HIV-1 escape from a small molecule, CCR5-specific entry inhibitor does not involve CXCR4 use.
ABSTRACT To study HIV-1 escape from a coreceptor antagonist, the R5 primary isolate CC1/85 was passaged in peripheral blood mononuclear cells with increasing concentrations of the CCR5-specific small molecule inhibitor, AD101. By 19 passages, an escape mutant emerged with a >20,000-fold resistance to AD101. This virus was cross-resistant to a related inhibitor, SCH-C, and partially resistant to RANTES but still sensitive to CCR5-specific mAbs. The resistant phenotype was stable; the mutant virus retained AD101 resistance during nine additional passages of culture in the absence of inhibitor. Replication of the escape mutant in peripheral blood mononuclear cells completely depended on CCR5 expression and did not occur in cells from CCR5-Delta32 homozygous individuals. The escape mutant was unable to use CXCR4 or any other tested coreceptor to enter transfected cells. Acquisition of CXCR4 use is not the dominant in vitro escape pathway for a small molecule CCR5 entry inhibitor. Instead, HIV-1 acquires the ability to use CCR5 despite the inhibitor, first by requiring lower levels of CCR5 for entry and then probably by using the drug-bound form of the receptor.
- SourceAvailable from: Emily J Platt[show abstract] [hide abstract]
ABSTRACT: In addition to the primary cell surface receptor CD4, CCR5 or another coreceptor is necessary for infections by human immunodeficiency virus type 1 (HIV-1), yet the mechanisms of coreceptor function and their stoichiometries in the infection pathway remain substantially unknown. To address these issues, we studied the effects of CCR5 concentrations on HIV-1 infections using wild-type CCR5 and two attenuated mutant CCR5s, one with the mutation Y14N at a critical tyrosine sulfation site in the amino terminus and one with the mutation G163R in extracellular loop 2. The Y14N mutation converted a YYT sequence at positions 14 to 16 to an NYT consensus site for N-linked glycosylation, and the mutant protein was shown to be glycosylated at that position. The relationships between HIV-1 infectivity values and CCR5 concentrations took the form of sigmoidal (S-shaped) curves, which were dramatically altered in different ways by these mutations. Both mutations shifted the curves by factors of approximately 30- to 150-fold along the CCR5 concentration axis, consistent with evidence that they reduce affinities of virus for the coreceptor. In addition, the Y14N mutation specifically reduced the maximum efficiencies of infection that could be obtained at saturating CCR5 concentrations. The sigmoidal curves for all R5 HIV-1 isolates were quantitatively consistent with a simple mathematical model, implying that CCR5s reversibly associate with cell surface HIV-1 in a concentration-dependent manner, that approximately four to six CCR5s assemble around the virus to form a complex needed for infection, and that both mutations inhibit assembly of this complex but only the Y14N mutation also significantly reduces its ability to successfully mediate HIV-1 infections. Although several alternative models would be compatible with our data, a common feature of these alternatives is the cooperation of multiple CCR5s in the HIV-1 infection pathway. This cooperativity will need to be considered in future studies to address in detail the mechanism of CCR5-mediated HIV-1 membrane fusion.Journal of Virology 09/2000; 74(15):7005-15. · 5.08 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: The beta-chemokine receptor CCR5 is considered to be an attractive target for inhibition of macrophage-tropic (CCR5-using or R5) HIV-1 replication because individuals having a nonfunctional receptor (a homozygous 32-bp deletion in the CCR5 coding region) are apparently normal but resistant to infection with R5 HIV-1. In this study, we found that TAK-779, a nonpeptide compound with a small molecular weight (Mr 531.13), antagonized the binding of RANTES (regulated on activation, normal T cell expressed and secreted) to CCR5-expressing Chinese hamster ovary cells and blocked CCR5-mediated Ca2+ signaling at nanomolar concentrations. The inhibition of beta-chemokine receptors by TAK-779 appeared to be specific to CCR5 because the compound antagonized CCR2b to a lesser extent but did not affect CCR1, CCR3, or CCR4. Consequently, TAK-779 displayed highly potent and selective inhibition of R5 HIV-1 replication without showing any cytotoxicity to the host cells. The compound inhibited the replication of R5 HIV-1 clinical isolates as well as a laboratory strain at a concentration of 1.6-3.7 nM in peripheral blood mononuclear cells, though it was totally inactive against T-cell line-tropic (CXCR4-using or X4) HIV-1.Proceedings of the National Academy of Sciences 06/1999; 96(10):5698-703. · 9.74 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Although HIV-1 gene expression is detected in naive, resting T cells in vivo, such cells are resistant to productive infection in vitro. However, we found that the endogenous microenvironment of human lymphoid tissues supports de novo infection and depletion of this population. Cell cycle analysis and DNA labeling experiments established that these cells were definitively quiescent and thus infected de novo. Quantitation of the "burst size" within naive cells further demonstrated that these cells were productively infected and contributed to the local viral burden. These findings demonstrate that lymphoid tissues support active HIV-1 replication in resting, naive T cells. Moreover, these cells are not solely reservoirs of latent virus but are permissive hosts for viral replication that likely targets them for elimination.Immunity 11/2001; 15(4):671-82. · 19.80 Impact Factor
HIV-1 escape from a small molecule, CCR5-specific
entry inhibitor does not involve CXCR4 use
Alexandra Trkola*†, Shawn E. Kuhmann†‡, Julie M. Strizki§, Elizabeth Maxwell‡, Tom Ketas‡, Tom Morgan‡,
Pavel Pugach‡, Serena Xu§, Lisa Wojcik§, Jayaram Tagat§, Anandan Palani§, Sherry Shapiro§, John W. Clader§,
Stuart McCombie§, Gregory R. Reyes§, Bahige M. Baroudy§, and John P. Moore‡¶
*Division of Infectious Diseases, Department of Medicine, University Hospital Zurich, 8091 Zurich, Switzerland;‡Weill Medical College of Cornell University,
New York, NY 10021; and§Schering Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033
Edited by Malcolm A. Martin, National Institutes of Health, Bethesda, MD, and approved November 14, 2001 (received for review October 2, 2001)
To study HIV-1 escape from a coreceptor antagonist, the R5 primary
isolate CC1?85 was passaged in peripheral blood mononuclear cells
with increasing concentrations of the CCR5-specific small molecule
inhibitor, AD101. By 19 passages, an escape mutant emerged with a
>20,000-fold resistance to AD101. This virus was cross-resistant to a
related inhibitor, SCH-C, and partially resistant to RANTES but still
sensitive to CCR5-specific mAbs. The resistant phenotype was stable;
the mutant virus retained AD101 resistance during nine additional
passages of culture in the absence of inhibitor. Replication of the
escape mutant in peripheral blood mononuclear cells completely
depended on CCR5 expression and did not occur in cells from CCR5-
?32 homozygous individuals. The escape mutant was unable to use
CXCR4 or any other tested coreceptor to enter transfected cells.
for a small molecule CCR5 entry inhibitor. Instead, HIV-1 acquires the
of CCR5 for entry and then probably by using the drug-bound form
of the receptor.
clinical development as potential therapies for HIV-1 infection (1,
2). These inhibitors include the gp41-targeted peptides T20 and
T1249, the gp120-targeted recombinant protein CD4-IgG2, the
peptides, and mAbs specific for the chemokine receptors CXCR4
and CCR5 (reviewed in refs. 1 and 2). The latter proteins act as
coreceptors with CD4 during the process of HIV-1 entry (3). If the
into clinically useful drugs, then a new group of compounds to
existing protease and reverse transcriptase inhibitors (1, 2).
There are many hurdles to overcome in the clinical development
of any compound that shows activity against HIV-1 replication in
vitro. As well as the traditional issues of toxicology and pharma-
cology, a problem common to all HIV-1 inhibitors is the rapid
development of drug resistance both in vitro and in vivo. Clinical
it is likely that the virus also will escape from combinations of
inhibitors, particularly if therapy is suboptimal. Therefore it is
prudent to study the escape pathways that are adopted by HIV-1 in
inhibitor is used clinically.
The issue of escape pathways is of particular importance with
inhibitors of HIV-1 entry via CCR5 because of a well documented
facet of HIV-1 pathogenesis. Almost all cases of HIV-1 transmis-
sion involve strains that use CCR5 for entry (R5 viruses); these
viruses persist throughout the course of HIV-1 infection in most
infected people and are pathogenic. However, in up to 50% of
also known as syncytium-inducing (SI) viruses, are associated with
a more rapid disease course exemplified by an accelerated rate of
new generation of antiviral compounds, collectively termed
entry inhibitors, is presently undergoing active preclinical and
CD4?T cell loss (reviewed in ref. 5). This loss may be because the
ability to use CXCR4 allows the virus to better target naive CD4?
T cells and?or more effectively inhibit T cell production (6, 7).
Because of the ability of R5 viruses to undergo phenotypic
evolution to acquire CXCR4 usage, there are concerns that block-
ing CCR5 with a specific inhibitor in vivo might force HIV-1 to
evolve to use CXCR4 instead (8). This outcome would be unde-
the escape pathways used by HIV-1 when replicating in peripheral
blood mononuclear cells (PBMCs) under the selection pressure of
a CCR5-specific small molecule inhibitor, AD101. We used an R5
virus isolate (HIV-1 CC1?85) that we knew to be capable of
undergoing phenotypic evolution to CXCR4 usage. We found that
the AD101 escape mutant of this virus did not use CXCR4 but
instead gained the ability to use CCR5 in an AD101-insensitive
Materials and Methods
Viruses and Other Reagents. Mitogen-activated PBMCs were pre-
Sciences University, Portland, OR; ref. 10). GHOST-coreceptor
cell lines were obtained from D. Littman (New York University,
New York) and maintained as described (11). HIV-1 CC1?85 and
CC2?86 isolates were from R. Connor (Aaron Diamond AIDS
Research Center, New York; ref. 12). Stocks of isolates NL4-3,
anti-CCR5 mAb 2D7 was from PharMingen (13), and the anti-
CCR5 mAb PA14 was from W. Olson (Progenics, Tarrytown, NY;
Generation of AD101 Escape Mutant. HIV-1 CC1?85 (1,000 tissue
culture 50% infective doses per ml) was added to 20 ml of
mitogen-activated PBMCs (2 ? 106?ml) with sufficient AD101 to
cause ?90% inhibition. Control cultures lacked AD101 but other-
The cultures were passaged weekly by adding a 5-ml aliquot of
PBMCs, maintaining a constant density of cells throughout the
experiment. On day 4 postpassage, AD101 was added to the
indicated final concentration. At each passage, p24 antigen pro-
duction was monitored to ensure that virus replication was occur-
ring and to determine the extent of inhibition by AD101. Because
This paper was submitted directly (Track II) to the PNAS office.
Abbreviations: PBMC, peripheral blood mononuclear cell; rfu, relative fluorescence unit.
†A.T. and S.E.K. contributed equally to this work.
¶To whom reprint requests should be addressed at: Joan and Sanford I. Weill Medical
College of Cornell University, Department of Microbiology and Immunology, 1300 York
Avenue, W-805, New York, NY 10021. E-mail: email@example.com.
The publication costs of this article were defrayed in part by page charge payment. This
article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C.
§1734 solely to indicate this fact.
www.pnas.org?cgi?doi?10.1073?pnas.012519099 PNAS ?
January 8, 2002 ?
vol. 99 ?
no. 1 ?
PBMCs from a different donor were used at each passage, p24
production varies in both the AD101-treated and control cultures.
when viral replication began to increase in the AD101-treated
for drug sensitivity and coreceptor utilization studies.
assayed in primary CD4?T cells prepared by positive selection
from freshly activated PBMCs (9); the results were comparable to
those obtained by using PBMC cultures (data not shown). Individ-
ual data points within an assay were derived from duplicate wells,
and the results presented represent an average of 3–6 assays. The
p24 concentration in the culture supernatants was monitored as
calculated as a percentage of that produced in the absence of
inhibitors. Assays using CCR5?32??32cells were performed as
described above but by using PBMCs from normal or CCR5?32??32
donors rather than CD4?T cells. Each data point is derived from
3–9 wells, with the average value shown. Infection of GHOST
indicator cells that stably express CD4 and different coreceptors
and inducibly express green fluorescent protein after HIV-1 infec-
tion was performed as described (11) except that green fluorescent
protein fluorescence in cell lysates prepared 3 days postinfection
was measured by using a microplate fluorometer. By adding
found that fluorescence increases ?5 relative fluorescence units
(rfus) above background could be detected reproducibly.
Sequence Analysis of env Genes Cloned from CC1?85 and CC101.19.
from genomic DNA purified from PBMCs infected for 4 days with
the CC1?85 and CC101.19 isolates. The env genes were amplified
with Pfu DNA polymerase (Stratagene) with the primers EnvF
agene), and individual clones were sequenced. The consensus
sequences (defined as greater than 50% amino acid identity) and
alignments were produced by using MACVECTOR (Oxford Molec-
ular, UK). The predicted gp120 amino acid sequences from the
available clones are published as supporting information on the
PNAS web site, www.pnas.org.
Infection of HeLa-CD4-CCR5 Cells. The single-round focal infectivity
assay was performed as described (10) except that the primary
antibody was an Ig fraction from HIV-1?human sera (from J.
Mascola, National Institutes of Health) at 17 ?g?ml, and the
secondary antibody was horseradish peroxidase-conjugated goat
anti-human Ig (BioSource International, Camarillo, CA). Foci of
infection were normalized to the counts derived from high CCR5
cells (clone RC.25) in the same assay. No stained cells were
observed in the parental HeLa-CD4 cell line (clone HI-R).
AD101: A Small Molecule Inhibitor of HIV-1 Entry via CCR5. SCH-C, a
small molecule CCR5 antagonist that is an effective inhibitor of
HIV-1 replication in vitro, is described elsewhere (16). SCH-C is
now being evaluated in phase I clinical trials. Similar to SCH-C,
receptor and has potent activity against a broad range of R5 HIV-1
isolates from different genetic subtypes (Fig. 1 a and b; data not
program dictated its choice for initial studies of CCR5 inhibitor
An AD101 Escape Mutant Can Be Generated in a PBMC Culture. In
designing the escape mutant experiment we wished to mimic as
closely as possible conditions relevant to the use of a CCR5
inhibitor in vivo. We therefore elected to use mitogen-stimulated
PBMCs as the target cells as opposed to a CCR5-expressing cell
line. We also chose to use a genetically heterogeneous HIV-1
primary isolate rather than a molecular or biological clone to allow
escape mutants to be selected from a quasi-species pool as well as
to be generated de novo. Moreover, we selected a primary R5
isolate, HIV-1 CC1?85 (isolated from patient case C in January
to evolve to use CXCR4; the February 1986 isolate from the same
individual (CC2?86) had an R5X4 or X4 phenotype, although the
patient’s CD4 counts remained high. By July 1986, his CD4 counts
with R5X4 or X4 phenotypes persisted in case C until his death
from AIDS in 1989 (12).
HIV-1 CC1?85 was chronically exposed to a CCR5 inhibitor by
culturing it for 22 weekly passages in the presence of increasing
AD101 concentrations. As a control, CC1?85 was passaged under
the same conditions but without AD101 to allow us to monitor
changes in the virus occurring spontaneously during prolonged
PBMC passage. The p24 concentration in the culture supernatants
was monitored with each passage. After six passages, partial
(?3-fold) resistance to AD101 was noted in the supernatants
harvested from the AD101-treated culture, and thus the AD101
peak in replication that had started after four passages with the
original AD101 concentration (Fig. 2). After 16 passages, p24
production rates in the control and AD101-treated cultures were
comparable. After this point, replication in the treated culture
could not be suppressed further by increasing the AD101 concen-
(Fig. 2). The viral supernatant from passage 19 of the AD101-
of AD101. (b) Ability of AD101 to inhibit HIV-1 replication in CD4?T cells. The
data point is the mean (?SEM) of five experiments. The viruses used were: JRFL,
?1.5 nM with a range of 0.53–3.5 nM.
www.pnas.org?cgi?doi?10.1073?pnas.012519099Trkola et al.
treated culture, designated CC101.19, was chosen for further study,
point, designated CCcon.19. Because partial resistance was appar-
ent at earlier time points, a virus from the six-passage culture also
was studied further along with its time-matched control isolate
(designated CC101.6 and CCcon.6, respectively). AD101 dose
escalation was terminated after 22 passages. The final isolate,
CC101.22, was cultured for nine passages in the absence of AD101,
yielding isolate CC101.22R9, to determine whether it had a stable
phenotype or had reverted to AD101 sensitivity (see below).
Phenotypic Properties of the AD101 Escape Mutant. In CD4?T cell
mutant was strongly (?20,000-fold) resistant to AD101, whereas
CC101.6 showed partial resistance (?3-fold; Fig. 3f), and CCcon.6
and CCcon.19 had the same sensitivity as the parental virus,
CC1?85 (Fig. 3 a and f). The escape mutant phenotype was stable;
the CC101.22R9 virus retained complete resistance to AD101
CC101.19 and CC101.22R9 viruses were cross-resistant to SCH-C,
which is in the same chemical family as AD101 (Fig. 3b). The
CC101.19, CCcon.19, and CC101.22R9 isolates also were less
sensitive (10–20-fold) to the chemically unrelated small molecule
CCR5 entry inhibitor, TAK-779 (ref. 17; data not shown). How-
ever, the parental isolate CC1?85 is only weakly inhibited by
TAK-779 (IC50? 370 nM), and all the above isolates are inhibited
by ?90% at TAK-779 concentrations of 25 ?M (data not shown).
Moreover, genetic drift in culture is a complicating factor, because
the control CCcon.19 isolate also developed partial resistance to
TAK-779. Thus, any cross-resistance to TAK-779 specifically gen-
erated by exposure to AD101 is marginal.
A modest decrease (?10-fold) in the sensitivity of CC101.19 to
the anti-CCR5 mAb, PA14, was noted. However, CC101.22R9 was
as sensitive as CC1?85 and CCcon.19 to PA14, thus the partial
resistance of CC101.19 to PA14 was not a stable phenotype (Fig.
3c). Both CC101.19 and CC101.22R9 retained the original sensi-
tivity of CC1?85 to a second anti-CCR5 mAb, 2D7 (Fig. 3d). Thus,
there may be subtle differences in how CC101.19 is inhibited by
similar anti-CCR5 mAbs.
The CC101.19 escape mutant also was substantially cross-
resistant (?20-fold) to RANTES. However, CC101.22R9 had
reverted such that its sensitivity to RANTES was comparable to
CC1?85 and CCcon.19 (Fig. 3e).
The AD101 Escape Mutant Is Unable to Use CXCR4. To determine
we evaluated its ability to replicate in PBMCs that lack CCR5, i.e.,
cells from donors homozygous for the CCR5-?32 allele (18, 19).
CC101.19, CCcon.19, CC101.22R9, or the parental isolate CC1?85
replication (p24 antigen production) at each time point is shown. When replica-
in the latter culture can no longer be suppressed by increasing the AD101
concentration, escape from AD101 has occurred.
An escape mutant for AD101. HIV-1 CC1?85 was passaged weekly by
(filled diamonds) were tested for their sensitivity to AD101 (a and f), SCH-C (b), PA14 (c), 2D7 (d), and RANTES (e). (f) The isolates used were CC1?85 (filled squares),
CC101.6 (open circles), and CCcon.6 (open triangles). The extent of virus replication is represented as a percentage of p24 antigen produced in the absence of any
and both CC1?85 and CCcon.6 is significant at an AD101 concentration of 5 nM, as indicated by a paired comparison t test (P ? 0.05, n ?5).
Trkola et al.
January 8, 2002 ?
vol. 99 ?
no. 1 ?
could replicate detectably in PBMCs from two different
CCR5?32??32donors (Fig. 4). In contrast, the reference X4 and
R5X4 viruses NL4-3 and DH123 replicated efficiently in the
the same person (Fig. 4). Replication of CC2?86, NL4-3, and
DH123 in the CCR5?32??32PBMCs was inhibited significantly by 1
in marked contrast to the CXCR4 usage of the later in vivo variant
CC2?86, replication of the CC101.19 in vitro escape mutant in
PBMCs absolutely depends on CCR5 and cannot be supported by
any other coreceptor expressed in these cells, including CXCR4.
We confirmed that CC101.19 and CC101.22R9 were unable to
use CXCR4 by showing they could not replicate in GHOST-
CXCR4 cells. In one experiment representative of three, green
fluorescent protein fluorescence increases over the background
level for uninfected GHOST-CCR5 cells (?5 rfus) was 65 rfus for
CC1?85, 10 rfus for CC101.19, 7 rfus for CCcon.19, 31 rfus for
CC101.22R9, and 64 rfus for CC2?86. None of the viruses caused
a fluorescence increase in GHOST-CXCR4 cells except for
CC2?86 (39 rfus). This infection was inhibited completely by the
GHOST cells expressing CCR1, CCR2b, CCR3, CCR4, CCR8,
GHOST-CCR3 cells (7 rfus) as reported previously (12). Overall,
we could find no evidence that CC101.19 or CC101.22R9 had
acquired the ability to use any other coreceptor when evolving to
escape from the selection pressure of AD101. We also confirmed
that CC101.6 and CC101.19 still required CD4 for entry; their
replication was fully inhibited by the anti-CD4 mAb RPA-T4 (data
Sequence Changes Associated with AD101 Resistance. To evaluate
what env sequence changes correlate with the AD101 resistance
phenotype of CC101.19, eight full-length env genes from CC1?85
and six from CC101.19 were cloned and sequenced. An additional
sequence of gp120 only was obtained also from each isolate. There
was considerable diversity among the gp120 and gp41 coding
regions of the CC1?85 clones (see the gp120 sequences, which are
published as supporting information on the PNAS web site).
Although there was slightly more diversity in the gp41 sequences of
CC1?85 than of CC101.19, no readily discernable pattern of dif-
ferences between the isolates was apparent. In contrast, the gp120
coding sequences of the CC101.19 clones showed little diversity,
suggesting that the selection pressure acts on this Env subunit
(S.E.K. and J.P.M., unpublished data). Alignment of the consensus
amino acid sequences of the gp120 subunits of the two isolates
reveals 24 differences; 22 single residue changes, one four-residue
insertion in V1, and one single-residue insertion in V5. The V5
region, which is extremely diverse in the CC1?85 isolate, has little
diversity in the CC101.19 clones (Fig. 5). When env genes from the
CC101.19 clones were inserted into the NL4-3 provirus in the
absence of any other changes, the chimeric, clonal viruses fully
recapitulate the AD101 resistance of the CC101.19 isolate (S.E.K.,
F. Lee, and J.P.M., unpublished data).
The CC101.6 and CC101.19 Isolates Have an Increased Ability to Use
Low Levels of CCR5. One possible mechanism for CC1?85 to escape
from AD101 is for the virus to develop an increased affinity for
CCR5 and thus be able to ‘‘scavenge’’ low levels of inhibitor-free
(CC1?85 and CCcon.19) and escape (CC101.6 and CC101.19)
of CD4 (10,000 molecules per cell) but different levels of CCR5
(10). The RC.10 clone of these cells expresses ?7,100 CCR5
molecules per cell, and the RC.25 clone expresses ?78,000 mole-
cules per cell (10). As a reference virus, we used the C3 variant of
HIV-1 JR-CSF, which was selected for its ability to replicate in the
expansion is the result of a single amino acid change in V1, which
present on Molt4 and SupT1 cells, while retaining the original R5
phenotype of JR-CSF (21, 22).
All the viruses infected the low CCR5 cells (RC.10) less effi-
ciently than the high CCR5 cells (RC.25) in a single-round focal
?1 (Fig. 6), as expected from previous studies (10). However, the
C3 variant more efficiently exploited the low levels of CCR5 on
RC.10 cells than did the parental JR-CSF isolate. Given the known
properties of these viruses (21, 22), this observation validates the
assay. Although all the tested isolates infected cells expressing high
levels of CD4 and CCR5 to similar efficiencies (data not shown),
the CC101.6 and CC101.19 isolates infected RC.25 cells with
?10-fold lower efficiencies than did the CC1?85 and CCcon.19
isolates. The reason for this difference is not yet clear. However, at
a constant CD4 concentration, both the early (CC101.6) and late
bars represent the SD. When 1 ?M AMD3100 was added to cultures to inhibit
entry via CXCR4, it is noted on the x axis.
The AD101 escape mutant is unable to replicate in PBMCs lacking CCR5
The sequence alignment shows the consensus amino acid sequences from nine
clones from the CC1?85 isolate (Upper) and seven clones from the CC101.19
isolate (Lower). Shaded amino acids are those which are identical between the
two consensus sequences. Dashes indicate gaps in the consensus sequences, and
Xs indicate amino acids where there was not ?50% amino acid identity among
the available clones.
www.pnas.org?cgi?doi?10.1073?pnas.012519099 Trkola et al.
could the parental CC1?85 and CCcon.19 isolates (Fig. 6).
The principal observation made in this study is that the dominant
pathway used by an HIV-1 isolate to escape from a CCR5-specific
of CCR5 in an inhibitor-insensitive manner. This observation was
made despite our use of PBMCs (which express CXCR4 and other
potential coreceptors) as the target cells during selection and
despite our choice of an HIV-1 strain that was capable of evolving
the ability to use CXCR4, based on what occurred with this virus
in vivo (12). Thus, although it can take only a few amino acid
substitutions to convert an X4 virus into an R5 virus (23, 24) and
use under the conditions of our experiment. Instead, multiple
mutations accumulated in gp120 over time, creating a virus that
selecting inhibitor, AD101. This counterintuitive finding speaks to
the overall efficiency gain for HIV-1 replication that is involved in
continued CCR5 use, compared with a switch to use of CXCR4.
Our observations are not unique to the inhibitor and virus
combination we used. An escape mutant with comparable prop-
erties was generated in response to a different small molecule
CCR5 inhibitor, SCH-C, in a PBMC culture inoculated with a
different R5 HIV-1 virus, the JR-FL molecular clone. Again, the
SCH-C escape mutant did not switch to CXCR4 use but developed
?50-fold resistance to SCH-C after 19 passages (S.X. and J.M.S.,
unpublished data). Moreover, in independent studies, the R5
HIV-1 molecular clone JR-CSF developed partial (5-fold) resis-
tance to MIP-1? when cultured in a cell line expressing both CCR5
and CXCR4 but did not acquire CXCR4 use (25). Escape mutants
with significant (?100-fold) resistance to the anti-CCR5 mAb 2D7
(26). Again, coreceptor switching did not occur, although the cell
line used for the selection process contains no known coreceptors
use was observed in a minority of human peripheral blood leuko-
cyte severe combined immunodeficient mice after infection with
JR-CSF in the prolonged presence of an N-terminally modified
chemokine derivative, NNY-RANTES (8). Whether this reflects a
specific property of this murine model is not yet known. It also may
be relevant that RANTES is not only an inhibitor of R5 virus
replication, but it also can actively enhance the replication of X4
CCR5 from cell surfaces (28), whereas AD101 and SCH-C do not
(S.X., A.T., J.P.M., J.M.S., unpublished data). Hence any receptor
a pathway unique to this class of compound. Overall, HIV-1 may
respond differently to the selection pressure of different CCR5
inhibitors, particularly to receptor agonists such as AD101 and
SCH-C and antagonists such as RANTES.
How can a CCR5 inhibitor escape mutant continue to use
CCR5? We can imagine two mechanisms that are not mutually
exclusive and may act sequentially. The first involves evolution of
the CCR5 binding site on gp120 such that the mutant virus has a
higher affinity for CCR5 and is better able to compete with AD101
or SCH-C. Under these conditions, the escape mutant would
are present despite the inhibitor, but it would do so more efficiently
to exploit the nearly undetectable levels of CCR5 on the Molt4 T
cell line, presumably by increasing the affinity of gp120 for CCR5
(21, 22). We have used this JR-CSF variant (C3) to validate a focal
infectivity assay with HeLa-CD4-CCR5 clones expressing low and
high levels of CCR5; both C3 and the CC101.6 and CC101.19
their parental isolates (Fig. 6). Hence the first stage of the AD101
escape pathway may involve the acquisition of a higher CCR5
affinity associated with partial (?3-fold) resistance, this process
of CCR5 may be better exploited by an escape mutant that had an
increased rate of fusion once it formed the virus?receptor complex
or required the presence of fewer CCR5 molecules in the
virus?receptor complex for fusion to occur (29).
Between passages 6 and 19, there is no further increase in the
ability of CC1?85 to use low levels of CCR5, but the escape mutant
acquires full resistance to AD101. We believe this second mode of
escape involves the creation of a substantially different binding site
even when the inhibitor is bound also to the receptor. This seems
to be the mechanism by which the X4 virus HIV-1 NL4-3 escapes
from the CXCR4 inhibitors AMD3100 and SDF-1? in cells that
express CXCR4 but not CCR5 (30, 31). We strongly suspect that
this is the dominant mechanism because of the magnitude of the
eventual AD101 resistance (?20,000-fold), although we have not
yet proven it.
Preliminary analyses of the sequences of viruses obtained at
intermediate time points are consistent with the escape process
proceeding by a two-step mechanism (S.E.K., J. Taylor, S. Wolin-
sky, and J.P.M., unpublished data). By passage six, 20 of the 24
differences between the consensus sequences shown in Fig. 5
dominate the gp120 sequences and apparently have been selected
from among preexisting variants in the CC1?85 population. This
selection is associated with partial (?3-fold) resistance to AD101
(Fig. 3 b and f). Between passages 6 and 16, three additional
single-residue changes occur in the V3 loop, probably de novo, and
correlate temporally with the development of full (?20,000-fold)
escape viruses and specific CCR5 mutants will enable further
studies of this process. If the later AD101 escape isolates do use the
drug-bound form of CCR5 as a coreceptor, they also must be able
to use the drug-free form efficiently, because they replicate both in
the absence and presence of AD101 and SCH-C (Fig. 3 a and b).
viruses were assayed for their ability to infect HeLa-CD4-CCR5 cells expressing
high (clone RC.25, ?78,000 CCR5 per cell) or low (clone RC.10, ?7,100 CCR5 per
cells was determined, and the ratio (low CCR5 cells?high CCR5 cells) was calcu-
lated (?SEM). The differences in the ratios between CC1?85 and CCcon.19 and
between CC101.6 and CC101.19 were not significant. However, the differences
between either of the isolates CC1?85 or CCcon.19 and the isolates CC101.6 or
CC101.19 were highly significant using an unpaired t test (P ? 0.03, n ? 8). The
shown are representative of one of three experiments.
Trkola et al.
January 8, 2002 ?
vol. 99 ?
no. 1 ?