was transferred to 37°C for 8 min to allow proper organization of raft microdomains and firm attachment of the recombinant viral envelope glycoproteins to the cells. Next, the cell system was transferred to 4°C, and unbound proteins were removed by extensive washing (4x) for 30 min. Cell monolayers were then fixed with 4% paraformaldehyde, processed for immunofluorescence, and analyzed by confocal microscopy as described previously (11). Primary Abs were the specific mAbs D1, directed against gp41 and C31, directed against gp120 (both at 5 µg/ml). Images presented are projections of consecutive sections (<10, with a step of 0.5 µm between two consecutive sections) corresponding to the apical domain of the epithelial cells above tight junctions (21). When indicated, a 100 molar excess ELDKWA or anti-Gal Cer mAb (68) at a concentration known to inhibit HIV-1 transcytosis (21) was added to the apical medium 10 min before addition of recombinant proteins. Addition of an irrelevant peptide at similar concentration into the apical medium did not affect the binding pattern of recombinant HIV-1 envelope glycoproteins (data not shown).
Ligand blotting
Ligand blotting was performed as previously described
(69). Briefly, lipids, Gal Cer (Avanti Polar Lipids,
Alabaster, AL), monosialoganglioside GM1 (Sigma), or
phosphatidylcholine (PC; Serdary Research Laboratories, London,
Ontario, Canada) were dissolved in chloroform/methanol (1:1) at 5 mg/ml
and spotted onto nitrocellulose (NC) supports (Bio-Rad, Hercules, CA).
The NC strips were incubated for 1 h at room temperature in PBS
containing 8% BSA to saturate nonspecific binding sites. Recombinant
gp160, gp41, or gp120 dissolved at 4 nM in PBS containing 1% BSA were
allowed to bind to immobilized lipids for 4 h at room temperature.
After extensive washing in PBS containing 1% BSA and 0.05% Tween 20,
ligand binding was detected with specific mAbs to gp41 (D1) or SU-gp120
(C31) (5 µg/ml) for 16 h at 4°C followed by the Vector Super
ELITE kit (Burlingame, CA) as described by the manufacturer using goat
anti-mouse biotinylated mAb and avidin-HRP. Finally, HRP activity
was detected with ECL (Amersham, Little Chalfont, U.K.) and
autoradiography on Amersham RNP6H film. Exposure time of the
autoradiograms was <1 min. For CT-B (the B subunit of cholera toxin)
ligand blotting (27), biotinylated CT-B (Sigma) at 2
µg/ml was used, and detection was conducted with streptavidin as
described above. When indicated, clostral IgG or S-IgA (#63 or 64) were
added together with the viral glycoproteins. In some cases, the various
Ig were preincubated with the ELDKWA peptide binding in an attempt to
block the corresponding recognition site on the Ig molecule. As results
obtained in the presence of the two Ab samples were identical, only
experiments with colostrum #63 are shown on Fig. 4.
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We previously reported that clostral IgG and S-IgA from
HIV-1-seropositive women bind to the HIV-1 envelope glycoprotein
precursor gp160 (21). We now investigate the epitope(s) on
gp160 that were recognized by the S-IgA and Ig, using Abs from both
colostrum and cervicovaginal secretions (Fig. 1, b and d). The
specific activity of S-IgA and IgG for four epitopes on the HIV-1
envelope glycoproteins implicated in HIV-1 infection was analyzed. Two
epitopes were on gp120: a V3 loop (MN isolate) hypervariable region and
the C5 constant region (22) shown to be associated with
neutralization or with slow HIV-1 progression. The other two epitopes
corresponded to conserved gp41 motifs: the principal immunodominant
domain (PID) (23) and the ELDKWA motif. The ELDKWA motif
(aa 662–667) is a conserved epitope recognized by human serum
IgG 2F5, the only Ig thus far identified that is directed against the
gp41 and capable of neutralizing infection by a large majority of
primary HIV-1 isolates (24). ELDKWA is part of the larger
peptide DP178 (also called T-20) that has been shown to inhibit HIV-1
infection at nanomolar concentrations in cell culture experiments and
to have an antiviral effect in a phase I clinical trial in humans
(25). We previously showed that this epitope was used by
both dimeric IgA and IgM isolated from HIV-1-seropositive patients to
neutralize intracellularly HIV-1 transcytosis by epithelial cells
(11). The results in Fig. 1 clearly indicate that for each
of the five colostrums and seven of eight cervicovaginal
secretions tested, S-IgA anti-HIV-1 envelope bound mainly to the
ELDKWA epitope of gp41. In contrast, IgG anti-HIV-1 envelope from
the same specimen of secretion failed to bind this gp41 domain. For the
one cervicovaginal sample (Fig. 1d, I18), the amount of
S-IgA specific for ELDKWA was still significant and 13 times more
abundant than the paired IgG. This difference in Ig subclass binding to
ELDKWA peptide is due to a higher S-IgA and lower IgG reactivity for
each sample. In contrast, the reactivity for gp160 was always higher
for IgG than for IgA (Fig. 1, a and c). Indeed,
the ratio of clostral S-IgA anti-gp160 to clostral IgG
anti-gp160 was 0.015–0.1 (Fig. 1a), indicating a much
lower quantity of S-IgA than IgG specific for HIV envelope glycoprotein
in colostrum. A similar analysis of cervicovaginal S-IgA and IgG
confirmed the ratio (ratio of S-IgA/IgG anti-gp160 <1) (Fig. 1c), indicative of the higher IgG anti-gp160 content as
compared with S-IgA anti-gp160 in mucosal secretions.
|
We next investigated whether polyclonal S-IgA and IgG purified
from the colostrum of HIV-1-seropositive women could neutralize
transcytosis of HIV-1. For these experiments, we used our previously
described transcytosis assay and PBMCs acutely infected with primary
isolates of HIV-1 of subtype A or B (HIV-1+
PBMCs). The ability of mucosal Ig to inhibit transcytosis of HIV-1
was higher for S-IgA (80% mean reduction) than for IgG (49% mean
reduction) (Fig. 2
). The amount of virus
produced at the apical pole of the epithelial barrier during the assay
was similar in the absence or presence of the Abs (data not shown),
indicating that the Abs did not alter PBMC virus production. In
addition, no difference in neutralization of transcytosis was detected
using PBMCs acutely infected with HIV-1 isolates of subtype A or B.
Clostral S-IgA and IgG were purified from seropositive women infected
with HIV-1 subtype A but could equally neutralize transcytosis of
primary isolates of subtypes A, B, and D. This strongly suggested that
the epitopes recognized by the neutralizing Abs are conserved regions
of HIV-1 envelope glycoprotein. No difference in transcytosis could be
detected using HEC-1 endometrial or HT29 (clone 19) intestinal cell
lines. Such inhibitory efficiency for S-IgA even though not complete,
is biologically significant, considering that the Ig samples were
polyclonal; consequently, titration of specific Ab-mediated inhibition
could not be performed.
Because S-IgA but not IgG from the different colostrums (Fig. 1b) and cervicovaginal secretions (Fig. 1d)
strongly recognized the ELDKWA peptide, we investigate whether this
epitope was recognized by the colostrum Igs that neutralized HIV-1
transcytosis.
In experiments designed to block neutralization, equal amounts of
purified IgG and S-IgA were preincubated with ELDKWA peptide or with a
ELDKWA scrambled irrelevant peptide for control (Fig. 2). After
dialysis to eliminate excess free peptide, the Ig samples were assayed
for their capacity to neutralize HIV-1 transcytosis. Neutralization was
significantly decreased when S-IgA, but not IgG, had been preincubated
with the ELDKWA peptide implicating ELDKWA as key peptide involved in
S-IgA inhibition of HIV-1 transcytosis. Transcytosis neutralization was
not inhibited by an excess of peptide inadequately removed by dialysis
because under the conditions referred to as "standard + ELDKWA"
(Fig. 2), HIV-1 transcytosis was not affected. In this condition,
before addition to the cell system, the ELDKWA peptide was incubated
alone (without Abs) at 4°C overnight followed by dialysis with a
cutoff at 10 kDa to remove the free peptide. Incubation with the ELDKWA
scrambled peptide did not alter the neutralization efficiency,
indicating the specificity of the ELDKWA peptide sequence in
Ab-mediated inhibition of transcytosis. The ability of ELDKWA to more
strongly diminish transcytosis inhibition by S-IgA than by IgG (55 vs
5%, respectively) is consistent with the ability of S-IgA to bind
ELDKWA more strongly than IgG (Fig. 1, b and d).
The amount of colostrum 48 and cervicovaginal fluids were
insufficient for analysis in the transcytosis assay.
Expressing HIV-1 transcytosis inhibition as a function of the specific
activity of the Ig for ELDKWA (Fig. 1) revealed that the capacity to
neutralize transcytosis and the involvement of ELDKWA in this process
is much more efficient for S-IgA than IgG. Prevention of neutralization
by ELDKWA was affected neither by the HIV-1 subtype used to infect
PBMCs nor by the origin of the epithelial cell line used in the
transcytosis assay. These results reinforce the concept that ELDKWA is
a crucial epitope in HIV-1 infection. Indeed, it is the minimal
hexapeptide recognized by the transcytosis-neutralizing clostral S-IgA
(present studies), dimeric IgA (11), and neutralizing 2F5
IgG (26).
Both HIV-1 envelope gp41 and gp120 subunits interact with epithelial cell Gal Cer
The ability of anti-ELDKWA Abs to inhibit HIV-1 transcytosis suggested that the hexapeptide epitope of gp41 participates in the interaction between HIV-1 and its epithelial cell target. The glycosphingolipid Gal Cer has been shown to act as the gp120 receptor on epithelial cells (14, 15). We have previously shown that this interaction likely plays a role in transcytosis because Abs to both anti-gp120 and anti-Gal Cer completely blocked HIV-1 transcytosis (10). Based on structural and thermodynamic considerations (A. A. and M. B., manuscript in preparation), we investigated whether gp41 also interacts with Gal Cer and whether this interaction involves hexapeptide ELDKWA.
First, as shown morphologically in Fig. 3, recombinant gp120 as well as gp41 bind
to the apical surface of a tight epithelial cell barrier. Both
interactions were antagonized by monoclonal anti-Gal Cer Abs. In
contrast, only the gp41 interaction with the apical epithelial surface
was inhibited by a 100 molar excess of the ELDKWA peptide, whereas the
binding of gp120 was not altered by this peptide. That gp41 bound
specifically to Gal Cer via ELDKWA is also supported by the inability
of an irrelevant or an ELDKWA scrambled peptide to impair this
interaction and the lack of binding of control human IgA to the apical
surface of the cell (data not shown).
|
a). No binding to other
glycolipids, such as the ganglioside GM1, or a phospholipid, such as
PC, was detected (Fig. 4a). As a control for specificity, we
also showed that CT-B bound only its known receptor GM1
(27) and not Gal Cer or PC (Fig. 4a).
Third, a 100 molar excess of the ELDKWA peptide completely inhibited
the interaction of gp41 with Gal Cer (Fig. 4a, arrow) but
not the interaction between gp120 and Gal Cer. ELDKWA scrambled peptide
did not inhibit binding of gp41 to Gal Cer, and the binding of CT-B to
GM1 was not sensitive to the presence of the ELDKWA peptide, further
indicating that gp41 binding with Gal Cer involved ELDKWA.
We next compared the capacity of S-IgA and its paired IgG to neutralize
HIV-1 transcytosis (Fig. 2) with their capacity to inhibit the binding
of gp41 to Gal Cer. In this experiment, ligand blotting was performed
in the presence of sample 63 or 64 S-IgA or IgG. Igs from these
patients were chosen because a relatively large quantity of material
was available, S-IgA bound ELDKWA substantially better than its paired
IgG (ratio of 
20, see Fig. 1b), and the S-IgA was highly
effective in neutralizing HIV-1 transcytosis (>84%, see Fig. 2).
Results obtained with samples from patients 63 and 64 were identical,
and only those obtained with sample from patient 63 are shown on Fig. 4b. As shown in Fig. 4b, S-IgA alone efficiently
blocked gp41 binding to Gal Cer (Fig. 4b, arrow) but not
gp120 binding to Gal Cer. Additionally, Gal Cer binding of both gp41
and gp120 was unaffected by paired clostral IgG. Finally, we addressed
whether the capacity of S-IgA to inhibit gp41 binding to Gal Cer could
be ascribed to the efficiency of S-IgA binding to ELDKWA (Fig. 1).
S-IgA (63 or 64) or paired IgG were preincubated with ELDKWA, and
unbound peptide was removed by dialysis, as in Fig. 2. The Igs were
assayed for inhibiting gp41 or gp120 binding to Gal Cer. S-IgA, not
matched IgG, preincubated with ELDKWA had lost its capacity to inhibit
gp41 binding to Gal Cer (Fig. 4b). As additional control for
specificity, an ELDKWA scrambled irrelevant peptide was used and had no
effect (data not shown). These results indicate that ELDKWA is a gp41
epitope functionally involved in HIV-1 binding to and transcytosis
across epithelial cells and that ELDKWA is the epitope used by the
S-IgA for neutralizing HIV-1 transcytosis.
Disruption of the organization of raft microdomains containing Gal Cer inhibits HIV-1 transcytosis
Glycolipids (mainly Gal Cer) and cholesterol form dynamic
microdomains functioning at the apical epithelial membrane (16, 28). These microdomains create floating rafts in otherwise
phospholipid-rich membrane environments and are involved in membrane
trafficking to and from the apical membrane, including apical
endocytosis (17, 20) and transcytosis (19, 20). These transient microdomains have been shown to be
stabilized by cholesterol (29), and depletion of
cholesterol from cell membrane destabilized these raft
microdomains (29), therefore impairing their functional
activity. To investigate the functional role of Gal Cer containing raft
microdomains in HIV-1 transcytosis, the raft organization was disrupted
by cholesterol depletion using mCD.
We first established the optimal conditions for mCD disruption of
raft microdomains without affecting the integrity and tightness of the
epithelial barrier (see Materials and Methods). Treatment
for 1 h at 37°C with 1–10 mM mCD, concentrations usually
used to disrupted the domain (29), did not alter the TER
of the epithelial barrier (Fig. 5). The
drug was then removed and HIV-1 transcytosis was assessed as described
above. As shown in Fig. 5, mCD inhibited HIV-1 transcytosis in a
concentration-dependent manner with almost total inhibition at 10 mM.
mCD pretreatment of epithelial monolayers did not modify the apical
production by HIV-1+-PBMCs of HIV-1 presented for
transcytosis (data not shown), and therefore could not account for the
inhibition of transcytosis. The effect of mCD on transcytosis was
reversible when the drug was removed and epithelial cells were allowed
to replenish their cholesterol content for 12 h prior the onset of
the transcytosis assay (data not shown). These data suggest that HIV-1
transcytosis required the interaction of HIV-1 envelope glycoprotein
subunits with the apical epithelial Gal Cer as well as the functional
organization of Gal Cer in raft microdomains stabilized by
cholesterol.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The HIV-1 envelope is organized as an oligomeric structure comprising multimers of gp120 and gp41 molecules, referred to as "spikes," in which gp120 interacts with and masks a large part of gp41 (32, 33, 34). Our results, combined with those of others (14, 15), indicate that both subunits interact with the epithelial target cell by binding Gal Cer at the apical surface of the epithelial cells. This agrees with the proposed physical organization of the spike in which both the Gal Cer binding sites, the one on gp120 (15, 32) and the one described here involving ELDKWA on the external portion of the gp41 C-terminal helix (26), are exposed and therefore accessible to neutralizing Abs.
The lipid composition of the apical epithelial membrane exhibits unique features. Its external leaflet is highly enriched in glycosphingolipids (35, 36). Gal Cer, a monohexosylceramide, is a glycosphingolipid characteristic of the apical membrane epithelial cell. Transient lateral assemblies of glycosphingolipids stabilized by cholesterol, referred to as raft microdomains (18, 37, 38), function in epithelial cells as platforms for trafficking toward the apical membrane as well as in apical endocytosis (17, 20). Raft microdomains are also involved in transcytosis across endothelial cells (39). Apical lipid rafts invaginate forming caveolar structures that contain the cholesterol-binding protein caveolin (40, 41). Such dynamic structures are endocytosed allowing endocytosis and transcytosis (20). Cholesterol depletion completely abolishes the formation of apical caveolae and blocks internalization and transcytosis of clustered raft proteins (20), indicating the stabilizing role of cholesterol in raft structure and the highly dynamic nature of raft microdomains.
Here we have shown that the interaction between HIV-1, and more specifically ELDKWA, on gp41 with epithelial cells is dependent on the raft organization on the apical epithelial membrane. Our studies reinforce the concept that HIV-1 is endocytosed at the apical surface and then transcytosed to the basolateral surface. First, HIV-1 envelope glycoprotein subunits bind specifically to Gal Cer, an essential component of raft structure and dynamics. Second, disruption of the organization of raft microdomains completely blocks transepithelial transcytosis of HIV-1. HIV-1 may enter epithelial cells via caveolar structures. The distance between the two Gal Cer binding sites on gp120 and gp41, estimated to be <7 nm (32, 34), is in good agreement with the proposed size of a stabilized raft microdomain having an estimated size between 3 and 50 nm in diameter (29, 37, 42). Therefore, the interaction of one spike of gp120-gp41 multimers at the surface of the virion with one domain of the target epithelial cell membrane could be a cooperative one. The primary contact of the gp120 part of the spike with a Gal Cer microdomain would stabilize this raft domain, allowing a favorable orientation and interaction of the gp41 exposed peptide within the same space. However, the involvement of clathrin-coated structures cannot be entirely excluded yet, as cholesterol depletion has also been shown to affect clathrin-dependent endocytosis in epithelial cells, although to a much lesser extent (20).
We and others have previously shown that transcytosis of the virus budding from HIV-1-infected mononuclear cells is dependent on the contact between the HIV-1-infected cell and the apical epithelial cell surface (10, 11, 12). Transcytosis, but not the viral budding event, is inhibited by Arg-Gly-Asp added to the apical surface. Arg-Gly-Asp peptides broadly inhibit cell-cell contact mediated by integrin/disintegrin interactions (M.B., unpublished data). Because epithelial cells are known to express adhesion molecules like ICAM-1 apically, these results suggest that the HIV-1 infection of mononuclear cells, which up-regulates surface expression of integrins like LFA-1 (43), interact with the target epithelial cell via integrin-disintegrin interactions. Such interactions would circumscribe a microenvironment between the two cells, which resembles the recently described immunological synapse (13, 44, 45). Actin is required for polarized HIV-1 budding from HIV-1-infected cells at the epithelial cell surface (46). Similarly, the dynamics of actin filaments polymerization, induced upon cell-cell contact and mediated by integrins, play an important role in the formation of the immunological synapse. Actin filaments organize the synapse into distinct supramolecular activation clusters that favor the initiation of signal transduction cascades. The actin scaffold would recruit or stabilize specialized membrane domains enriched in glycolipids (13, 47). The interaction of HIV-1-infected mononuclear cells with epithelial cell inducing HIV production and transcytosis may be similar to the formation of immunological synapse for T cell activation. This adhesion mechanism would overcome the resistance to cell-cell contact posed by the negatively charged glycocalyx present also on these cells. In summary, as described for the immunological synapse (13), adhesion of HIV-1-infected cells to the apical epithelial cell surface may induce a rearrangement of the epithelial actin cytoskeleton that, in turn, would facilitate recruitment and stabilization of the glycolipid raft microdomains by the HIV-1 envelope and initiate the cascade of events leading to endocytosis of the virus.
Several gp120 and gp41 epitopes have been shown to be the target of neutralizing Abs. 2F5, a monoclonal IgG purified from peripheral B cells of an HIV-1-seropositive patient (24, 48) that recognizes the epitope ELDKWA on gp41, neutralizes infection by a large array of HIV-1 primary isolates both in vitro and in vivo by passive immunization (30, 31). However, 2F5-like Abs that possess both immunochemical specificity for ELDKWA and neutralization activity are seldom present in the sera of HIV-1-infected patients, and numerous attempts to produce such neutralizing IgG in vitro or in animals in vivo have failed (49). In contrast, our results show that S-IgA, but not IgG, specific for ELDKWA and possessing neutralizing activity for HIV-1 transcytosis is usually present in secretions from HIV-1-seropositive patients. Thus, ELDKWA appears to be immunodominant for a mucosal but not for a systemic immune response.
Why are neutralizing systemic IgG not generated against ELDKWA, whereas in the context of TM-gp41 this peptide appears immunogenic for humoral mucosal immunity and elicits S-IgAs that neutralize transcytosis of HIV? Several non-mutually exclusive parameters can be invoked.
First, the structure of neutralizing Abs: the multimeric structure of S-IgA, as compared with monomeric IgG, most likely plays a role in the affinity/avidity of the Ab for the target (50) and, therefore, for the neutralizing function. HIV-neutralizing IgGs do not have a remarkably rapid mutational rate nor do they reveal striking pressure for replacement mutations in V region framework regions and complementary-determining regions. However, both the neutralizing 2F5 and 2G12 IgGs that recognize, respectively, the ELDKWA peptide and a region on SU-gp120 carry an unusual heavy chain complementary-determining region 3 that might mediate their respective functions and probably also account for their rare occurrence in patient sera (49). The existence of such specific structures in neutralizing S-IgA remains to be established.
Second, the compartmentalized induction of mucosal vs systemic humoral immunity (51, 52), and within the mucosal compartment itself, as has been recently demonstrated (53, 54, 55). However, mechanisms responsible for compartmentalization remain unclear.
Third, the tolerance induced by self-peptide (56): self-peptide could originate from a soluble protein like the human complement factor H present at high concentration in the plasma, which has been shown to bind to TM-gp41 at a site overlapping the 2F5 epitope (57).
Fourth, the structure of the epitope that elicits neutralizing Abs: ELDKWA is the minimal hexapeptide recognized by neutralizing 2F5 IgG (26) by clostral S-IgA (present studies) and by dimeric IgA as we have shown earlier (11). However, injection of this hexapeptide in animal models never leads to the induction of neutralizing Abs, whatever the site of induction and the class of Ab (IgG or IgA) induced. This indicates that either the immunogenically functional epitope is larger than the hexapeptide and/or that the conformational structure of this peptide within TM-gp41 is of great importance for its immunogenicity. Furthermore, even if the ELDKWA peptide has been found identically conserved in >85% of HIV of the main viral subtypes (A-H), some rare isolates harbor in this region a partial conservation with the leucine (L) and the tryptophan (W), as is also the case for HIV-2 (EHO) and SIV (mac239) (58).This suggests that not only the sequence but also the structure of the region encompassing ELDKWA is of importance and could define a mimotope, essential for HIV transcytosis via Gal Cer binding and induction of humoral immunity. It is striking that the overlapping peptide DP178/T20 (aa 638–673) is able to neutralize infection of cells in vitro by primary HIV isolates, and has been recently shown significantly to reduce the level of circulating HIV when administered to patients for 14 days (25). Indeed, DP178/T20 has been shown to adopt a helical structure and to assemble in tri- or tetrameric oligomers. Furthermore, the amino acid sequence 650–684 of TM-gp41 contains a stack of aromatic hydrophobic residues (W and F) that extend the DP178 sequence at its C-terminal end up to the transmembrane region (aa 684) (59). These tryptophans have been very recently found to be essential for HIV interaction with its target, at a post binding but pre-fusion step (58) as is the case here, because HIV do not fuse with the epithelial cell membrane during transcytosis (10). This suggests that this stack contributes to the structure of the sequence starting at the ELDKWA peptide and extending N terminally to the charged polar amino acid up to the reported Ca2+ binding sites (60) by organizing the covering water surface layer. It would create a water layer between the peptide and the cell membrane that in turn would facilitate charge transfer processes (A. A. and M. B, manuscript in preparation) being at the origin of the signal for endocytosis (61).
The data presented here suggest that the ELDKWA hexapeptide of gp41 is a potential candidate Ag for a mucosal HIV-1 vaccine. First, the peptide is a conserved epitope. Second, ELDKWA is involved in the binding of gp41 to its epithelial receptor Gal Cer and is functionally involved in HIV-1 translocation across a tight epithelial barrier. Third, ELDKWA is an immunodominant Ag in the mucosal humoral response in seropositive patients. Fourth, S-IgA developed against the hexapeptide functionally blocks transcytosis, likely by interfering with the binding of gp41 via ELDKWA to epithelial Gal Cer. Fifth, as part of the DP178/T20 peptide that inhibits HIV-1 infection in cell culture experiments, ELDKWA is reported to have an antiviral effect in a phase I clinical trial in humans (25). Finally, the peptide is recognized by the 2F5 IgG that recently was shown to participate in the protection of macaque against SIV or chimeric HIV-1/SIV mucosal infection (30, 31) when used in a passive immunization protocol. Although the precise structure of the ELDKWA containing peptide that elicits in vivo such neutralizing S-IgA has to be determined, the information presented here indicates the critical role the ELDKWA plays in the interaction of HIV-1 envelope glycoproteins and epithelial cells, the first step in epithelial cell translocation of HIV-1.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Morgane Bomsel, Institut Cochin de Genetique Moleculaire, 22 rue Mechain, 75014 Paris, France.
3 Abbreviations used in this paper: Gal Cer, glycosphingolipid galactosyl ceramide; HIV-1+ PBMC, PBMC infected with HIV-1 primary isolates; mCD; methyl--cyclodextrin; TER, transepithelial resistance; NC, nitrocellulose; CT-B, the B subunit of cholera toxin; PC, phosphatidylcholine.
Received for publication December 26, 2000. Accepted for publication March 14, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1-acidic glycoprotein in the continuous microvascular endothelium. Proc. Natl. Acad. Sci. USA 95:6175.
synergistically increase total polymeric IgA receptor levels in human intestinal epithelial cells: role of protein tyrosine kinases. J. Immunol. 156:4807.[Abstract]
This article has been cited by other articles:
|
|
 |
|
  S. Bryson, J.-P. Julien, R. C. Hynes, and E. F. Pai Crystallographic Definition of the Epitope Promiscuity of the Broadly Neutralizing Anti-Human Immunodeficiency Virus Type 1 Antibody 2F5: Vaccine Design Implications J. Virol., November 15, 2009; 83(22): 11862 - 11875. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. F. Arnold, P. K. Velasco, A. K. Holmes, T. Wrin, S. C. Geisler, P. Phung, Y. Tian, D. A. Resnick, X. Ma, T. M. Mariano, et al. Broad Neutralization of Human Immunodeficiency Virus Type 1 (HIV-1) Elicited from Human Rhinoviruses That Display the HIV-1 gp41 ELDKWA Epitope J. Virol., May 15, 2009; 83(10): 5087 - 5100. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Wright, M. E. Lamm, and Y. T. Huang Excretion of Human Immunodeficiency Virus Type 1 through Polarized Epithelium by Immunoglobulin A J. Virol., December 1, 2008; 82(23): 11526 - 11535. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Coutant, H. Yu, M.-J. Clement, A. Alfsen, F. Toma, P. A. Curmi, and M. Bomsel Both lipid environment and pH are critical for determining physiological solution structure of 3-D-conserved epitopes of the HIV-1 gp41-MPER peptide P1 FASEB J, December 1, 2008; 22(12): 4338 - 4351. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-H. Huang, Z. Qi, F. Wu, L. Kotula, S. Jiang, and Y.-H. Chen Interaction of HIV-1 gp41 Core with NPF Motif in Epsin: IMPLICATION IN ENDOCYTOSIS OF HIV J. Biol. Chem., May 30, 2008; 283(22): 14994 - 15002. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Montero, N. E. van Houten, X. Wang, and J. K. Scott The Membrane-Proximal External Region of the Human Immunodeficiency Virus Type 1 Envelope: Dominant Site of Antibody Neutralization and Target for Vaccine Design Microbiol. Mol. Biol. Rev., March 1, 2008; 72(1): 54 - 84. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Muratori, A. Sistigu, E. Ruggiero, M. Falchi, I. Bacigalupo, C. Palladino, E. Toschi, and M. Federico Macrophages Transmit Human Immunodeficiency Virus Type 1 Products to CD4-Negative Cells: Involvement of Matrix Metalloproteinase 9 J. Virol., September 1, 2007; 81(17): 9078 - 9087. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. Lewis, R. J. Pleass, M. R. Batten, J. D. Atkin, and J. M. Woof Structural Requirements for the Interaction of Human IgA with the Human Polymeric Ig Receptor J. Immunol., November 15, 2005; 175(10): 6694 - 6701. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Alfsen, H. Yu, A. Magerus-Chatinet, A. Schmitt, and M. Bomsel HIV-1-infected Blood Mononuclear Cells Form an Integrin- and Agrin-dependent Viral Synapse to Induce Efficient HIV-1 Transcytosis across Epithelial Cell Monolayer Mol. Biol. Cell, September 1, 2005; 16(9): 4267 - 4279. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. A. Schneeman, M. E. C. Bruno, H. Schjerven, F.-E. Johansen, L. Chady, and C. S. Kaetzel Regulation of the Polymeric Ig Receptor by Signaling through TLRs 3 and 4: Linking Innate and Adaptive Immune Responses J. Immunol., July 1, 2005; 175(1): 376 - 384. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Buonaguro, M. L. Visciano, M. L. Tornesello, M. Tagliamonte, B. Biryahwaho, and F. M. Buonaguro Induction of Systemic and Mucosal Cross-Clade Neutralizing Antibodies in BALB/c Mice Immunized with Human Immunodeficiency Virus Type 1 Clade A Virus-Like Particles Administered by Different Routes of Inoculation J. Virol., June 1, 2005; 79(11): 7059 - 7067. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. T. Huang, A. Wright, X. Gao, L. Kulick, H. Yan, and M. E. Lamm Intraepithelial Cell Neutralization of HIV-1 Replication by IgA J. Immunol., April 15, 2005; 174(8): 4828 - 4835. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Devito, B. Zuber, U. Schroder, R. Benthin, K. Okuda, K. Broliden, B. Wahren, and J. Hinkula Intranasal HIV-1-gp160-DNA/gp41 Peptide Prime-Boost Immunization Regimen in Mice Results in Long-Term HIV-1 Neutralizing Humoral Mucosal and Systemic Immunity J. Immunol., December 1, 2004; 173(11): 7078 - 7089. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-M. Kang, L. Guo, Q. Yao, I. Skountzou, and R. W. Compans Intranasal Immunization with Inactivated Influenza Virus Enhances Immune Responses to Coadministered Simian-Human Immunodeficiency Virus-Like Particle Antigens J. Virol., September 15, 2004; 78(18): 9624 - 9632. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Matoba, A. Magerus, B. C. Geyer, Y. Zhang, M. Muralidharan, A. Alfsen, C. J. Arntzen, M. Bomsel, and T. S. Mor A mucosally targeted subunit vaccine candidate eliciting HIV-1 transcytosis-blocking Abs PNAS, September 14, 2004; 101(37): 13584 - 13589. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Zhang, Y. Huang, R. Fayad, G. T. Spear, and L. Qiao Induction of Mucosal and Systemic Neutralizing Antibodies against Human Immunodeficiency Virus Type 1 (HIV-1) by Oral Immunization with Bovine Papillomavirus-HIV-1 gp41 Chimeric Virus-Like Particles J. Virol., August 1, 2004; 78(15): 8342 - 8348. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-M. Kang, Q. Yao, L. Guo, and R. W. Compans Mucosal Immunization with Virus-Like Particles of Simian Immunodeficiency Virus Conjugated with Cholera Toxin Subunit B J. Virol., September 15, 2003; 77(18): 9823 - 9830. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. E. Cummins Jr., J. M. Villanueva, T. Evans-Strickfaden, S. M. Sesay, S. R. Abner, T. J. Bush, T. A. Green, J. L. Lennox, T. Wright, T. M. Folks, et al. Detection of Infectious Human Immunodeficiency Virus Type 1 in Female Genital Secretions by a Short-Term Culture Method J. Clin. Microbiol., September 1, 2003; 41(9): 4081 - 4088. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. I. Albu, A. Jones-Trower, A. M. Woron, K. Stellrecht, C. C. Broder, and D. W. Metzger Intranasal Vaccination Using Interleukin-12 and Cholera Toxin Subunit B as Adjuvants To Enhance Mucosal and Systemic Immunity to Human Immunodeficiency Virus Type 1 Glycoproteins J. Virol., May 15, 2003; 77(10): 5589 - 5597. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Wolbank, R. Kunert, G. Stiegler, and H. Katinger Characterization of Human Class-Switched Polymeric (Immunoglobulin M [IgM] and IgA) Anti-Human Immunodeficiency Virus Type 1 Antibodies 2F5 and 2G12 J. Virol., April 1, 2003; 77(7): 4095 - 4103. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B.-L. Waarts, R. Bittman, and J. Wilschut Sphingolipid and Cholesterol Dependence of Alphavirus Membrane Fusion. LACK OF CORRELATION WITH LIPID RAFT FORMATION IN TARGET LIPOSOMES J. Biol. Chem., October 4, 2002; 277(41): 38141 - 38147. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. van Meer and Q. Lisman Sphingolipid Transport: Rafts and Translocators J. Biol. Chem., July 12, 2002; 277(29): 25855 - 25858. [Full Text] [PDF]
|
|
|
|
|
|
A. Alfsen and M. Bomsel HIV-1 gp41 Envelope Residues 650-685 Exposed on Native Virus Act as a Lectin to Bind Epithelial Cell Galactosyl Ceramide J. Biol. Chem., July 5, 2002; 277(28): 25649 - 25659. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) of transcytosis and is expressed as
the percentage of inhibition of the TER compared with standard
condition of transcytosis in the absence of drug. Bars correspond to
mean ± SEM of data pooled from at least three independent
experiments performed in duplicate.