Immunologic strategies for HIV-1
remission and eradication
Dan H. Barouch1,2and Steven G. Deeks3
Antiretroviral therapy (ART) is able to suppress HIV-1 replication indefinitely in individuals
who have access to these medications, are able to tolerate these drugs, and are
motivated to take them daily for life. However, ART is not curative. HIV-1 persists
indefinitely during ART as quiescent integrated DNA within memory CD4+T cells and
perhaps other long-lived cellular reservoirs. In this Review, we discuss the role of the
immune system in the establishment and maintenance of the latent HIV-1 reservoir. A
detailed understanding of how the host immune system shapes the size and distribution of
the viral reservoir should lead to the development of a new generation of immune-based
therapeutics, which may eventually contribute to a curative intervention.
to be established during acute infection (4–6),
although precisely when, where, and how the res-
ervoir is seeded remains to be determined. The
reservoir has a remarkably long half-life and is
resistant to antiretroviral therapy (ART), result-
ing in lifelong infection and viral rebound in the
are currently under way to understand the biol-
ogy of the viral reservoir, the mechanism of viral
latency, and the potential of various therapeutic
approaches to target the reservoir (Fig. 1). Re-
cent data indicate that the size of the viral
previously anticipated, suggesting the profound
scope of this challenge (8).
IV-1 is a retrovirus that integrates into the
host genome, primarily in memory CD4+
T cells that can harbor latent, replication-
competent HIV-1 DNA for many years
(1–3). This latent HIV-1 reservoir is thought
Role of the immune system in shaping
and maintaining the viral reservoir
HIV-1 infection causes profound and often irre-
versible changes to the adaptive and innate im-
mune system. In the absence of ART, CD4+T
cells are progressively depleted, CD8+T cells are
often expanded, and much of the immune sys-
tem is chronically activated. Some of these ab-
normalities improve during long-term ART, but
the immune system rarely returns to normal,
and chronic inflammation persists despite ART.
This state of heightened inflammation is driven
by multiple factors, including HIV-1 production,
irreversible loss of the mucosal integrity, expo-
sure to gut microbes, and an excess burden of
other pathogens such as cytomegalovirus (CMV).
The implications of this chronic inflammatory
state for overall health and the HIV-1 reservoir
are the focus of intense investigation (9).
HIV-1 preferentially infects activated memory
CD4+T cells that express the chemokine recep-
tor CCR5, although resting CD4+ T cells, naïve
CD4+T cells, and macrophages can also be in-
fected. The majority of infected and activated
CD4+T cells die quickly (10), but a small fraction
revert to a resting state and persist indefinitely
as the latent reservoir. Because ART blocks all or
nearly all new infection events, the reservoir that
exists at the time ART is initiated becomes the
reservoir that persists for the life of the indi-
vidual. This viral reservoir is maintained during
ART by the long half-life of infected memory T
cells, homeostatic proliferation of these cells
(11, 12), and perhaps low levels of cell-to-cell vi-
rus transfer (cryptic replication) (13). Recent
studies suggest that the HIV-1 genome is often
found integrated in host genes associated with
cell growth (14), indicating that HIV-1 may pro-
mote its own persistence in part by promoting
the continual expansion of latently infected cells
The viral reservoir in peripheral blood exists
predominantly in memory CD4+T cells endowed
with regenerative potential, including memory
stem cells and central memory cells (11, 17). The
reservoir also persists in potentially shorter-lived
CD4+T effector cell populations, but whether
these cells represent a stable reservoir or one
that is constantly being regenerated via prolif-
eration and differentiation is unknown. The dis-
tribution of the viral reservoir differs in tissues
compared with blood, with the frequency of in-
fection generally higher on a per-cell basis in
lymphocyte-rich tissues such as peripheral lymph
nodes, the ileum, and perhaps the spleen (18, 19).
The higher frequency of target cell infection
in these tissues may reflect cell-to-cell virus
spread (20) and/or the presence of other im-
mune cells that contribute to the maintenance
of latency (21).
In the absence of therapy, the frequency
of activated T cells is associated with the level
of viremia. During long-term suppressive
ART, a similar albeit less consistent association
11 JULY 2014 • VOL 345 ISSUE 6193
1Center for Virology and Vaccine Research, Beth Israel
Deaconess Medical Center, Harvard Medical School, Boston,
MA 02215, USA.2Ragon Institute of Massachusetts General
Hospital (MGH), Massachusetts Institute of Technology, and
Harvard University, Cambridge, MA 02139, USA.3University
of California, San Francisco, San Francisco, CA 94110, USA.
E-mail: email@example.com (D.H.B.); sdeeks@php.
“Kill” (Effector mechanism)
•Immune checkpoint blockade
Resting CD4 T cells
“Shock” (Reservoir activator)
CD4 T cell
Fig. 1. “Shock and kill” strategies. Latent HIV-1
reservoirs in resting CD4+Tcells can be activated
(“shocked”), which might make them more sus-
ceptible to be eliminated (“killed”) by immuno-
logic effector mechanisms.
on September 13, 2015
on September 13, 2015
on September 13, 2015
on September 13, 2015
on September 13, 2015
on September 13, 2015
exists, with the frequency of activated CD4+T
cells expressing human leukocyte antigen–DR
(HLA-DR), CCR5, and programmed cell death
protein-1 (PD-1) correlated with the frequency of
cells containing HIV-1 RNA or DNA (22, 23). The
mechanism for this association is unknown but
likely is bidirectional and multifactorial. Low
levels of virus replication and/or production may
lead to T cell activation. When stable ART is
intensified with the addition of a potent HIV-
1 integrase inhibitor, markers of low-level HIV-
1 replication and inflammation decline at least in
some individuals, indicating that replication
persists during ART and triggers an inflammatory
response (13, 24).
Alternatively, an inflammatory immune en-
vironment may contribute to the persistence of
the viral reservoir by a number of mechanisms
(25). T cell activation promotes cell-to-cell virus
transfer because activated cells are more likely
to produce virus and to become infected. T cell
receptor engagement by cognate antigen or cyto-
kines [e.g., interleukin (IL)–7] stimulates CD4+T
cell proliferation and the expansion of the infected
cell population (12). CD4+T cell proliferation may
in fact be the most important mechanism lead-
ing to the stability of the reservoir (Fig. 2) (11). A
chronic inflammatory environment would also
be expected to prevent the generation of optimal
HIV-1–specific immune responses. Chronic inflam-
mation stimulates potent and sustained immu-
noregulatory responses, including expansion of
T regulatory cells and up-regulation of PD-1 and
other negative regulators on effector cells (11, 26).
HIV-1–associated inflammation also stimulates
the deposition of collagen in secondary lymphoid
organs, which in turn leads to tissue fibrosis, per-
sistent immunodeficiency, and poor host clear-
ance mechanisms (27).
Understanding the complex virus-host inter-
actions that lead to the establishment and main-
tenance of the latent HIV-1 reservoir will require
advances in our fundamental understanding of
T cell biology. How are memory CD4+T cells
generated and maintained? What are the life
spans of the various CD4+T cell subsets? How
are the properties of T cells in the blood different
from those in lymphoid tissues? How do the fates
of CD4+T cells activated via cognate antigen dif-
fer from those stimulated to proliferate by homeo-
static mechanisms? How do myeloid cells, including
dendritic cells, contribute to the generation and
maintenance of T cell memory and latency? How
will inhibitors of the pathways that regulate T cell
survival and function change the life span of in-
dividual cell types?
There are also a number of critical unanswered
questions related to the immunology of HIV-1
persistence. Because the size, distribution, and
stability of the viral reservoir during ART are
determined in part by the host immune system,
factors that influence immune function—including
gender, ethnicity, chronic inflammatory diseases,
chronic infectious diseases, obesity, and substance
abuse—are all likely to have important but un-
known effects on the viral reservoir. The age
at which the virus is acquired and ART is ini-
tiated is also likely to be important. ART ad-
ministered at 30 hours of age in a perinatally
infected infant resulted in sustained remission
and a possible cure (28), and the administration
of ART in the first 3 months of life was asso-
ciated with a sustained decay in the reservoir
(29). Given that the development of memory
perhaps limited by immunoregulatory responses
(30), it is possible that seeding the viral re-
servoir may be less efficient in infants (31). This
hypothesis remains a largely unexploredareaof
HIV-1 cure research.
The Berlin and Boston patients
Much of the enthusiasm for HIV-1 cure research
is derived from the case of the “Berlin patient”
(32). In 2007, an HIV-1–infected adult living in
Berlin developed acute myelogenous leukemia
and received an allogeneic hematopoietic stem
cell transplant from a donor who carried a homo-
zygous deletion of CCR5, which renders cells
highly resistant to HIV-1 infection. The procedure
worked, and the patient has appeared to be free
of both cancer and HIV-1 for over 5 years (33).
Despite intense efforts, this outcome has not yet
been repeated. Finding suitable HLA-matched
donors homozyogous for the CCR5 deletion is
challenging, and the few transplants that have
been performed have failed because of recur-
rence of the underlying cancer or emergence of
CXCR4-tropic viruses. Nevertheless, this case has
galvanized gene therapy efforts to modify CD4+T
cells (34) and potentially also stem cells (35, 36) to
render them resistant to HIV-1 infection.
A related approach is to use ART to protect
transplanted donor cells until full chimerism
occurs. This strategy is illustrated by two cases
of HIV-1–infected adults in Boston who received
hematopoietic stem cell transplants for treat-
ment of refractory lymphoma (37). Under the
coverage of ART, full donor chimerism was ap-
parently achieved, with the donor immune cells
eventually replacing the original immune cells
over a period of years. As that occurred, HIV-1
DNA gradually declined to undetectable levels.
Therapy was then interrupted in both individ-
uals. Virus failed to rebound within the first few
weeks, leading to initial optimism that these
individuals may have been cured. Unfortunate-
ly, virus then dramatically rebounded at weeks
12 and 32 after treatment interruption, indicating
that the transplant caused a profound but in-
complete elimination of the viral reservoir. This
failure highlights the need either to eradicate all
replication-competent virus or to enhance the
capacity of the host immune system to contain
the limited residual virus that may persist after
a curative intervention. These observations also
illustrate the limitation of our current biomarkers
for HIV-1 persistence. The development and val-
idation of a highly sensitive biomarker that can
reliably detect and quantify the total body bur-
den of replication-competent HIV-1 during ART
is one of the highest current priorities of the field.
Comparing the outcomes of the Berlin and
Boston cases is informative. Graft versus host dis-
ease (GVHD) occurred in all three cases and likely
contributed to reductions of the viral reservoirs.
As has been observed with beneficial graft-versus-
tumor effects, alloreactive donor cells targeting
hematopoietic cells likely reduced the number of
recipient CD4+T cells harboring latent HIV-1.
These cases have also led to increased interest in
modifying host responses with immune-based
therapeutics as part of a curative strategy (38).
Assuming that the Berlin patient achieved a
complete sterilizing cure and that ART was fully
effective in the Boston cases, why was the virus
eradicated only in the former case? One key dif-
ference in the transplant protocols was the ex-
tent of pretransplant myeloablation, which would
be expected to reduce the size of the reservoir.
The Berlin patient received an aggressive regi-
men of chemotherapy and total body irradia-
tion, whereas a milder approach was used in the
Boston cases. The Berlin patient also received
more aggressive immunosuppressive therapies
for GVHD, and, as outlined below, suppressing
T cell activation and proliferation might result
in further control of the reservoir.
“Shock and kill”
Hematopoietic stem cell transplantation is too
risky and too complex for the majority of HIV-1–
infected individuals worldwide. Other approaches
that have generated considerable enthusiasm in-
clude pharmacologic strategies to induce latently
infected cells to produce virus (“shock”) together
with interventions that would enhance the abil-
ity of the host to clear these virus-producing cells
(“kill”) (Fig. 1). Histone deacetylase (HDAC) in-
hibitors have been shown to increase production
of HIV-1 RNA and, to a lesser degree, virus parti-
cles from the viral reservoir in vivo (39). However,
the magnitude of the effect of HDAC inhibitors
has to date been modest, and this class of drugs has
not yet demonstrated a consistent effect on the
frequency of cells that harbor replication-competent
HIV-1 (40–42). Other classes of antilatency drugs
and immunomodulators are therefore being ex-
plored for their capacity to stimulate the viral
reservoir (40, 42, 43).
To augment the capacity of the host to elim-
inate reservoir cells after activation, several im-
munologic strategies are being explored. These
strategies include therapeutic vaccines, mono-
clonal antibodies, and immune checkpoint inhib-
itors (Fig. 2).
It is likely that a latency-reversing agent will need
to be coupled with an immunologic strategy to
clear virus-producing cells. One approach is to
expand HIV-1–specific CD8+ T lymphocyte re-
sponses (44). As has been well-established in
natural control of HIV-1 (i.e., “elite” control), potent
CD8+T lymphocytes can control HIV-1 by selec-
tively killing virus-producing cells (45). This con-
cept has led to a renewed interest in therapeutic
vaccines, namely the administration of candidate
HIV-1 vaccines to HIV-1–infected individuals typ-
ically on suppressive ART, with the goal of aug-
menting virus-specific immune responses and either
11 JULY 2014 • VOL 345 ISSUE 6193
accelerating the decay of the reservoir during ART
or improving the control of viral rebound after
interruption of ART. Several therapeutic vaccines
have been evaluated in HIV-1–infected subjects,
including vector-based vaccines that express HIV-1
antigens from harmless or attenuated viruses such
as canary pox (ALVAC) or adenovirus serotype 5
(Ad5), as well as plasmid DNA vaccines (46–51).
Although these vaccines generally proved immu-
nogenic, most therapeutic vaccines have to date
failed to show virologic or clinical benefit, and a
few have suggested harm, potentially as a result of
increasing immune activation. Infusion of antigen-
pulsed autologous dendritic cells has shown poten-
tially greater promise but requires confirmation (52).
Most of these early-generation therapeutic
vaccines involved relatively weak or suboptimal
immunogens. Moreover, as was demonstrated
in a study of DNA priming followed by Ad5 boost-
ing, these therapeutic vaccines typically expand
preexisting T cell clones, which by definition had
already failed to control virus replication before
ART (53), either because they were dysfunctional
or because they selected for a virus population
containing CD8+T cell escape mutations. A suc-
cessful therapeutic vaccine would ideally induce
functional CD8+T cells specific for novel HIV-1
epitopes. The next generation of therapeutic vac-
cines will also likely be combined with reservoir-
Several novel therapeutic vaccines may be
evaluated in clinical trials over the next several
prime, modified vaccinia Ankara (MVA) boost
regimens; and lymph node–targeted amphiphilic
peptide vaccines. Prophylactic vaccination with
CMV vectors has been shown to result in the
induction of broad cellular immune responses
against novel epitopes and apparent clearance
of a stringent challenge with simian immunode-
ficiency virus (SIV) in about half of vaccinated
rhesus monkeys (54–56). The persistently repli-
cative nature of CMV vectors is likely critical for
the highly functional effector memory T cells that
appear to persist indefinitely. Moreover, the abil-
II–restricted CD8+T cell responses provides a
potentially important benefit because they target
ject to immune selection pressure (56). Clinical
development of CMV vectors will presumably
require attenuated vectors, and manufacturing
and regulatory issues for CMV vectors remain to
Another therapeutic vaccine platform that will
likely be evaluated in clinical trials involves Ad26
vectors. Prophylactic vaccination with Ad26-based
prime-boost vaccine regimens, such as priming
with Ad26 and boosting with the pox virus MVA,
has proven highly immunogenic and has afforded
as well as reduced setpoint viral loads afterstrin-
gent SIV challenges in rhesus monkeys (57, 58).
Ad26 has several advantages over Ad5 as a vac-
cine vector, including the induction of different
innate immune profiles that may reduce unde-
11 JULY 2014 • VOL 345 ISSUE 6193
Immune modulating drugs
CD8+ T cells
Type 1 IFN
Fig. 2. Immunotherapeutic strategies. Tcell proliferation and activation have complex effects on the
size and distribution of theviral reservoir. Homeostatic cytokines and antigen-driven proliferation of Tcells
maintain a population of latently infected cells. Activation of Tcells can either lead to virus production and
virus spread (reservoir maintenance) ordeath of these cells because of virus-induced cytopathic effect or
host clearance mechanisms (reservoir depletion). Therapies that alter the proliferation, maturation,
activation, and effector functions of T cells could impact these processes. Immunotherapies, such as
neutralizing antibodies and therapeutic vaccines, as well as immune modulating drugs, such as PD-1
inhibitors, sirolimus, type I interferon (IFN), and IL-7, may lead to reservoir depletion.
functional T cell phenotypes. Moreover, Ad26-
based prime-boost regimens have demonstrated
partial protective efficacy in the stringent rhesus
monkey challenge models in which Ad5-based
regimenshavefailed(57–59),suggesting their po-
tential for greater clinical utility.
Use of amphiphilic peptides, which have been
shown to target antigen to lymph nodes and aim
to induce cellular immune responses simulta-
neously to two or more coevolving regions of
immunologic vulnerability of HIV-1, is another
potential therapeutic vaccine strategy (60, 61).
The extent to which these and other next-
generation vaccine candidates will afford ther-
apeutic efficacy in preclinical and clinical studies
remains to be determined.
Potent broadly neutralizing HIV-1–specific mono-
clonal antibodies (mAbs) are also being explored
as a potential HIV-1 eradication strategy. Prior
studies using the earlier generation of neutral-
izing mAbs were generally unsuccessful in both
preclinical and clinical studies (62–64), but
the identification of a newgenerationof mAbs
with improved potency and breadth has led to
a resurgence of interest in this approach. In par-
ticular, two studies in rhesus monkeys demon-
strated that infusion of mAb cocktails, as well as
certain individual mAbs, in chronically SHIV-
infected rhesus monkeys resulted in substantial,
albeit transient, suppression of viremia (65, 66),
building on previous studies in humanized mice
(67). Moreover, a subset of animals with the lowest
starting viral loads did not exhibit viral rebound
even aftermAbtiters declined toundetectablelevels,
suggesting the potential for a durable therapeutic
effect in certain circumstances (65). Infusion of
one particular mAb, PGT121, resulted in not only
rapid and profound suppression of plasma viral
RNA but also substantial reductions of proviral
DNA in peripheral blood, lymph nodes, and gas-
trointestinal mucosa (65). These data suggest that
certain mAbs may be able to target virus-infected
cells in tissues, although it remains to be deter-
mined whether these mAbs can impact the viral
A growing number of potent and broadly neu-
tralizing mAbs now exist that target various epi-
topes on the HIV-1 envelope protein. However,
a central question for HIV-1 eradication strat-
egies is whether mAbs will be able to target the
viral reservoir and clear virally infected cells,
potentially via Fc-mediated mechanisms such
as antibody-dependent cellular cytotoxicity or
antibody-dependent cell-mediated virus inhibi-
tion. Another possibility is that direct neutrali-
zation of free virions will reduce chronic antigen
stimulation and augment the functionality of
host virus-specific T cell responses, resulting in
improved virologic control. Supporting this lat-
ter possibility is the observation that, after PGT121
infusion in SHIV-infected rhesus monkeys, HIV-1
Gag–specific T cells exhibited decreased activation
and exhaustion and improved capacity to sup-
press virus replication in vitro (65). A potential
caveat regarding broadly neutralizing mAbs is
their limited accessibility to certain anatomic re-
servoir sites, such as the central nervous system.
Immune checkpoint blockade
Another potential strategy to augment host im-
mune control involves the blockade of immune
checkpoint molecules. Activated T cells express a
series of receptors that when engaged by their
ligands result in suppression of function and return
to a nonactivated state. The best-characterized
immune checkpoint molecule is PD-1, which is a
marker of functional T cell exhaustion. The ligands
for PD-1, PDL-1 and PDL-2, are widely expressed
in tissues. Inhibitors of the PD-1 pathway restore
T cell function and have shown efficacy in the
cancer field (68). These inhibitors may also en-
hance the capacity of the host immune system
to clear chronic viral infections. Indeed, a single
dose of an antibody against PD-1 appeared to
cure hepatitis C virus infection in a small subset
of individuals (69). It is thought that inhibitors
of other immune checkpoint molecules (such as
CTLA-4, LAG-3, TIM-3, TIGIT, and 2B4) may also
PD-1 inhibitors and other checkpoint blockers
may also have direct effects on the establishment
and maintenance of the viral reservoir (11). Be-
cause activated cells are more likely to become
infected than resting cells, previously activated
cells—including those expressing PD-1—might be
expected to be enriched for HIV-1 (11). Indeed,
the size of the reservoir is positively correlated
and HIV-1 is enriched in PD-1–expressing mem-
ory cells (11). Clinical trials of PD-1 inhibitors in
treated HIV-1 disease have been initiated.
After years of suppressive ART, the vast majority
of the residual population of replication-competent
virus resides in long-lived memory CD4+T cells.
The estimated size of the reservoir during ART
is directly associated with the frequency of ac-
tivated and proliferating cells (11, 12, 22, 23, 25),
and in some studies activated and proliferating
cells also contain more HIV-1 (11, 70). Therapeu-
tic interventions that target T cell activation or
proliferation may therefore impact the viral res-
ervoir (Fig. 2).
T cell activation and proliferation are controlled
by a number of signaling pathways, including those
involving the mammalian target of rapamycin
(mTOR), signal transducer and activator of tran-
scription 5a (STAT5a), and forkhead box O3a
(71). Specific inhibitors of these pathways might
therefore reduce the size of the latent viral reser-
voir. For example, sirolimus (rapamycin) is a nat-
urally occurring macrolide that inhibits mTOR,
and as a consequence it blocks cell cycle progres-
sion from G1to S phase in activated T cells and
reduces CCR5 expression (72). Sirolimus also en-
hances memory T cell formation in response to
vaccines in preclinical studies (73). We performed
a retrospective analysis of ART-treated HIV-1–
infected kidney transplant recipients who received
sirolimus or non–mTOR-targeting drugs to pre-
vent graft rejection. Only treatment with sirolimus
was temporally associated with lower levels of cell-
associated HIV-1 DNA (74), suggesting that mTOR
inhibition may prove useful in a combination
eradication strategy. A prospective clinical trial is
being developed to evaluate this concept further.
Other drugs that limit T cell activation, includ-
ing those that inhibit the Janus kinase/STAT
pathway (75), are also being developed. Because
chronic inflammation may contribute to excess
morbidity in ART-suppressed, HIV-1–infected in-
dividuals (9), a number of therapies that target
the inflammatory pathways are being developed
as adjuncts to therapy (e.g., methotrexate, anti-
fibrotic drugs, anti-CMV therapy, and a number of
agents aimed at gut mucosa and microbial trans-
location). Measures of the viral reservoir are in-
creasingly being incorporated into these studies.
The role of the type I interferon family of cyto-
kines in the context of chronic viral infections is
currently a focus of intense investigation and
debate. An acute viral infection triggers imme-
diate and potent type I interferon production by
a variety of cells, leading to up-regulation of hun-
dreds of genes (the interferon-stimulated genes)
(76). In the context of chronic HIV-1 infection, per-
sistent interferon signaling has been associated
with increased CCR5 and PD-1 expression, up-
regulation of indoleamine 2,3-dioxygenase and
other inflammatory pathways, reduced thymo-
poieses, the generation of dysfunctional T cells,
and altered T cell homeostasis (77–80). Higher
levels of interferon signaling are also correlated
with poor immune reconstitution (80, 81). In
murine models of chronic lymphocytic chorio-
meningitis virus (LCMV) infection, inhibition of
type I interferon leads to decreased immune
activation, decreased PD-1 and PD-L1 activity,
decreased amounts of the immunomodulatory
cytokine IL-10, restored lymphoid architecture,
improved CD4+T cell responses, and ultimately
enhanced viral clearance (82, 83). Although chronic
type I interferon stimulation might be harmful,
supratherapeutic doses of type I interferon have
also been shown to have sustained anti–HIV-1
effects in untreated chronic infection (84) and
may affect reservoir size in treated disease (85).
Proof-of-concept studies involving interferon-a
and inhibitors of interferon-a are both planned.
As a result of the correlation between markers
of T cell activation or dysfunction and the size
of the viral reservoir, it is possible that immuno-
suppressive interventions may also contribute
to a cure. It should be noted, however, that the
most direct way to “shock” HIV-1 out of a state
of latency may be nonspecific and potent ac-
tivation of infected CD4+T cells (40, 86–87). The
first generation of HIV-1 eradication studies per-
formed over a decade ago involved the use of
antibodies against CD3 (88), which trigger T cell
activation, but the inflammatory response proved
too toxic to pursue. An ideal immune modifying
regimen would reverse HIV-1 latency in a specific
manner while preventing the negative conse-
quences of T cell activation and T cell proliferation.
Given the inherent complexity of the immune
response and the fact that any immune-modifying
11 JULY 2014 • VOL 345 ISSUE 6193
therapeutic agent will invariably lead to com-
plex and difficult-to-predict counterregulatory
responses, it is difficult to predict how such inter-
ventions will affect the viral reservoir. The com-
plexity of attempting to use an activating agent
to reverse latency is illustrated by recent studies
with IL-7. Preclinical studies suggested that this
approach would reverse latency (89). Although a
proof-of-concept clinical study found that IL-7
caused production of virions in vivo (90), the
overall impact of the approach was an increase
in the estimated reservoir size, presumably be-
cause of proliferation and expansion of infected
cells (12). Experiments in improved preclinical
models of ART-suppressed virus infection and
proof-of-concept clinical trials will hopefully pro-
vide clarity on these and other issues.
The PrEP, PEP, and cure continuum
Treating HIV-1 with ART shortly after infection
may also contribute to HIV-1 eradication. ART
during acute HIV-1 infection reduces the size of
the reservoir (91, 92), limits the generation of es-
cape mutants, and preserves immune function.
For these reasons, it is thought that HIV-1–infected
individuals who initiate ART during acute infec-
tion have the best chance for HIV-1 eradication,
although this group represents only a small frac-
tion of total HIV-1–infected individuals. It is also
possible that very early ART may even be curative,
as illustrated by the case of the HIV-1–infected
infant who was started on ART at 30 hours of
life. Therapy was discontinued after 18 months,
and virus remained undetectable (through at
least month 30), suggesting that very early ini-
tiation of ART may prevent establishment of a
long-lived reservoir (28). However, the applica-
bility of these findings to sexual HIV-1 transmis-
sion in adults remains uncertain.
The potentially curative role of ART when
administered within days of infection blurs the
traditional distinction between preexposure pro-
phylaxis (PrEP), postexposure prophylaxis (PEP),
and a cure (Fig. 3). It is well established that treat-
ing adults with ART within hours of HIV-1 ex-
posure (e.g., a needlestick injury in a healthcare
worker) substantially reduces the risk of acquir-
ing HIV-1. However, antiretroviral drugs inhibit
active virus replication in host cells, and thus the
clinical success of PEP strategies indicates that
the first HIV-1–infected cells after viral exposure
can be eradicated. Indeed, if intensive virologic
monitoring were used in these individuals, then
it is possible that some might exhibit transient
low levels of virus in blood or tissues and thus
would be considered “cured.” From this per-
spective, the use of very early ART to eradicate
HIV-1 infection may not be uncommon.
Early ART that fails to block the establish-
ment of the viral reservoir might still prevent
some of the immunologic damage that typically
occurs during acute HIV-1 infection, thus aug-
menting the capacity of the host immune system
to control viral replication. Such an ART-induced
shift toward a more effective host immune re-
sponse and a smaller viral reservoir may in rare
cases lead to sustained control of the virus (a
functional cure). Among a cohort of adults in
France who started ART during acute HIV-1 in-
fection and then discontinued therapy after sev-
eral years (the VISCONTI cohort), about 10 to
15% did not exhibit detectable viral rebound,
although replication-competent virus still per-
sisted in these individuals and they lacked pro-
tective HLA haplotypes (93). This observation
has not yet been confirmed, and the mechanism
for the posttreatment controller phenotype re-
mains to be determined. As compared with elite
without ART, the posttreatment controllers dem-
onstrate remarkably small viral reservoirs and low
levels of T cell activation, both of which may have
contributed to the lack of rebound viremia when
therapy was discontinued.
Conclusions and perspectives
Advances over the past several years have sug-
under specific conditions. Thus, the increasing
enthusiasm for HIV-1 eradication research and
the growing public and private investment in
the HIV-1 cure agenda are justifiable. However,
as described in detail elsewhere (94, 95), major
scientific challenges remain. A more detailed
understanding of the biology of the latent viral
reservoir and the partially effective virus-specific
immune responses is critical. As the Boston pa-
tients demonstrate, improved assays are needed
to quantify the viral reservoir, and predictive
biomarkers for viral rebound are also needed.
Reliable and predictive animal models should
also be developed to evaluate concepts and to
inform clinical research strategies.
The establishment and maintenance of the
viral reservoir appears to be affected at least in
part by the immune system. Immunotherapy
approaches will therefore likely have an increas-
ing role in HIV-1 eradication strategies in the
future. However, the optimal strategies to stim-
ulate viral release from latency, to augment host
immune responses, and to limit negative inflam-
matory responses remain to be determined. Al-
though a small number of case reports suggest
that it might be possible to eradicate HIV-1 in-
fection in unusual circumstances, no proof of con-
cept yet exists that chronic HIV-1 infection can
be cured by a safe and scalable intervention.
Over the next few years, multiple novel and
promising HIV-1 eradication concepts will be
evaluated. It is likely that progress will be steady
but unpredictable, and reaching the final goal
may take many years. Regardless, substantial
expansion of rigorous basic research, preclinical
studies, and clinical trials in this field will un-
doubtedly lead to important advances in our
understanding of the biology of the HIV-1 res-
ervoir and the challenges that face HIV-1 eradi-
REFERENCES AND NOTES
1. D. Finzi et al., Science 278, 1295–1300 (1997).
2. D. Persaud, Y. Zhou, J. M. Siliciano, R. F. Siliciano, J. Virol. 77,
3. T. W. Chun et al., Proc. Natl. Acad. Sci. U.S.A. 94, 13193–13197
4. T. W. Chun et al., Proc. Natl. Acad. Sci. U.S.A. 95, 8869–8873
5. T. W. Chun et al., Nature 387, 183–188 (1997).
6. T. W. Chun, R. T. Davey Jr., D. Engel, H. C. Lane, A. S. Fauci,
Nature 401, 874–875 (1999).
7. D. Finzi et al., Nat. Med. 5, 512–517 (1999).
8. Y. C. Ho et al., Cell 155, 540–551 (2013).
9. S. G. Deeks, R. Tracy, D. C. Douek, Immunity 39, 633–645
10. G. Doitsh et al., Nature 505, 509–514 (2013).
11. N. Chomont et al., Nat. Med. 15, 893–900 (2009).
12. C. Vandergeeten et al., Blood 121, 4321–4329 (2013).
11 JULY 2014 • VOL 345 ISSUE 6193
Fig. 3. PrEP/PEP/cure continuum. ART initiated before exposure is termed preexposure prophylaxis (PrEP), whereas ART initiated shortly after exposure is
postexposure prophylaxis (PEP) and forms a continuum with efforts aimed at virus eradication (cure). Even if early ART is not curative, it may reduce the size of
the viral reservoir and preserve immune function.
13. M. J. Buzón et al., Nat. Med. 16, 460–465 (2010).
14. J. Cohen, Science 343, 1188 (2014).
15. L. Josefsson et al., Proc. Natl. Acad. Sci. U.S.A. 110,
16. C. V. Fletcher et al., Proc. Natl. Acad. Sci. U.S.A. 111,
17. M. J. Buzon et al., Nat. Med. 20, 139–142 (2014).
18. S. A. Yukl et al., J. Infect. Dis. 208, 1212–1220 (2013).
19. T. W. North et al., J. Virol. 84, 2913–2922 (2010).
20. A. Sigal et al., Nature 477, 95–98 (2011).
21. V. A. Evans et al., PLOS Pathog. 9, e1003799 (2013).
22. H. Hatano et al., J. Infect. Dis. 208, 50–56 (2012).
23. J. M. Murray et al., J. Virol. 88, 3516–3526 (2014).
24. H. Hatano et al., J. Infect. Dis. 208, 1436–1442 (2013).
25. N. R. Klatt, N. Chomont, D. C. Douek, S. G. Deeks,
Immunol. Rev. 254, 326–342 (2013).
26. M. F. Chevalier, L. Weiss, Blood 121, 29–37 (2013).
27. M. Zeng et al., PLOS Pathog. 8, e1002437 (2012).
28. D. Persaud et al., N. Engl. J. Med. 369, 1828–1835 (2013).
29. K. Luzuriaga et al., J. Infect. Dis. (2014).
30. J. E. Mold et al., Science 330, 1695–1699 (2010).
31. R. F. Siliciano, Nat. Med. 20, 480–481 (2014).
32. G. Hütter et al., N. Engl. J. Med. 360, 692–698 (2009).
33. S. A. Yukl et al., PLOS Pathog. 9, e1003347 (2013).
34. P. Tebas et al., N. Engl. J. Med. 370, 901–910 (2014).
35. N. Holt et al., Nat. Biotechnol. 28, 839–847 (2010).
36. P. M. Younan et al., Blood 122, 179–187 (2013).
37. T. J. Henrich et al., J. Infect. Dis. 207, 1694–1702 (2013).
38. P. M. Cannon, D. B. Kohn, H. P. Kiem, Nat. Biotechnol. 32,
39. N. M. Archin et al., Nature 487, 482–485 (2012).
40. C. K. Bullen, G. M. Laird, C. M. Durand, J. D. Siliciano,
R. F. Siliciano, Nat. Med. 20, 425–429 (2014).
41. J. Blazkova et al., J. Infect. Dis. 206, 765–769 (2012).
42. A. M. Spivak et al., Clin. Infect. Dis. 58, 883–890 (2014).
43. C. Katlama et al., Lancet 381, 2109–2117 (2013).
44. L. Shan et al., Immunity 36, 491–501 (2012).
45. S. A. Migueles et al., Immunity 29, 1009–1021 (2008).
46. E. S. Rosenberg et al., PLOS ONE 5, e10555 (2010).
47. R. T. Schooley et al., J. Infect. Dis. 202, 705–716 (2010).
48. J. B. Angel et al., AIDS 25, 731–739 (2011).
49. R. T. Gandhi et al., Vaccine 27, 6088–6094 (2009).
50. B. Autran et al., AIDS 22, 1313–1322 (2008).
51. J. M. Kilby et al., J. Infect. Dis. 194, 1672–1676 (2006).
52. F. García et al., Sci. Transl. Med. 5, 166ra2 (2013).
53. J. P. Casazza et al., J. Infect. Dis. 207, 1829–1840 (2013).
54. S. G. Hansen et al., Nature 473, 523–527 (2011).
55. S. G. Hansen et al., Nature 502, 100–104 (2013).
56. S. G. Hansen et al., Science 340, 1237874 (2013).
57. D. H. Barouch et al., Cell 155, 531–539 (2013).
58. D. H. Barouch et al., Nature 482, 89–93 (2012).
59. N. L. Letvin et al., Sci. Transl. Med. 3, 81ra36 (2011).
60. H. Liu et al., Nature 507, 519–522 (2014).
61. V. Dahirel et al., Proc. Natl. Acad. Sci. U.S.A. 108, 11530–11535
62. P. Poignard et al., Immunity 10, 431–438 (1999).
63. A. Trkola et al., Nat. Med. 11, 615–622 (2005).
64. S. Mehandru et al., J. Virol. 81, 11016–11031 (2007).
65. D. H. Barouch et al., Nature 503, 224–228 (2013).
66. M. Shingai et al., Nature 503, 277–280 (2013).
67. F. Klein et al., Nature 492, 118–122 (2012).
68. S. L. Topalian et al., N. Engl. J. Med. 366, 2443–2454
69. D. Gardiner et al., PLOS ONE 8, e63818 (2013).
70. T. W. Chun et al., J. Clin. Invest. 115, 3250–3255 (2005).
71. C. Riou et al., J. Exp. Med. 204, 79–91 (2007).
72. A. Heredia et al., Proc. Natl. Acad. Sci. U.S.A. 100, 10411–10416
73. K. Araki et al., Nature 460, 108–112 (2009).
74. P. G. Stock et al., Am. J. Transplant. 14, 1136–1141
75. C. Gavegnano et al., Antimicrob. Agents Chemother. 58,
76. M. J. de Veer et al., J. Leukoc. Biol. 69, 912–920 (2001).
77. C. A. Stoddart, M. E. Keir, J. M. McCune, PLOS Pathog. 6,
78. D. Favre et al., Blood 117, 2189–2199 (2011).
79. D. Favre et al., Sci. Transl. Med. 2, 32ra36 (2010).
80. C. Le Saout et al., PLOS Pathog. 10, e1003976 (2014).
81. S. Fernandez et al., J. Infect. Dis. 204, 1927–1935 (2011).
82. E. B. Wilson et al., Science 340, 202–207 (2013).
83. J. R. Teijaro et al., Science 340, 207–211 (2013).
84. S. K. Pillai et al., Proc. Natl. Acad. Sci. U.S.A. 109, 3035–3040
85. L. Azzoni et al., J. Infect. Dis. 207, 213–222 (2013).
86. A. Bosque, M. Famiglietti, A. S. Weyrich, C. Goulston,
V. Planelles, PLOS Pathog. 7, e1002288 (2011).
87. A. R. Cillo et al., Proc. Natl. Acad. Sci. U.S.A. 111, 7078–7083
88. J. Kulkosky et al., J. Infect. Dis. 186, 1403–1411 (2002).
89. F. X. Wang et al., J. Clin. Invest. 115, 128–137 (2005).
90. I. Sereti et al., Blood 113, 6304–6314 (2009).
91. N. M. Archin et al., Proc. Natl. Acad. Sci. U.S.A. 109,
92. J. Ananworanich et al., PLOS ONE 7, e33948 (2012).
93. A. Sáez-Cirión et al., PLOS Pathog. 9, e1003211 (2013).
94. D. D. Richman et al., Science 323, 1304–1307 (2009).
95. S. G. Deeks et al., Nat. Rev. Immunol. 12, 607–614 (2012).
The authors thank B. Walker, N. Michael, J. Kim, M. Robb, R. Geleziunas,
M. McCune, R. Sekaly, N. Chomont, and S. Lewin for helpful discussions.
The authors acknowledge support from the NIH [AI078526, AI084794,
AI095985, AI096040, AI096109 (Delaney AIDS Research Enterprise),
AI100663, and OD011170]; the Bill and Melinda Gates Foundation
(OPP1033091, OPP1040741, and OPP1083689); the Ragon Institute
of MGH, MIT, and Harvard; the Henry M. Jackson Foundation; and the
American Foundation for AIDS Research. D.H.B. is a co-inventor on
Ad26 vector, mosaic antigen, and Env immunogen patents. S.G.D.
has served as an ad hoc consultant for Bristol-Myers Squib,
GlaxoSmithKline, and Janssen on HIV cure–related activities.
11 JULY 2014 • VOL 345 ISSUE 6193
DOI: 10.1126/science.1255512 Download full-text
, 169 (2014);
Dan H. Barouch and Steven G. Deeks
Immunologic strategies for HIV-1 remission and eradication
This copy is for your personal, non-commercial use only.
clicking here. colleagues, clients, or customers by
, you can order high-quality copies for your
If you wish to distribute this article to others
The following resources related to this article are available online at
here. following the guidelines
can be obtained by
Permission to republish or repurpose articles or portions of articles
Updated information and services,
): September 13, 2015 www.sciencemag.org (this information is current as of
version of this article at:
including high-resolution figures, can be found in the online
related to this article
A list of selected additional articles on the Science Web sites
, 44 of which can be accessed free:
cites 95 articles
5 articles hosted by HighWire Press; see:
This article has been
This article appears in the following
registered trademark of AAAS.
is aScience 2014 by the American Association for the Advancement of Science; all rights reserved. The title
CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005.
(print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the Science
on September 13, 2015