hiv pathogenesis: dynamics and genetics of viral...

HIV Pathogenesis: Dynamics and Geneticsof Viral Populations and Infected Cells

John Coffin1 and Ronald Swanstrom2

1Department of Molecular Biology and Microbiology, Tufts University, Boston, Massachusetts 021112Department of Biochemistry and Biophysics, University of North Carolina, Chapel Hill, North Carolina 27599

Correspondence: [email protected]

In the absence of treatment, HIV-1 infection, usually starting with a single virion, leads inex-orably to a catastrophic decline in the numbers of CD4þ T cells and to AIDS, characterized bynumerous opportunistic infections as well as other symptoms, including dementia andwasting. In the 30 years since the AIDS pandemic came to our attention, we have learneda remarkable amount about HIV-1, the responsible virus—the molecular details abouthow it functions and interacts with the host cell, its evolution within the host, and the counter-measures it has evolved to overcome host defenses against viral infection. Despite theseadvances, we remain remarkably ignorant about how HIV-1 infection leads to disease andthe death of the human host. In this brief article, we introduce and discuss importantlessons that we have learned by examining the dynamics of viral populations and infectedcells. These studies have revealed important features of the virus–host interaction that nowform the basis of our understanding of the importance and consequence of ongoing viral rep-lication during HIV-1 infection.

TIME COURSE OF INFECTION

As discussed elsewhere (Shaw and Hunter2011), HIV-1 infection is usually initiated

with a single virion infecting a single targetcell at the portal of entry. The subsequent courseof infection can be monitored in several ways:overt symptoms, such as fever, wasting, op-portunistic infections, neurological symptoms,and so on; blood levels of the CD4þ T-cell tar-get and antiviral antibodies; and viremia (virusin blood), measured by infectivity, immunoas-say for viral proteins, and, most accurately, byPCR for viral RNA. A typical time course of in-fection relating these properties to one anotheris shown in Figure 1. Although their timing can

vary considerably from individual to individ-ual, as can the levels of viremia, the general out-line is essentially the same in virtually everyinfected person who does not receive effectiveantiviral therapy.

1. (�1–2 wk) An eclipse phase, during whichthe virus is freely replicating and spreadingfrom the initial site of infection to themany tissues and organs that provide thesites for replication. Viremia is undetectable,and neither immune response nor symp-toms of infection are yet visible.

2. (�2–4 wk) The acute (or primary) infectionphase, characterized by relatively high levels

Editors: Frederic D. Bushman, Gary J. Nabel, and Ronald Swanstrom

Additional Perspectives on HIV available at www.perspectivesinmedicine.org

Copyright # 2013 Cold Spring Harbor Laboratory Press; all rights reserved; doi: 10.1101/cshperspect.a012526

Cite this article as Cold Spring Harb Perspect Med 2013;3:a012526

1

ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

Press on November 9, 2020 - Published by Cold Spring Harbor Laboratoryhttp://perspectivesinmedicine.cshlp.org/Downloaded from

http://perspectivesinmedicine.cshlp.org/

of viremia (up to 107 or more copies of viralRNA per milliliter of blood), and large frac-tions of infected CD4þ T cells in blood andlymph nodes. This phase is often, but notalways, accompanied by “flu-like” symp-toms—fever, enlarged lymph nodes, andthe like. Around the time of peak viremia,the immune response begins to appear,both in the form of antibodies against allviral proteins, and a CD8þ T-cell responseagainst HIV-1 antigens expressed on infectedcells. The high levels of viremia that charac-terize this phase most likely result from theabsence of the early immune response andthe generation, as part of the host response,of large numbers of activated CD4þ T cells,providing a wealth of targets for viral replica-tion. At the end of the acute phase, the levelof viremia declines sharply, 100-fold ormore, a result of both partial control by theimmune system and exhaustion of activatedtarget cells. This phase is also characterizedby a transient decline in the numbers ofCD4þ T cells in the blood.

3. (�1–20 yr) Chronic infection, or “clinicallatency,” is characterized by a constant orslowly increasing level of viremia, usuallyon the order of 1–100,000 copies/mL,sometimes referred to as the “set point,”and steady, near normal (�1000 cells/mL),or gradually falling levels of CD4þ T cells.As a rule, patients in this phase are asympto-matic and usually unaware that they havebeen infected. Despite the term “latency,”the viral infection is far from latent, withlarge numbers of CD4þ T cells becominginfected and dying every day.

4. Finally, the number of CD4þ T cells declinesto the point (�200 cells/mL) at whichimmune control of adventitious infectiousagents can no longer be maintained, andopportunistic infections (discussed in Lack-ner et al. 2011) begin to appear. Control ofthe HIV-1 infection itself is also lost, andthe level of viremia rises during the AIDSphase, culminating in death of the infectedpatient. Indeed, untreated HIV-1 infection

12108

Years

Time after infection

Weeks

6421296300

200

400CD

4+ T

cel

ls (

cells

/μL)

600

800

1000

1200

Primaryinfection Death

Appearance ofanti-HIV CTLs

Clinical latency

Opportunisticinfections

Acute HIV infectionWide distribution of virusSeeding of lymphoid organs

102

103

104

105

106

HIV

RN

A (

copi

es/m

L pl

asm

a)

107

Onset ofsymptoms

Figure 1. Time course of typical HIV infection. Patterns of CD4þT cell decline and viremia vary greatly from onepatient to another. (From Fauci and Desrosiers 1997; reprinted, with permission, from Cold Spring Harbor Lab-oratory Press # 1997.)

J. Coffin and R. Swanstrom

2 Cite this article as Cold Spring Harb Perspect Med 2013;3:a012526

ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



is one of the most uniformly lethal infectiousdiseases known, with a mortality rate wellover 95%.

SIGNIFICANCE OF VIREMIA

As clinical researchers were developing tools fordiagnosis and prognosis of HIV-1 infection, itbecame apparent that the most powerful ofthem is the measurement of viremia usingquantitative nucleic acid hybridization or PCRassays for HIV-1 genomic RNA in blood. RNAmeasurements have the virtue of an extraordi-narily wide dynamic range. Routine commer-cial assays can now accurately measure levelsof viremia as low as 50 copies/mL and as highas 107 or more copies/mL. More sensitive re-search assays have reduced the lower end tomuch less than 1 copy/mL with good quantita-tive accuracy (Palmer et al. 2003).

The first recognition of the prognosticimportance of viremia came from analysis of alarge prospective study of gay men at risk forAIDS, the Multicenter AIDS Cohort Study(MACS), in which volunteers provided bloodsamples and health information at frequentintervals over long periods of time, duringwhich many men who were initially infected,but still healthy, progressed to AIDS and death(Mellors et al. 1996). When the levels of viremiaat study entry were compared with outcome, astrong inverse correlation between the level of

viremia and the time of progression to AIDSwas observed (Fig. 2). The data showed thatthe group with the highest level of viremia pro-gressed to AIDS on average about fourfold fasterthan the lowest group. Thus, the level of viruspresent in the blood provides a measure ofthe rate at which viral infection damages theimmune system of the infected host.

The second important insight made possi-ble by the ability to accurately measure levelsof viremia was in analyzing the consequencesof antiviral therapy, where it was found thatthe ability of an antiretroviral drug to halt oreven reverse progression of an infected patientto AIDS was well correlated with its ability tosuppress viremia. Effective therapies, such ascombinations of antiviral drugs discussed byArts and Hazuda (2011), which can lead to sus-tained reductions in viremia to undetectablelevels of ,50 copies/mL by standard clinicalassays, can forestall progression to AIDS indef-initely, although they do not cure the infection.Indeed, if therapy is suspended, even after manyyears, the level of viremia will rapidly return tothe prior set point.

The correlation between viremia and diseasehas a straightforward underlying explanation(Coffin 1995). Infected cells produce virions,some fraction of which are released into theblood, from which they are rapidly removed ordegraded. Although different cells may produceand release different amounts of virus into the

00

0.2

0.4

0.6

0.8

1

CD4 count (cells/μL)B

2 4 6 8 1000

0.2

0.4

0.6

0.8

1

Viral load (RNA copies/mL)A

2 4 6

>36000

<300

4500–13000

<4500

13000–36000

300–500

500–800

>800

8 10

Years Years

Pro

port

ion

AID

S-f

ree

surv

ivor

s

Figure 2. Relationship between viral load (viremia) and clinical progression. Shown are Kaplan–Meier plots ofAIDS-free survival divided into quartiles according to virus load (A) or CD4 count (B) at the time of diagnosis.(From Fauci and Desrosiers 1997; reprinted, with permission, from Cold Spring Harbor Laboratory Press# 1997.)

HIV Pathogenesis: Viral Populations and Infected Cells

Cite this article as Cold Spring Harb Perspect Med 2013;3:a012526 3

ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



blood, their numbers are very large, and, onaverage, the amount of virus per infected cellwill be about the same from time to time inthe same patient and similar from one patientto another. Also, the rate of removal of virionsfrom the blood will be similar (and, as hasbeen measured, quite rapid) from time totime and patient to patient. Because the levelof viremia is determined by the balance betweenvirion release into and removal from the blood-stream, assuming a constant rate of release percell and a constant rate of removal, the level ofviremia will be an instantaneous relative meas-ure of the number of productively infected cellsat any one time.

INSIGHTS FROM THERAPEUTICDRUG STUDIES

This understanding of the significance of vire-mia in HIV-1 infection, combined with thenew availability of potent antiretroviral drugs

(Arts and Hazuda 2011), led to an experimentthat provided considerable insight into theprocess of HIV-1 infection and pathogenesisin vivo (Fig. 3) (Coffin 1995; Ho et al. 1995;Wei et al. 1995). The study was based on theprinciple that all antiviral drugs affect the abil-ity of the virus to infect cells and do not affectthe lifetime of infected cells or their ability toproduce viral particles (albeit noninfectiousones in the case of protease inhibitors). There-fore, any decrease in viremia following initiationof suppressive therapy must reflect a combina-tion of the length of time that productivelyinfected cells live and produce virus after infec-tion and the lifetime of the virus in the blood.Because the latter factor is very small, it can beignored with little effect on the overall result.Thus, viremia was assayed by quantitative PCRat closely spaced time points after initiation ofsuppressive therapy, with striking results: Aftera brief lag, viremia drops rapidly, to �1%–10% of the set point level with a half-time of

108

Weeks of treatment

64

Macrophages??Virus load

T1/2 ~ 20 d

T1/2 ~ 6 mo

Quiescent infected CD4 cells

T1/2 ~ 1 dActivated infected CD4 cells

T1/2 ~ 100 d Viral DNA positive cells

2

Initiate suppressive antiviral therapy

0

–3

–2

–1

Log

rela

tive

valu

es

0

Figure 3. Decay of circulating virus and infected cells after initiation of suppressive antiviral therapy. Theseresults imply the presence of at least four classes of HIV-infected cells: productively infected, a second class, per-haps macrophages, with a half-life of �2 wk; latently infected resting CD4 cells, with a half-life of �6 mo; andlong-lived DNA positive cells, most of which are nonproductive. A fourth class of infected cells, inferred to havean approximately infinite half-life, is not shown. (Figure modified from Fauci and Desrosiers 1997.)



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



�1–2 d. Following this first phase of decay,there is a second phase in which most of theremaining viremia decays with a half-time of�2 wk or so, down to the detection limit ofstandard assays. Two more phases, which havebeen defined with more sensitive assays, are dis-cussed below.

The first phase of decay almost certainlyreflects the lifetime of productively infectedCD4þ T cells, although the exact cells and theirlocation have not been identified. The secondphase has been speculatively attributed to mac-rophages (Perelson et al. 1997), but this point isfar from established. Given that this phase ismuch less prominent when the treatmentincludes an integrase inhibitor, it may reflectat least in part a population of infected cells inwhich integration is delayed for some reason(Murray et al. 2007).

THE HIV-1 STEADY STATE

The results discussed above, combined with thefact that the characteristic level of viremia (setpoint) in any infected individual is essentiallyconstant, led to the conclusion that HIV-1 existsin the infected host as a quasi-steady state, inwhich the level of viremia is largely determinedby the number of virus-producing cells, whichtherefore must also be constant day in and dayout throughout most of the course of infection.In any given patient, therefore, there are aboutthe same number of infected cells today as therewere yesterday and last year, but they must bedifferent cells arising from infection of freshtarget cells, which then rapidly die after pro-ducing sufficient virus, on average, to infectthe same number of cells again. This conceptis key to our current understanding of theHIV–host interaction. It has many importantimplications.

1. The replication cycle time of the virus is, onaverage, �1–2 d. Thus, virus in an individ-ual at a year after infection is 200–300 gener-ations removed from the initial infectingvirion.

2. The population size is determined by thesteady-state number of productively infected

cells, not by the number of virions, exceptthat the latter must be at least large enoughto ensure that the same number of cells isinfected every day.

3. The steady state is remarkably robust, chang-ing little throughout the clinically latentphase, and returning to the same value aftereven prolonged periods (7 yr or more) whenit is undetectable on suppressive therapy, orafter transient increases due to immunestimulation by vaccination or by infectionwith some other agent.

4. The virus exists as a replicating populationthat is carried forward by repeated infectionof large numbers of cells with no evidence ofbottlenecks or other major alterations inpopulation size from time to time.

5. The level of viremia is determined by a com-bination of the number of available targetcells, the infectivity of the virus, the numberof virions produced per cell, the productivelifetime of the infected cell, and the clearancerate of the virus.

6. We know little about the infected cells thatare responsible for producing the virusobserved in the blood and can only infertheir properties indirectly.

7. The extraordinary extent of genetic diversitythat accumulates during HIV-1 infection,relative to other RNA viruses, is not due toits unusually high error rate, but rather toreplication of a large population of viruswith an average error rate repeated over andover again, accumulating thousands of gen-erations over the course of infection of a sin-gle individual.

With these properties in mind, we can lookat some specific issues.

Nature and Lifetimes of Infected Cells

The initial therapeutic studies that led to thesteady-state concept implied two populationsof target cells: one with a half-life of �1–2 dand another of �2 wk. These conclusionsgenerated considerable optimism because the



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



proportionality of the level of viremia to thenumber of productively infected cells suggestedthat sufficiently prolonged treatment couldeventually reduce the number of infected cellsto ,1, effectively curing the patient of theHIV-1 infection. Unfortunately, these hopeswere dashed when further studies, using moresensitive techniques, revealed the presence ofat least two more phases of decay of HIV-1 vir-emia on fully suppressive therapy, apparentlydue to small populations of infected cells capa-ble of surviving and releasing virus many yearsafter infection (Siliciano and Green 2011).Although such cells constitute only a tiny frac-tion of the total number of infected cells presentbefore therapy, they decay very slowly (with ahalf-life of �6 mo by one estimate) (Zhanget al. 1999) after treatment. Also, much moresensitive PCR assays, with a limit of detectionwell below 1 copy/mL of viral RNA in plasma(Palmer et al. 2003), revealed that the majoritypatients on therapy with “undetectable” vire-mia for more that 1 yr in fact had an averageof about 3 copies of HIV-1 RNA per milliliter,or about 1/10,000 of the level before therapy,and that this level declined further with a curveconsistent with two more phases of decay: onewith a half-life of �39 wk, not significantlydifferent from that of inducible cells, perhapsfrom the same source, and another with a half-life not different from infinity, suggestingthe possibility of a dividing reservoir. Viremiawas still detectable in all patients as long as 7yr after the start of therapy (Palmer et al.2008). The existence of these populations ofcells, apparently infected before therapy andcapable of producing virus many years later,rules out eradicating infection by antiviral ther-apy alone.

One more population of infected cells bearsdiscussion. In a typical chronically infectedindividual, about one CD4þ T cell per 100 toone per 1000 contains HIV-1 DNA, largely inthe form of an integrated provirus. Very few cellscontain more than one provirus (Josefsson et al.2011). After the initiation of therapy, there is aninitial decline in the numbers of these cells,reflecting the decay in viremia and implyingthat at least some of the virus in the blood comes

from populations of cells also found in theblood. However, as can be seen in Figure 3,the level of viral DNA positive cells soon levelsoff, typically somewhere around 1%–10% ofthe starting value, implying that a significantfraction of these cells is not involved in phase1 or 2 viral production. Furthermore, the num-bers of DNA positive cells after prolongedtherapy is typically 100-fold greater than thenumber of cells inducible to produce virus exvivo (Chun et al. 1997), implying that most ofthem are incapable of ever producing infectiousvirus, perhaps because of mutations in the pro-virus, unfavorable sites of integration, or someundefined cellular factor that prevents expres-sion of the silent provirus. Simple modelingimplies that over time, because of their muchlonger lifetime, cells containing inactive orlatent proviruses should accumulate to a con-siderable extent relative to productively infectedcells, most of which die within a few days. Thus,only a small fraction of the population of cellscontaining detectable proviruses represent cellsthat are latently infected, that is, that contain aprovirus that is not expressed but is capable ofbeing induced to express infectious virus bysome means, such as stimulation to reenterthe cell cycle. For this reason, it cannot beassumed (although it often is) that the natureof cells in the so-called latent reservoir pre-sumed responsible for the persistence of HIV-1 during years of therapy can be inferred fromthe properties of all provirus-containing cells(such as sites of integration, expression of tran-scription factors, etc.), the large majority ofwhich must contain dead proviruses incapableof reincarnation. The physical identificationand manipulation of the tiny fraction of latentlyinfected cells in suppressed infection is a majorunsolved problem in HIV-1 pathogenesis (Mal-darelli 2011; Siliciano and Green 2011).

Persistence of HIV-1 Infection on Therapy

Although it is well accepted that the largemajority of patients on suppressive therapyhave low levels of viremia as well as latentlyinfected cells capable of being reactivated toproduce HIV-1 ex vivo, the link between the



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



two phenomena remains somewhat controver-sial, with some investigators arguing that thepersistent viremia observed in patients on long-term therapy could also be due to a smallamount of “escape” replication of virus in ana-tomic sites or cells that are poorly accessible toantiviral therapy (referred to as “sanctuaries”)(Sharkey and Stevenson 2001). In support ofthis claim, they point to the ongoing presenceof two-LTR circles (Craigie and Bushman2011) in PBMCs in a fraction of patients (Buzonet al. 2011), an observation interpreted asmeaning that productive infection of new cellsis continuing to occur. However, the underlyingassumption that the circular forms of viral DNAare inherently unstable in cells in vivo remainsunsupported, and there are several lines of evi-dence in favor of the idea that the low-level vir-emia observed in the majority of patients onlong-term suppressive therapy is derived fromcells infected before the start of treatment.

1. The distribution of the levels of viremiaamong treated patients is independent ofthe drugs used for treatment (Maldarelliet al. 2007). This result is not what wouldbe expected for replication in a sanctuarybecause it is improbable that such sites orcells would be equally unavailable to differ-ent drugs.

2. The level of viremia is not affected by theinclusion of another drug. Several studieshave been performed in which a suppressivethree-drug regimen was intensified by theaddition of a fourth inhibitor of a class previ-ously unseen by the patient (Dinoso et al.2009; Gandhi et al. 2010; McMahon et al.2010). Despite some differences in designand inhibitors used, all studies gave thesame result: There was no difference in theamount of virus during and after the intensi-fication period as compared with before.Again, it is highly improbable that sanctuariesof replication would be equally inaccessible toall drugs, and these results are inconsistentwith low-level ongoing replication.

3. As is discussed in more detail in the next sec-tion, accumulation of genetic variation due

to the rapid turnover of infected cells is ahallmark of HIV-1 infection. In the chronicinfection phase, there is a slow but percepti-ble turnover of the viral population such thatgenetically distinct populations arise everyfew years (F Maldarelli, M Kearney, S Palmer,et al., unpubl.). In contrast, in patients onmaximally suppressive therapy, evolution ofviral genomes is frozen at the point at whichtherapy began (Ruff et al. 2002; Tobin et al.2005), and no evidence for continuing evo-lution can be seen even after many years. Inpatients for which therapy is partially sup-pressive, ongoing evolution can still beobserved (Tobin et al. 2005), as is also truein patients known as “elite controllers,” whonaturally control HIV-1 infection at levelssimilar to those seen in patients on suppres-sive antiviral therapy (Mens et al. 2010). Theapparent lack of evolution on therapy isconsistent with the absence of new cycles ofinfection.

4. As noted, the appearance of 2-LTR circles in aminority of patients undergoing intensifica-tion with an integrase inhibitor raises thepossibility of a low level of infection occur-ring in the face of antiviral drugs. However,this result remains to be fully understoodor reproduced.

5. It has recently been suggested, on the basis ofmathematical and cell culture modeling, thatthe high multiplicity of infection that wouldresult from cell–cell spread could greatlyreduce the effectiveness of antiviral therapyand allow persistent viral replication (Sigalet al. 2011). However, this model predictsthat cells infected in vivo should have multi-ple proviruses and that genetic variationshould continue to accrue on suppressivetherapy, neither of which is observed.

Regarding point 2, in the small amount ofvirus in the blood of some patients whose vire-mia has been suppressed for more than 3 yr orso, significant subpopulations of virus withgenomes identical in sequence can be observedand, in some cases, can become a large fractionof the total population (Bailey et al. 2006).



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



These subpopulations have been called pre-dominant plasma clones (PPCs). Their originis unknown, but it can be speculated that theymay arise from a very small fraction of infectedcells that is capable of dividing and releasingvirus but that do not die as a result of the expan-sion. Such clonal populations of virus can alsobe seen in the blood of patients who havestopped suppressive therapy after long-termsuppression (Joos et al. 2008), although thetwo phenomena have not yet been firmly con-nected. It seems likely that the last, apparentlystable, phase of viremia on therapy reflects clo-nal viral populations, but this point remains tobe established.

It is important to keep in mind that theinfected cell populations that give rise to long-lived, persistent viremia on therapy are not cre-ated by the therapy; rather, they must existthroughout the course of infection, only to beexposed as the shorter-lived cells of phases 1and 2 die off and are not replenished and thevirus produced by them disappears. The cellsresponsible for the production of the persistentvirus must represent an extremely small fractionof the total population of productively infectedcells in chronic infection—somewhere between1024 and 1025 is a reasonable estimate. Giventhat HIV-1 infection is carried forward, day inand day out, by repeated cycles of infection,viral production, and death of target cells, theselong-lived cell populations must be irrelevant tothe natural history of HIV-1 infection (as well asSIV infection of the natural hosts). They are ofcritical importance to attempts to treat HIV-1infection, however, because the virus they pro-duce is capable of rekindling full-blown infec-tion from the faintest embers remaining afterprolonged completely suppressive therapy.

GENETIC VARIATION AND EVOLUTION

One of the first unusual features of HIV-1 infec-tion to become apparent was the remarkableaccumulation of genetic diversity during thecourse of infection of a single individual.Indeed, when first observed, it represented thefastest rate of evolution in any natural eukary-otic system; for example, the extent of HIV-1

diversity that accumulates in one infected indi-vidual exceeds that of all influenza virus isolatesworldwide in any given year (Korber et al.2001). (Actually, the record diversity no longerbelongs to HIV, having since been eclipsed byhepatitis C virus.) Although it was first thoughtthat the remarkable diversification reflected anextraordinarily high underlying error rate asso-ciated with HIV-1 replication, we now knowthat this rate is about average for an RNA virusand that it is the pace of replication, the dura-tion of infection, and the size of the replicatingpopulation, allowing it to evolve rapidly inresponse to selective influences, that set HIV-1(and HCV) apart from all other known viralinfections. These features, in addition to theease of generating large numbers of sequenceseither directly from infected individuals andpopulations of individuals with well-timedinfections or from massive databases of suchsequences (Kuiken et al. 2003), have madeHIV-1 a favorite topic of study for evolutionarybiologists, and they have used it effectively todevelop and test new mathematical models forevolution in general, not just for HIV-1 or justfor viruses. We do not present any specific mod-els here, but the next section discusses the role ofspecific factors of evolution that all modelsmust include and how they apply to HIV-1 inan infected human host.

Factors of HIV-1 Evolution

By “evolution,” we mean the process by whichgenetic change accumulates during generationsof replication. For any replicating entity, or-ganisms or viruses, the general principles ofevolution are similar, although their applicationcan differ greatly. As a rule, models are based onconsideration of four principal factors. We dis-cuss them here in the context of HIV.

Mutation

In the case of HIV-1 and other retroviruses, mu-tations are primarily attributed to error-pronereverse transcription, but there are actuallytwo other possible sources of error: transcrip-tion by host RNA polymerase II, and G-to-Ahypermutation mediated by ABOBEC3G (or



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



F). As discussed in other articles in this col-lection (Hu and Hughes 2011; Karn and Stoltz-fus 2011; Malim and Bieniasz 2011), the relativecontribution of these mechanisms is a matter ofsome debate. Whatever the source, the overallsingle-step point mutation rate for HIV-1 is�3 � 1025 mutations per base per replicationcycle, and about 10-fold less for transversionsthan transitions (Mansky and Temin 1995).Thus, about one genome in three contains amutation after a single round of replication.Other types of mutation, including frameshifts(insertion or deletion of small numbers ofbases), as well as large-scale deletions, inser-tions, duplications, and inversions, are also com-mon (Svarovskaia et al. 2003). The error rate ofretrovirus replication, while much greater thanthat of cellular and viral DNA polymerases, isabout average for RNA viruses.

Selection results from the ability of one var-iant to replicate more effectively than another, aproperty referred to as “fitness.” In the case ofviruses, fitness differences lead to differencesin the number of infectious progeny at each rep-lication cycle. Fitness differences can result fromdifferences at any stage of the series of eventsleading from one infected cell to the next,including efficiency of entry, rate of infectiousprogeny production by an infected cell, and itsproductive lifetime, stability of the infectiousvirion, and so on. Note that there is no suchthing as absolute fitness: The concept is mean-ingful only in relative terms and only in a spe-cific context (e.g., in vitro vs. in vivo, in thepresence or absence of neutralizing antibodiesor antiviral drugs). It is often expressed as aselection coefficient, the difference in the rela-tive number of progeny produced by one var-iant compared with another per generation.Thus, for a variant that yields 1% more in-fectious progeny than another, the selectioncoefficient is 0.01. For large populations under-going multiple repeated cycles of infection, evensmall selection coefficients can have majoreffects on the frequency of slightly advanta-geous or disadvantageous mutations, with theformer leading to positive selection and thelatter to purifying selection. Because synony-mous mutations (those that do not change the

corresponding amino acid) generally have muchsmaller selection coefficients than nonsynony-mous ones, the ratio between the two in a pop-ulation (dN/dS) is a useful indicator of thenature of selective forces acting at any givensite or set of sites.

Drift

“Drift” is a term for the change in frequency of avariant independent of selection or mutation,resulting from sampling error that arises whenreplicating population sizes are small. In thecase of viral populations, it is important tobear in mind that the population size is deter-mined by the number of cells infected at eachround of replication, not by the number of viri-ons produced by them. In the case of HIV-1infection, perhaps 1011 virions are produceddaily; the number of cells infected in the sametime span is still a matter of some debate, butis unlikely to exceed 109. A further importantconsideration is the effective population size:in simple terms, the number of infected cellsmultiplied by the probability that any one cellgives rise to progeny that go on to infect othertarget cells. If there is an infinite supply of targetcells and the population of infected and targetcells is well mixed, then the effective populationsize can be similar to the census population size(defined as the total number of infected cells).These conditions obviously do not apply ininfected individuals, and the effective popula-tion size is probably much smaller. At very largeeffective population sizes, evolution can be con-sidered to be deterministic, in that the fre-quency of a mutation at any given time (ingenerations) in the future can be exactly deter-mined from its present frequency, mutationrate, and selection coefficient, as long as thesefactors remain constant. At smaller sizes, evolu-tion becomes more and more stochastic, that is,subject to random factors, so that the frequencycan only be predicted as an average, with thevariance increasing with decreasing size. A use-ful relationship to keep in mind is that, aftermany generations of deterministic evolution,slightly deleterious mutations will reach anequilibrium frequency equal to the mutation



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



rate divided by their selection coefficient. In thestochastic case, this calculation will give theaverage value over many observations but willvary considerably from one individual to thenext. This consideration is of more than aca-demic interest because it defines the probabilitythat drug-resistance mutations, which areslightly deleterious in the absence of the drug,will already be present in the replicating viralpopulation before initiation of antiviral ther-apy (the so-called genetic barrier) (see Artsand Hazuda 2011). Although transition fromdeterministic to stochastic evolution is continu-ous with decreasing population size, a usefulapproximate crossover point to keep in mindis a size equal to the inverse of the mutationrate, at which it becomes probable that all pos-sible single mutations will arise at each round ofreplication (Rouzine and Coffin 2005).

Linkage

To complicate matters further, selection doesnot act on individual mutations, but rather onwhole genomes, which may contain othermutations as well. In very large deterministicpopulations, where any mutation is alwayspresent in many genomes, this effect is unim-portant; in smaller populations, it can preventselection from operating on individual muta-tions, and even, in extreme cases (and at leastin theory) lead to declining fitness of a popula-tion due to gradual accumulation of deleteriousmutations (Muller’s ratchet). Linkage can thusbe visualized as increasing the role of stochas-tic effects on evolution, even at large populationsizes. The effect of linkage on evolution can bereduced by recombination to allow mutationsto be redistributed among genomes. It is im-portant to keep in mind that recombinationdoes not generate diversity, but rather redistrib-utes it among replicating genomes, allowingeffects of selection to be felt more strongly. Asdiscussed elsewhere (Hu and Hughes 2011),recombination between genetically distinct ge-nomes in retrovirus replication occurs at veryhigh rates, but only during reverse transcriptionof genomes from heterozygous progeny of cellscontaining two or more proviruses. Although

it was originally reported that such multiplyinfected cells are very frequent in HIV-infectedindividuals (Jung et al. 2002), more recentresults suggest that they are much rarer, consti-tuting perhaps 10% of infected CD4þ T cellsin blood (Josefsson et al. 2011). Nevertheless,there is ample evidence for the occurrence andimportance of recombination in HIV-1 infec-tion, particularly in rare individuals coinfectedwith two or more genetically distinct viruses(Keele et al. 2008; Kearney et al. 2009).

Course of Evolution following Infection

Monitoring of genetic diversity of the virusfollowing HIV-1 infection has provided a veryimportant tool for understanding many impor-tant aspects of the virus–host interaction. Forreasons discussed above, the sequences of viralgenomes in the blood provide the most accurateinstantaneous indicator for the genetics of thereplicating viral population, particularly whensingle-genome sequencing (SGS) technology,which avoids many PCR artifacts (Palmeret al. 2005), is used. Figure 4 shows, in cartoonform, the important events in virus evolutionduring the course of a “typical” infection, start-ing with a clonal population and proceedingtoward the highly diverse populations charac-teristic of chronic infection. The initial accumu-lation of diversity is at a rate very close to themutation rate, with little evidence for selection,except for highly deleterious mutations; as timegoes on, purifying selection takes hold, and theincrease in diversity begins to level off, ap-proaching an asymptotic value of an averagepairwise distance �1%–2% for gag, pro, andpol, which are largely subject to purifying selec-tion, as indicated by low dN/dS ratios. Rates aremuch higher for env, where the greater, and con-stantly increasing, diversity and higher dN/dS

ratios point to constant positive selection result-ing from an ever-changing antibody response(Richman et al. 2003). In the case of gag, pro,and pol (as well as accessory genes, particularlynef ), even though dN/dS ratios are low, positiveselection can be discerned at sites correspond-ing to T-cell epitopes. Indeed, virtually all ofthe nonsynonymous diversity accumulating in



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



these genes in untreated infection can be attrib-uted to escape from the cytotoxic T-cell (CTL)response (Kearney et al. 2009; Walker andMcMichael 2011). In env, escape from the anti-body response is characterized not only bypoint mutation, particularly affecting the dis-play of glycosylation sites, but also by grossstructural changes—insertions and duplica-tions—in the hypervariable loops V1, V2, andV4, which tend to increase in length and num-bers of glycosylation sites (the “glycan shield”)

during the course of infection. It is importantto bear in mind that, given the conditions underwhich HIV populations are evolving, even weakselective forces (on the order of a few percent)can have dramatic effects on the genetics ofthe viral population. Thus, it cannot be inferredfrom these results alone that the immune re-sponse plays a significant role in controllingHIV infection. The partial control of viralreplication by the host is likely the sum of effectsof the immune response and the fitness cost

12108

Time since infection (yr)

Reference (consensus) HIV-1

Genetic distance

64200

0.5

1.0

2.0A

B

Initiate suppressive therapy

1.5

Ave

rage

pai

rwis

e di

ffere

nce

(per

cent

)

103

102

101

104

105

106

Vire

mia

(R

NA

cop

ies/

mL)

Figure 4. Typical evolution of HIV after infection. (A) The level of viremia (orange) and the genetic diversity,measured as the average pairwise difference (black) of the HIV population as a function of time after infection.(B) Phylogenetic relationship of HIV-1 sequences at various stages of infection. The colored circles correspondto sequences of plasma virus RNA sampled at the corresponding time points in A: At peak viremia (yellow),the viral population (in most patients) is highly uniform, diversifying gradually over the first few years ofinfection. Samples taken from chronic infection (green and blue) are considerably more diverse, but take 3yr or more to diverge perceptibly (red). Following initiation of suppressive therapy, divergence ceases, evenafter long times (orange, gray). After years of therapy, clonal populations of virus arise in some patients(black).



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



incurred with the initial escape by the virus, afitness cost that is in part alleviated over timethrough compensatory mutations.

A remarkable aspect of HIV genetic diver-sity during chronic infection is its stability.Except in the case of escape from antiretroviraltherapy, genetic turnover of the viral populationin the blood is very slow, with no evidence ofa genetic or physical bottleneck (Shankarappaet al. 1999; F Maldarelli, M Kearney, and SPalmer, unpubl.). This observation is consistentwith a large replicating population that can besteered by immune selection, but without everhaving to recover from significant depletion.Because mutations leading to immune escapewill often reduce fitness, when that immune re-sponse is not present, such as after transmissionto another person, these mutations will tend torevert to the consensus sequence, although athighly variable rates. For example, mutationsincreasing the length of the variable loops tendto reset to shorter length following transmission(Shaw and Hunter 2011).

In addition to immune escape, resistance toantiretroviral therapy is an important conse-quence of genetic diversity accumulating duringchronic HIV-1 infection. Both theory (dis-cussed above) and observation are consistentwith the idea that drug-resistance mutationsthat accumulate to low levels before therapyare an important cause of failure. Indeed, in vir-tually all cases, treatment of HIV-1 infectionwith a single agent (monotherapy) leads to rapidreappearance of viremia, rebounding to levelsnear the pretherapy set point. A striking exam-ple is in women treated with a single dose of thenonnucleoside RT inhibitor (NNRTI) nevira-pine to prevent mother-to-child transmission(Chersich and Gray 2005). A single pill takenduring labor prevents a significant fraction ofinfections of the newborn but leads to a veryhigh frequency of drug-resistant virus in themother a few weeks later. The reproducibilityof this dramatic effect implies a low (and sofar undetectable) frequency of NNRTI-resistantmutants in the replicating viral populationcombined with a relatively long lifetime of thedrug, allowing their selection during the ensu-ing cycles of HIV-1 replication.

Unlike the immune response, antiviral ther-apy can impose a significant physical and genet-ic bottleneck on the HIV-1 population, withdrug-resistant virus that rebounds after mono-therapy often showing substantially less diver-sity than the starting population (Kitrinos et al.2005).

Another important consequence of HIVgenetic variation is the appearance of popula-tions of virus infecting specific anatomicalcompartments and cell types. As discussedby Swanstrom and Coffin (2011), HIV can befound in a large number of organs and tissues.In some sites, such as the central nervoussystem (CNS), there is a clear genetic separationof independently evolving populations, oftenleading to macrophage-tropic viruses that likelyreplicate in either macrophages or microglialcells in the brain. Macrophage-tropic virusescan also be found late in infection at otheranatomical sites, but their pathophysiologicalsignificance is much less clear. The final impor-tant consequence of HIVevolution in vivo is theappearance of virus capable of using alternativecoreceptors, such as CXCR4, distinct from theCCR5 chemokine receptor used by the foundervirus.

HOW DOES HIV-1 CAUSE AIDS?

As is apparent from this article and the rest ofthe collection, in the 25þ years since its dis-covery, we have learned an enormous amountabout HIV, but we still cannot answer the onebig question: How does HIV-1 cause AIDS? Atfirst thought, the answer is obvious. HIV-1infects and kills CD4þ T cells. AIDS is a resultof loss of CD4þ T cells. QED. Unfortunately, itis not that simple.

For starters, although HIV-1 infection killsinfected cells ex vivo, by a mechanism that is stillcontroversial, there is still considerable debateabout how infected cells die in vivo. There arethree alternatives: virus-mediated killing ofinfected cells, as in culture systems; immune-mediated effects due to HIV-specific CD8þ

T cells or antibodies; or one or more of a varietyof indirect effects. In support of the role of suchdirect killing is the short half-life of infected



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



cells early in infection, before the appearance ofantibodies or CTL, ruling out the adaptiveimmune response as essential for cell killing invivo. The innate (i.e., interferon) responsecould still play some role at this point, however.Although the antibody response appears to havelittle effect on the life of the infected cell, there isevidence for an important role of the CTLresponse in cell killing because treatment ofSIV-infected macaques with antibody directedagainst CD8 leads to a significant (100-fold)but transient increase in the level of viremia(Jin et al. 1999; Schmitz et al. 1999). The moststraightforward explanation for this result isthat CTLs shorten the productive lifetime ofthe infected cell, although more indirect effectshave also been proposed (Davenport and Pet-ravic 2010; Klatt et al. 2010; Wong et al. 2010).Two features of this hypothesis are worth not-ing. First, the productive lifetime of an infectedcell is the time from the onset of virus produc-tion to cell death by CTL and is likely to benot much shorter than the lifetime defined byvirus-mediated killing, perhaps only an houror so. Second, it is likely that part of the advant-age conferred on HIV-1 (and other complexviruses) by the two-step replication strategymediated by tat and rev is the delay of the syn-thesis of virion proteins as long as possible inthe life cycle to minimize exposure to the CTLresponse and maximize virion output per cell.Although cells can die from the effects of infec-tion alone, their productive lifetime is some-what shortened by the CTL response, that is,specifically during the time when virus is beingproduced. Thus, CTLs can modulate infection,but they do not eliminate it. Effects due toimmune activation could include not only deg-radation of immune function late in infection,but also stimulation of production of activatedmemory CD4þ T cells, increasing the numberof targets available to the virus.

Even if we knew the mechanism of HIV-mediated cell killing, we would not know howHIV-1 causes CD4þ T-cell decline and AIDSin humans. The observation that virus and cellturnover rates in various SIVs in their naturalhosts (such as SIVsm in sooty mangabeys),which do not progress to AIDS, are essentially

identical to those in humans, who do progress,implies that cell killing alone cannot accountfor AIDS pathogenesis. Indeed, this result isconsistent with the high natural turnover rateof activated effector memory helper T cells,the primary target for HIV-1 infection, on theorder of 1010 cells per day, of which only a smallfraction are infected after the initial primaryinfection phase. Clearly, in the natural host,these viruses have evolved to replicate in a pop-ulation of cells that is replenished constantlyand in excess over what is needed for goodhealth. In unnatural hosts, such as HIV-1 infec-tion of humans, or SIVsm infection of ma-caques, something else must be happening tolead to AIDS. Two features that distinguish thepathogenic from nonpathogenic infection arechronic immune activation, discussed above,and greater levels of infection of other CD4þ

T-cell subsets, particularly central memoryhelper T cells. The former may lead eventuallyto immune exhaustion and damaged lymphoidarchitecture, the latter (which might, itself, be aresult of immune activation) to loss of criticalcells necessary for maintenance of a diverseresponse to common environmental pathogens.Sorting out the roles of these and other sideeffects of HIV-1 infection in the causation ofAIDS will provide subject matter for HIV-1researchers for a long time to come.

SUMMARY AND PERSPECTIVE

Quantitation of viral RNA in plasma, of viralDNA in cells, and of cells capable of beingstimulated to express quiescent proviruses hasled to a conceptual framework of the dynamicsof HIV-1 replication and persistence in the host.The availability of suppressive antiviral therapyhas brought disease progression under control,and it has also provided a means to perturbthe dynamics of viral replication that hasrevealed additional features of viral persistence,features that show the necessity for lifelong ther-apy given the current treatment strategies. Thedynamic changes in the sequence of the viralgenome provide another critical source of infor-mation about viral replication and the selectivepressures at work.



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



Given the ongoing generation of HIV-1sequence information and the dynamic natureof viral populations, this virus will continue toprovide the basis for studies of viral evolutionand evolution in general. Deep sequencingstrategies will reveal features of the viral popu-lation that have previously been hidden. Therewill be continuing efforts to define selectivepressures in the context of sequence changes,thus providing a link between the selective pres-sure and the affected viral protein. Viral dynam-ics and evolution will continue to be centraltools as our understanding of HIV-1 deepens.The lessons learned to date provide the founda-tion for continuing studies of virus–host inter-actions and for efforts to develop strategies toeradicate the virus completely.

REFERENCES�Reference is also in this collection.

� Arts EJ, Hazuda DJ. 2011. HIV-1 antiretroviral drug therapy.Cold Spring Harb Perspect Med doi: 10.1101/cshperspect.a007161.

Bailey JR, Sedaghat AR, Kieffer T, Brennan T, Lee PK, Wind-Rotolo M, Haggerty CM, Kamireddi AR, Liu Y, Lee J, et al.2006. Residual human immunodeficiency virus type 1viremia in some patients on antiretroviral therapy isdominated by a small number of invariant clones rarelyfound in circulating CD4þ T cells. J Virol 80: 6441–6457.

Buzon MJ, Massanella M, Llibre JM, Esteve A, Dahl V, Puer-tas MC, Gatell JM, Domingo P, Paredes R, Sharkey M,et al. 2011. HIV-1 replication and immune dynamicsare affected by raltegravir intensification of HAART-suppressed subjects. Nat Med 16: 460–465.

Chersich MF, Gray GE. 2005. Progress and emerging chal-lenges in preventing mother-to-child transmission.Curr Infect Dis Rep 7: 393–400.

Chun TW, Carruth L, Finzi D, Shen X, DiGiuseppe JA, Tay-lor H, Hermankova M, Chadwick K, Margolick J, QuinnTC, et al. 1997. Quantification of latent tissue reservoirsand total body viral load in HIV-1 infection. Nature387: 183–188.

Coffin JM. 1995. HIV population dynamics in vivo: Impli-cations for genetic variation, pathogenesis, and therapy.Science 267: 483–488.

� Craigie R, Bushman FD. 2011. HIV DNA integration. ColdSpring Harb Perspect Med doi: 10.1101/cshperspect.a006890.

Davenport MP, Petravic J. 2010. CD8þ T cell control ofHIV—A known unknown. PLoS Pathog 6: e1000728.doi: 10.1371/journal.ppat.1000728.

Dinoso JB, Kim SY, Wiegand AM, Palmer SE, Gange SJ,Cranmer L, O’Shea A, Callender M, Spivak A, BrennanT, et al. 2009. Treatment intensification does not reduce

residual HIV-1 viremia in patients on highly active anti-retroviral therapy. Proc Natl Acad Sci 106: 9403–9408.

Fauci AS, Desrosiers RC. 1997. Pathogenesis of HIVand SIV.In Retroviruses (ed. Coffin JM, et al.), pp. 587–635.Cold Spring Harbor Laboratory Press, Cold SpringHarbor, NY.

Gandhi RT, Bosch RJ, Aga E, Albrecht M, Demeter LM,Dykes C, Bastow B, Para M, Lai J, Siliciano RF, et al.2010. No evidence for decay of the latent reservoir inHIV-1-infected patients receiving intensive enfuvirtide-containing antiretroviral therapy. J Infect Dis 201:293–296.

Ho DD, Neumann AU, Perelson AS, Chen W, Leonard JM,Markowitz M. 1995. Rapid turnover of plasma virionsand CD4 lymphocytes in HIV infection. Nature 373:123–126.

� Hu WS, Hughes SH. 2011. HIV-1 reverse transcription. ColdSpring Harb Perspect Med doi: 10.1101/cshperspect.a006882.

Jin X, Bauer DE, Tuttleton SE, Lewin S, Gettie A, BlanchardJ, Irwin CE, Safrit JT, Mittler J, Weinberger L, et al. 1999.Dramatic rise in plasma viremia after CD8þ T cell deple-tion in simian immunodeficiency virus-infected ma-caques. J Exp Med 189: 991–998.

Joos B, Fischer M, Kuster H, Pillai SK, Wong JK, Boni J, Hir-schel B, Weber R, Trkola A, Gunthard HF. 2008. HIVrebounds from latently infected cells, rather than fromcontinuing low-level replication. Proc Natl Acad Sci105: 16725–16730.

Josefsson L, King MS, Makitalo B, Brannstrom J, Shao W,Maldarelli F, Kearney MF, Hu WS, Chen J, Gaines H,et al. 2011. Majority of CD4þ T cells from peripheralblood of HIV-1-infected individuals contain only oneHIV DNA molecule. Proc Natl Acad Sci 108: 11199–11204.

Jung A, Maier R, Vartanian JP, Bocharov G, Jung V, FischerU, Meese E, Wain-Hobson S, Meyerhans A. 2002. Recom-bination: Multiply infected spleen cells in HIV patients.Nature 418: 144.

� Karn J, Stoltzfus CM. 2011. Transcriptional and post-tran-scriptional regulation of HIV-1 gene expression. ColdSpring Harb Perspect Med doi: 10.1101/cshperspect.a006916.

Kearney M, Maldarelli F, Shao W, Margolick JB, Daar ES,Mellors JW, Rao V, Coffin JM, Palmer S. 2009. Humanimmunodeficiency virus type 1 population genetics andadaptation in newly infected individuals. J Virol 83:2715–2727.

Keele BF, Giorgi EE, Salazar-Gonzalez JF, Decker JM, PhamKT, Salazar MG, Sun C, Grayson T, Wang S, Li H, et al.2008. Identification and characterization of transmittedand early founder virus envelopes in primary HIV-1infection. Proc Natl Acad Sci 105: 7552–7557.

Kitrinos KM, Nelson JA, Resch W, Swanstrom R. 2005.Effect of a protease inhibitor-induced genetic bottleneckon human immunodeficiency virus type 1 env gene pop-ulations. J Virol 79: 10627–10637.

Klatt NR, Shudo E, Ortiz AM, Engram JC, Paiardini M, Law-son B, Miller MD, Else J, Pandrea U, Estes JD, et al. 2010.CD8þ lymphocytes control viral replication in SIV-mac239-infected rhesus macaques without decreasing



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



the lifespan of productively infected cells. PLoS Pathog 6:e1000747.

Korber B, Gaschen B, Yusim K, Thakallapally R, KesmirC, Detours V. 2001. Evolutionary and immunologicalimplications of contemporary HIV-1 variation. Br MedBull 58: 19–42.

Kuiken C, Korber B, Shafer RW. 2003. HIV sequence data-bases. AIDS Rev 5: 52–61.

� Lackner AA, Lederman MM, Rodriguez B. 2011. HIVpathogenesis—The host. Cold Spring Harb Perspect Meddoi: 10.1101/cshperspect.a007005.

Maldarelli F. 2011. Targeting viral reservoirs: Ability of anti-retroviral therapy to stop viral replication. Curr Opin HIVAIDS 6: 49–56.

Maldarelli F, Palmer S, King MS, Wiegand A, Polis MA,Mican J, Kovacs JA, Davey RT, Rock-Kress D, Dewar R,et al. 2007. ARTsuppresses plasma HIV-1 RNA to a stableset point predicted by pretherapy viremia. PLoS Pathog 3:e46. doi: 10.1371/journal.ppat.0030046.

� Malim MH, Bieniasz PD. 2011. HIV restriction factors andmechanisms of evasion. Cold Spring Harb Perspect Meddoi: 10.1101/cshperspect.a006940.

Mansky LM, Temin HM. 1995. Lower in vivo mutation rateof human immunodeficiency virus type 1 than thatpredicted from the fidelity of purified reverse transcrip-tase. J Virol 69: 5087–5094.

McMahon D, Jones J, Wiegand A, Gange SJ, Kearney M,Palmer S, McNulty S, Metcalf JA, Acosta E, Rehm C,et al. 2010. Short-course raltegravir intensification doesnot reduce persistent low-level viremia in patients withHIV-1 suppression during receipt of combination antire-troviral therapy. Clin Infect Dis 50: 912–919.

Mellors JW, Rinaldo CRJ, Gupta P, White RM, Todd JA,Kingsley LA. 1996. Prognosis in HIV-1 infection pre-dicted by the quantity of virus in plasma. Science 272:1167–1170.

Mens H, Kearney M, Wiegand A, Shao W, Schonning K,Gerstoft J, Obel N, Maldarelli F, Mellors JW, Benfield T,et al. 2010. HIV-1 continues to replicate and evolve inpatients with natural control of HIV infection. J Virol84: 12971–12981.

Murray JM, Emery S, Kelleher AD, Law M, Chen J, HazudaDJ, Nguyen BY, Teppler H, Cooper DA. 2007. Antiretro-viral therapy with the integrase inhibitor raltegravir altersdecay kinetics of HIV, significantly reducing the secondphase. AIDS 21: 2315–2321.

Palmer S, Wiegand AP, Maldarelli F, Bazmi H, Mican JM,Polis M, Dewar RL, Planta A, Liu S, Metcalf JA, et al.2003. New real-time reverse transcriptase-initiated PCRassay with single-copy sensitivity for human immunode-ficiency virus type 1 RNA in plasma. J Clin Microbiol 41:4531–4536.

Palmer S, Kearney M, Maldarelli F, Halvas EK, Bixby CJ,Bazmi H, Rock D, Falloon J, Davey RT Jr, Dewar RL,et al. 2005. Multiple, linked human immunodeficiencyvirus type 1 drug resistance mutations in treatment-expe-rienced patients are missed by standard genotype analy-sis. J Clin Microbiol 43: 406–413.

Palmer S, Maldarelli F, Wiegand A, Bernstein B, Hanna GJ,Brun SC, Kempf DJ, Mellors JW, Coffin JM, King MS.2008. Low-level viremia persists for at least 7 years in

patients on suppressive antiretroviral therapy. Proc NatlAcad Sci 105: 3879–3884.

Perelson AS, Essunger P, Cao Y, Vesanen M, Hurley A, Sak-sela K, Markowitz M, Ho DD. 1997. Decay characteristicsof HIV-1-infected compartments during combinationtherapy. Nature 387: 188–191.

Richman DD, Wrin T, Little SJ, Petropoulos CJ. 2003. Rapidevolution of the neutralizing antibody response to HIVtype 1 infection. Proc Natl Acad Sci 100: 4144–4149.

Rouzine IM, Coffin JM. 2005. Evolution of human immu-nodeficiency virus under selection and weak recombina-tion. Genetics 170: 7–18.

Ruff CT, Ray SC, Kwon P, Zinn R, Pendleton A, Hutton N,Ashworth R, Gange S, Quinn TC, Siliciano RF, et al.2002. Persistence of wild-type virus and lack of tem-poral structure in the latent reservoir for human im-munodeficiency virus type 1 in pediatric patients withextensive antiretroviral exposure. J Virol 76: 9481–9492.

Schmitz JE, Kuroda MJ, Santra S, Sasseville VG, Simon MA,Lifton MA, Racz P, Tenner-Racz K, Dalesandro M, ScallonBJ, et al. 1999. Control of viremia in simian immunode-ficiency virus infection by CD8þ lymphocytes. Science283: 857–860.

Shankarappa R, Margolick JB, Gange SJ, Rodrigo AG,Upchurch D, Farzadegan H, Gupta P, Rinaldo CR, LearnGH, He X, et al. 1999. Consistent viral evolutionarychanges associated with the progression of humanimmunodeficiency virus type 1 infection. J Virol 7312:10489–10502.

Sharkey ME, Stevenson M. 2001. Two long terminal repeatcircles and persistent HIV-1 replication. Curr Opin InfectDis 14: 5–11.

� Shaw GM, Hunter E. 2011. HIV transmission. Cold SpringHarb Perspect Med doi: 10.1101/cshperspect.a006965.

Sigal A, Kim JT, Balazs AB, Dekel E, Mayo A, Milo R, Balti-more D. 2011. Cell-to-cell spread of HIV permitsongoing replication despite antiretroviral therapy. Nature477: 95–98.

� Siliciano RF, Greene WC. 2011. HIV latency. Cold SpringHarb Perspect Med doi: 10.1101/cshperspect.a007096.

Svarovskaia ES, Cheslock SR, Zhang WH, Hu WS, PathakVK. 2003. Retroviral mutation rates and reverse tran-scriptase fidelity. Front Biosci 8: D117–D134.

� Swanstrom R, Coffin J. 2011. HIV-1 pathogenesis: The virus.Cold Spring Harb Perspect Med doi: 10.1101/cshperspect.a007443.

Tobin NH, Learn GH, Holte SE, Wang Y, Melvin AJ,McKernan JL, Pawluk DM, Mohan KM, Lewis PF, Mul-lins JI, et al. 2005. Evidence that low-level viremias duringeffective highly active antiretroviral therapy result fromtwo processes: Expression of archival virus and replica-tion of virus. J Virol 79: 9625–9634.

� Walker B, McMichael A. 2011. The T-cell response to HIV.Cold Spring Harb Perspect Med doi: 10.1101/cshperspect.a007054.

Wei X, Ghosh S, Taylor ME, Johnson VA, Emini EA, DeutschP, Lifson JD, Bonhoeffer S, Nowak MA, Hahn BH, et al.1995. Viral dynamics in human immunodeficiency virustype 1 infection. Nature 373: 117–122.



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



Wong JK, Strain MC, Porrata R, Reay E, Sankaran-Walters S,Ignacio CC, Russell T, Pillai SK, Looney DJ, Dandekar S.2010. In vivo CD8þ T-cell suppression of siv viremia isnot mediated by CTL clearance of productively infectedcells. PLoS Pathog 61: e1000748.

Zhang L, Ramratnam B, Tenner-Racz K, He Y, Vesanen M,Lewin S, Talal A, Racz P, Perelson AS, Korber BT, et al.1999. Quantifying residual HIV-1 replication in patientsreceiving combination antiretroviral therapy. N Engl JMed 340: 1605–1613.



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



2013; doi: 10.1101/cshperspect.a012526Cold Spring Harb Perspect Med John Coffin and Ronald Swanstrom Infected CellsHIV Pathogenesis: Dynamics and Genetics of Viral Populations and

Subject Collection HIV

Viral Populations and Infected CellsHIV Pathogenesis: Dynamics and Genetics of

John Coffin and Ronald Swanstrom

HIV-1 Pathogenesis: The VirusRonald Swanstrom and John Coffin

Human Immunodeficiency Virus Vaccine Trials

Corey, et al.Robert J. O'Connell, Jerome H. Kim, Lawrence

The T-Cell Response to HIVBruce Walker and Andrew McMichael

HIV TransmissionGeorge M. Shaw and Eric Hunter

HIV-1 Reverse TranscriptionWei-Shau Hu and Stephen H. Hughes

Novel Cell and Gene Therapies for HIVJames A. Hoxie and Carl H. June

HIV Pathogenesis: The Host

RodriguezA.A. Lackner, Michael M. Lederman and Benigno

Strategies for HIV PreventionBehavioral and Biomedical Combination

QuinnLinda-Gail Bekker, Chris Beyrer and Thomas C.

HIV: Cell Binding and EntryCraig B. Wilen, John C. Tilton and Robert W. Doms

HIV-1 Assembly, Budding, and MaturationWesley I. Sundquist and Hans-Georg Kräusslich

Innate Immune Control of HIVMary Carrington and Galit Alter

HIV-1 Assembly, Budding, and MaturationWesley I. Sundquist and Hans-Georg Kräusslich

HIV DNA IntegrationRobert Craigie and Frederic D. Bushman

Vaccine Research: From Minefields to MilestonesLessons in Nonhuman Primate Models for AIDS

Jeffrey D. Lifson and Nancy L. Haigwood TreatmentCurrent Issues in Pathogenesis, Diagnosis, and HIV-1-Related Central Nervous System Disease:

Serena Spudich and Francisco González-Scarano

http://perspectivesinmedicine.cshlp.org/cgi/collection/ For additional articles in this collection, see

Copyright © 2013 Cold Spring Harbor Laboratory Press; all rights reserved


http://perspectivesinmedicine.cshlp.org/cgi/collection/

http://perspectivesinmedicine.cshlp.org/cgi/collection/


hiv pathogenesis: dynamics and genetics of viral...

Documents