hiv vaccines: prospects and challenges

MACROPHAGE ORRELATED ANTIGEN-PRESENTING CELL MACROPHAGE

INFECTED BY HIV

CD8 T CELL

HIV

CD4 T CELL

STIMULATORYSIGNALING

MOLECULES

B CELL

a

b

c

d

e

i

j

k

o

Scientists know more aboutHIV—the human immunodefi-ciency virus that causes AIDS—

than any other virus. Yet designing avaccine able to protect against it re-mains as much of a challenge today aswhen the virus was discovered. Part ofthe problem is that, unlike the body’sresponse to most acute viral infections,the natural immune response does notdestroy HIV. This failure makes it diffi-cult for investigators to know what typeof immune activity an effective vaccineshould evoke.

At the same time, researchers have tobe extremely cautious about using thepreparations that have become standardfor warding off other infectious dis-eases—such as whole, killed viruses orlive, attenuated versions. If HIV vaccinesin these forms managed to cause infec-tions, the consequences could be devas-tating. Vaccinologists therefore have hadto search for alternative ways to immu-nize people against HIV.

Vaccines protect individuals by prim-

ing the immune system to recognize dis-ease-causing organisms when they areencountered. In the case of HIV, a suc-cessful vaccine should be able to elimi-nate incoming virus and destroy quick-ly any cells that become infected.

Most vaccines activate what is calledthe humoral arm of the immune sys-tem, stimulating formation of protec-tive antibodies: molecules that mark freevirus (which circulates outside cells) fordestruction. The antibodies recognizeand bind to a unique part of the infec-tious agent. This unique structure,called an antigen, is often a proteinon the viral surface.

Foreign antigens on an in-vading virus or in a vaccineactivate two types of whiteblood cells involved in anti-

body manufacture. After contactingantigens, cells known as B lymphocytesmature and produce antibodies. In ad-dition, helper, or CD4, T lymphocytesdirect B cells to manufacture more anti-bodies or to take the form of memory Bcells. The memory cells do not produceantibodies immediately but respondvigorously to subsequent exposures.Following vaccination, the long-termproduction of small amounts of anti-body and the persistence of memory

98 Scientific American July 1998

HIV Vaccines:Prospects and Challengesby David Baltimore and Carole Heilman

DE

FE

AT

ING

AID

S

Copyright 1998 Scientific American, Inc.

MEMORY CD8 T CELL

PRIMED CD8 T CELL

KILLEDINFECTED CELL

MEMORYB CELL

ACTIVATED B CELL

VIRUS BEINGDESTROYED

n

f

h

l

m

g

HIV-INFECTED CELL

ANTIBODIES

Scientific American July 1998 99

cells allow the body to mount a rapiddefense if ever it encounters the virus.

No vaccines have been designed spe-cifically to stimulate the other arm of theimmune system, known as the cellularcomponent. But many AIDS researchersare working on just that aim because,thus far, vaccines designed to generateantibodies to HIV have failed to elicitimmunity against the strains of the virus

commonly found in infected patients.In cellular immunity, activated white

blood cells called cytotoxic T lympho-cytes (CD8 T cells) multiply and cruisethrough the bloodstream and tissues,searching for and eliminating virus-in-fected cells. Some also become memorycells, ready to leap into action after a

later exposure to a pathogen. Unlikeantibodies, cytotoxic T lymphocytes rec-ognize infected cells, rather than the in-fectious agent itself. Like B cells in thehumoral arm of the immune system,however, cytotoxic T cells are activatedin part by signals from helper T cells. In

HIV STIMULATES two kinds of immune responses. In the humoralresponse (upper sequence), a macrophage or related cell degradesHIV particles (a–c) and presents the fragments, or antigens, to whiteblood cells known as CD4, or helper, T lymphocytes (d ). These cellsthen release molecules (e) that direct B lymphocytes to mature ( f )and produce antibody molecules (g) able to mark HIV particles fordestruction (h).The cellular response (lower sequence) begins after anHIV-infected macrophage displays viral fragments that are recognizedby CD8, or cytotoxic, T lymphocytes (i–j). Helper T cells then promptthe cytotoxic cells to destroy other HIV-infected cells (k–m). Some ofthe B and T cells eventually become long-lived memory cells (n and o)that react promptly to future exposures to HIV. A successful HIV vac-cine may need to elicit both humoral and cellular immunity.

Unlike vaccines for many viruses, those for HIV may have to gobeyond generating antibodies. Devising approaches that will fully activate the immune system is far from simple

TOM

O N

AR

ASH

IMA


the long run, the most effective HIV vac-cines may well be the ones that stimu-late both the humoral and cellular armsof the immune system, generating anti-bodies and activated cytotoxic T cells.

Efforts to design an HIV vaccine thatmaximizes production of antibodies orstimulation of cytotoxic T cells have beenhampered by a lack of basic knowledgeabout how the immune system func-tions. Until investigators can learn howto induce the body to generate and main-tain memory cells and cytotoxic T cells,those attempting to develop HIV vac-cines will have to rely on a certainamount of trial and error, hoping to hiton an approach that will work.

The Antibody Approach

Vaccines that stimulate the produc-tion of protective antibodies have

proved successful for combating diseasessuch as poliomyelitis, measles and in-fluenza. At present, the most extensivelytested HIV vaccine candidates containsome part of the envelope protein (Env),the molecule that coats the surface ofthe virus. Because the virus uses Env asa kind of key for gaining entry to humancells, generating antibodies that attachto the business end of this protein shouldprevent HIV from binding to and in-fecting cells.

The Env protein, also called gp160, isactually an association of two units:gp120, a sugar-shrouded protein thatjuts out of the virus membrane and in-teracts with receptors on the surface ofhuman T lymphocytes, and gp41, the

small protein that anchors gp120 to themembrane. Both gp120 and gp160 havebeen tested as HIV vaccine candidatesin human volunteers.

In tests, the proteins elicited the pro-duction of antibodies, a result that raisedhopes that they might form the basis foran effective HIV vaccine. Further, the re-sulting antibodies effectively neutralizedlive HIV in a test tube, blocking its abilityto infect cultured human lymphocytes.

Unfortunately, the antibodies only rec-ognized strains of HIV that were similarto those used to generate the vaccines.The gp120 and gp160 proteins in thepreparations were made from HIVstrains that had been cultivated in thelaboratory. The antibodies elicitedagainst proteins from such lab-adaptedvirus strains were ineffective at neutral-izing HIV strains isolated directly frominfected patients; the isolates were quiteable to infect cultured cells.

Why did the antibodies fail to neutral-ize the HIV obtained directly from pa-tients? The structure of the Env proteinin laboratory-grown strains appears tobe somewhat looser than that of the sur-face protein in patient isolates; those inisolates are folded tightly. Antibodies tolaboratory strains of HIV may recognizeparts of the Env protein that are not nor-mally exposed in viruses from patients,probably because the recognition sitesare buried within the more folded formof the protein. Antibodies to laboratory-grown virus, then, would not “see” theirtargets on HIV isolated from patients.

Researchers are currently developingvaccines based on surface proteins pre-

pared from patient isolates. Such prepa-rations may present Env in the confor-mation found in patients. Yet even thesevaccines may not work. The Env proteinon such isolates may be very denselypacked and highly camouflaged by sug-ars. As a result, B cells may be unable tofind many antigens and so may producerelatively few kinds of antibodies. Suchan outcome would be consistent withthe finding that people who are infectedwith HIV generally produce a limitedrepertoire of antibodies that react withthe surface of HIV.

When Env binds to a cell, the proteinchanges its shape somewhat. A vaccinethat duplicates the conformation adopt-ed by gp120 as it attaches to receptorson the cell surface may succeed best atraising antibodies able to block HIVfrom infecting human cells.

Individuals who are infected with HIVbut remain healthy and keep viral repli-cation in check may offer some hopefor guiding the design of an effectiveHIV vaccine. Some of these long-termsurvivors make a very small amount ofantibody, which, when isolated, can neu-tralize HIV from patient isolates. Fur-ther, those antibodies can neutralizeviruses from many different patient iso-lates—a necessity for an AIDS vaccinethat will be effective against a broadspectrum of HIV strains. Unfortunately,even these antibodies may not be thewhole answer. Tests of cells in cultureindicate that the antibodies must be pres-ent at surprisingly high concentrationsto block HIV entry into cells effectively.

Pure protein vaccines may not be the

HIV Vaccines: Prospects and Challenges 100 Scientific American July 1998

ANTIBODIES HAVE NO EFFECT

ANTIBODY BLOCKS HIV FROMINFECTING CELL

CD4 CO-RECEPTOR

LABORATORY-GROWN HIV

ENV PROTEIN FROM LABORATORY-GROWN HIV

ANTIBODIESAGAINST ENV

ACTIVE B CELL

B CELL

a b

CELL

PURE ENV PROTEIN, isolated from virus grown in the labora-tory, has been studied as a vaccine. The protein successfully in-duced B lymphocytes to make antibodies that recognized the Envprotein (left panel). Further, the antibodies prevented laboratory-

grown HIV from infecting cultured cells (a), perhaps by blockingbinding to cell-surface receptors or by enhancing elimination ofthe virus. Disappointingly, though, those antibodies have not beenable to bar infection by virus isolated directly from patients (b).

TOM

O N

AR

ASH

IMA

HIVFROMPATIENT


best way to stimulate antibody produc-tion: in isolation, gp120 does not appearto have a precise conformation, andgp160 clumps into an ineffective aggre-gate. To get around these difficulties, re-searchers are currently testing two dif-ferent vaccine strategies designed topresent the Env proteins in a more nat-ural conformation.

One plan of attack involves usingwhole, killed virus particles. This dis-abled form of HIV, incapable of multi-plying, might present the immune sys-tem with more natural forms of Envproteins. With a better target, B cellsmight produce a better quality and ahigher quantity of protective antibody.

Making a killed-virus vaccine requiresa rigorous inactivation procedure, be-cause residual virus and even residualviral genetic material could potentiallybe dangerous. Harsh treatment makesthe vaccine less effective, however; theinactivation process can cause HIV toshed its weakly attached gp120. Manyresearchers have therefore been movingaway from this design, although thegp120 stability problem may ultimatelybe solvable.

Env proteins can also be presented tothe immune system embedded in “pseu-dovirions,” artificial structures that re-semble virus particles. These empty lipidshells could be made to carry nothingbut gp160. Pseudovirions would be saferthan whole, killed virus, because theylack the genes that could propagate HIVinfection. Unfortunately, pseudovirionsare very difficult to manufacture andproduce in a stable form. Researchershope, however, to have sturdier versionsready for safety trials in humans shortly.

Recruiting Cytotoxic T Cells

Different vaccine strategies are re-quired to generate activated cyto-

toxic T lymphocytes. Although surfaceproteins or even whole, killed virus par-ticles can elicit antibody production,they are poor stimulants of cellular im-munity. Cytotoxic T cells recognize shortpieces of foreign protein that appear onthe surface of an infected cell. Infectedimmune cells generate these antigenicpeptides as they digest samplings of vi-ral proteins—surface proteins such asEnv as well as the internal proteins that

drive viral reproduction and assembly.A carrier protein then escorts the pro-tein fragments to the cell membrane,where they are displayed on the outsideof the cell.

For an HIV vaccine to stimulate cell-based immunity, it must direct selectedcells to synthesize and display one ormore peptides from the proteins nor-mally made by the virus. These cellswould trick the body into mounting animmune response against all cells dis-playing the viral peptides, including onestruly invaded by HIV.

The Sabin vaccine against polio, whichconsists of a live poliovirus, turns out toevoke cytotoxic T cell activity againstpolio-infected cells, yet it does not causepolio, because the virus has been weak-ened in the laboratory by certain geneticmutations. So far, though, no mutationshave been identified that will transformHIV into a vaccine that will be com-pletely safe.

Investigators are, however, develop-ing other methods for inducing cells toproduce and display HIV proteins. Oneapproach, construction of a so-calledlive vector vaccine, takes advantage of

HIV Vaccines: Prospects and Challenges Scientific American July 1998 101

Vaccine Constituents

Viral surface proteins,such as gp120

Whole, killed HIV

Pseudovirions(artificial viruses)

Live vector viruses(non-HIV viruses engineered to carry genes encoding HIV proteins)

Naked DNAcontaining one or more HIV genes

HIV peptides(protein fragments)

Combinations of elements, such as pure gp120 protein plus canarypox vector

Live, attenuated HIV

Status

In phase I and II trials, which examine safety

Not under study in humans

Close to phase I trials

In phase II trials

In phase I trials

In phase I trials

In phase II trials

Not under study in humans; being assessed in nonhuman primates

Advantages

Safe and simple to prepare

Should present HIV surface proteins in a relatively natural conformation; simple to prepare

Present HIV surface proteins in a relatively natural conformation

Makers can control amount and kinds of viral proteins produced

Simple and inexpensive to prepare

Simple to prepare

Should stimulate both arms of the immune responseat once

Most closely mimics HIV; may interfere with infectious HIV’s ability to replicate

Disadvantages

Vaccine-elicited antibodies have failed to recognize HIV from patients

Slight risk that preparations might include some active virus; inactivated virus might shed its proteins and become ineffective

Difficult to produce and to ensure long-term stability

Complicated to prepare; current vaccines elicit modest immune response

Some worry that integration of HIV genes into human cells could harm patients

Do not elicit strong immune response

Complicated to prepare

Virus could potentially cause disease

Vaccines Eliciting Anti-HIV Antibodies

Vaccines Eliciting Cellular Responses

Vaccines Eliciting Antibody and Cellular Responses

Vaccine Strategies under Study

SLIM

FIL

MS


the ability of different viruses to invadecells. Researchers insert selected HIVgenes into a virus that is not harmful andthen allow the benign virus, or vector,to deliver the DNA to cells in the body.Because genes are the blueprints for pro-teins, the infected cells will produce HIVproteins. These viral proteins are thenchopped and shipped to the cell surface,where they can attract the attention ofwandering cytotoxic T lymphocytes.The T cells, in turn, should multiply inresponse to the antigenic stimulationand stand ready to kill any cells that ac-tually become infected with HIV.

The most extensively tested live vectorvaccines are based on the canarypoxvirus. This nonpathogenic relative of thesmallpox virus enters human cells but isincapable of assembling new viral parti-cles. Researchers have engineered ca-narypox viruses to deliver the genesthat direct the production of Env andgp120 and a variety of nonsurface HIVproteins, such as Gag (the core protein)and protease.

To date, the canarypox vaccines testedin humans have proved safe and haveelicited modest cytotoxic T cell–basedimmune responses. To stimulate a morevigorous immune response, researchersare developing viruses that will producegreater quantities or varieties of HIVproteins inside infected cells. Admin-

istering multiple doses of these vaccinesmay help generate and maintain highnumbers of activated cytotoxic T cells.

Other researchers are looking into ad-ministering viral peptides—fragments ofviral proteins—to induce an immune re-sponse. Because antigenic peptides de-rived from viral proteins activate cyto-toxic T lymphocytes, perhaps peptideswould work as a vaccine. Unfortunately,peptides by themselves do not elicit astrong immune response, cellular or an-tibody-based, in humans. The peptidesmay be degraded before they reach thetarget cells, or they may not be present-ed efficiently by the cells that encounterthem. Peptide vaccines may benefit fromthe development of better adjuvants,materials delivered along with a vaccinethat induce the immune system to re-spond more strongly.

A rather novel approach to eliciting acellular immune response involves in-jecting “naked” HIV DNA—genetic ma-terial with no proteins or lipids to deliveror protect it. At one time, scientists be-lieved that naked DNA would be de-graded too rapidly to be effective as avaccine. In reality, the DNA does get intocells and can direct the production ofviral proteins. In animal studies in miceand nonhuman primates, DNA vaccineshave successfully generated cytotoxic Tlymphocytes that recognize HIV pro-

teins. In some but not all experiments,the DNA vaccine protected animals fromsubsequent infection with HIV. Furtherstudies in animals and humans are eval-uating the safety and effectiveness ofthis approach.

Combination Strategies

The most effective strategies—andthe ones that are furthest along in

human testing—incorporate elementsthat will stimulate both arms of the im-mune response. For example, a patientmight receive a vaccine containing a ca-narypox virus carrying the Env gene tostimulate cellular immunity. Months lat-er the same patient might receive puregp120 to elicit the generation of anti-bodies. This combination strategy iscalled a prime boost, because the canary-pox vector primes the cytotoxic T cells,and the gp120 protein then strength-ens, or boosts, the immune response byeliciting antibody production.

Early trials have demonstrated thathumans vaccinated using such a combi-nation strategy develop both humoraland cellular immunity. But the antibodiesgenerated have been against laboratory-adapted HIV strains, and the cytotoxicT cell response has not been strong. Thenext generation of combination vaccineswill use canarypox viruses that carry

HIV Vaccines: Prospects and Challenges 102 Scientific American July 1998

a

b c

d

e

POX-INFECTED CELL

HIVFRAGMENTS

HIVPROTEIN

CELL BECOMES PRIMED TO ATTACK HIV-INFECTED CELLS

CD8 T CELL

CANARYPOX

HIV GENE

RECOMBINANT CANARYPOX VACCINE is among thosebeing studied as a way to elicit cell-based immunity against HIV.Such vaccines deliver HIV genes to human cells (a). The viralgenes are translated into proteins (b), which are subsequently di-

gested into fragments (c) and displayed on the cell surface (d ).These fragments stimulate HIV-specific cytotoxic, or CD8, Tlymphocytes, thereby priming them to kill any cells that may ac-tually be infected with HIV (e).

TOM

O N

AR

ASH

IMA


more HIV genes capable of producinggreater quantities of viral protein, andthe boost may contain gp120 proteinsmade from HIV isolated from patients.Such vaccines are being produced andmay soon be ready for testing in humans.

Many researchers also continue tolook into developing a live, attenuatedHIV vaccine. Because such a vaccinewould closely mimic active HIV, itshould theoretically be effective at in-ducing cellular immunity, antibody-based immunity and perhaps other un-known modes of protection. By system-atically deleting genes critical for HIVreplication, scientists hope to develop avariant of the virus that can elicit astrong immune response without givingrise to AIDS.

Recently a group of physicians volun-teered to participate in the first clinicaltrial of a live, attenuated HIV vaccine.Such a protocol would allow researchersto monitor the volunteers’ immune re-sponses and study the long-term safetyof the vaccine. The physician volunteersbelieve the value of testing this approachoutweighs the potential risks to theirhealth. Their plan remains highly con-troversial, and we and many other re-searchers feel that attenuated HIV vi-ruses should be more fully investigatedin nonhuman primates before anymovement into human trials.

Monkeys and AIDS

Vaccines based on a live, attenuatedsimian immunodeficiency virus

(SIV)—a relative of HIV that infectsmonkeys—have been tested in macaquesand other nonhuman primates. Mon-keys infected with pathogenic strains ofSIV will develop an AIDS-like syndrome.By studying this monkey model, scien-tists are able to test live, attenuated vac-cines for their safety and their ability toprotect animals when they are challengedby subsequent exposure to pathogenicstrains of SIV. Several different attenuat-ed SIV vaccines have proved remarkablyeffective at suppressing the growth of awild-type virus.

The basis of this immunity in mac-aques is unclear: animals that are effec-tively protected from SIV challenge donot necessarily have high levels of neu-tralizing antibodies or activated cytotox-ic T lymphocytes. The protective effectsmay be a consequence of some combina-tion of antibody, helper T cell and cyto-toxic T cell activity, or the effects mayderive from other aspects of immunity.

Further work is needed to determineexactly how the SIV vaccines manageto confer protection.

Although initial studies suggested ahigh degree of safety for the live, atten-uated SIV, extended and expanded safetystudies are beginning to show increasednumbers of vaccinated animals progress-ing to AIDS-like syndromes, even in theabsence of exposure to wild-type virus.

The studies are now starting to look ata greater number of animals, but the re-sults suggest that live, attenuated vac-cines may not provide full, long-termimmunity and may even cause disease.The findings also imply that investiga-tors should proceed with caution be-fore testing such vaccines in humans.

Prognosis

If the immune system in HIV-infectedindividuals cannot wipe out the virus,

why should a vaccine that activates thesame immune responses be expected toblock infection? Vaccines may give thebody an immunological “head start” bypriming the immune system to attack

HIV as soon as it appears, rather thantaking time to initiate a defense fromscratch. By doing so, vaccine-inducedimmunity may succeed in containing thevirus where the naturally infected bodydoes not.

At present, however, there is no proofthat vaccination against HIV is possible,because no protective vaccine candidatehas yet moved into Phase III trials, whichare large-scale tests designed to evaluateeffectiveness in humans. In addition, thewide genetic variability of HIV may re-duce the utility of any vaccine under de-velopment, because HIV strains isolat-ed from patients in different parts of theworld have distinctly different structuresin their Env and, to a lesser extent, otherproteins. Whether these differences, oradditional ones we have yet to appreci-ate, will significantly hamper vaccinedevelopment remains uncertain.

But there is hope. As the pathogenesisof HIV infection has become better un-derstood, investigators have realizedthat if the virus can be kept at low con-centrations in the blood, an infected per-son may never progress to AIDS. Thisinsight is encouraging because it suggeststhat even a partially effective vaccinecould be valuable in limiting the amountof virus in patients, thus potentially re-ducing their infectiousness and the symp-toms they suffer.

It is unlikely that we will develop avaccine suitable for wide-scale use inhumans within the next five years. Evenif the prime-boost combination ap-proach appears to stimulate cellular im-munity and generate good broad-spec-trum antibodies, large clinical trials willstill be needed to demonstrate its value.Those trials alone will take several years.In the meantime, researchers will contin-ue to pursue every approach that mighthelp the immune system combat HIV.

HIV Vaccines: Prospects and Challenges Scientific American July 1998 103

RHESUS MACAQUE is a type of mon-key being examined in vaccine studies.Animals receiving a live, attenuated simi-an version of HIV have been able to limitsubsequent infection by the natural mon-key virus. But in a worrisome finding, thevaccine itself has eventually caused dis-ease in some monkeys.

The Authors

DAVID BALTIMORE and CAROLEHEILMAN work together on the NationalInstitutes of Health’s AIDS Vaccine Re-search Committee, a group charged with re-assessing the priorities of the vaccine initia-tive and identifying new and innovative ar-eas of vaccine research. Baltimore, chairmanof the committee since its formation in1996, is president of the California Instituteof Technology. Heilman, a microbiologist, isdeputy director of the Division of AIDS atthe National Institute of Allergy and Infec-tious Diseases in Bethesda, Md.

SA

FRED

WH

ITEH

EAD

Ani

mal

s A

nim

als


hiv vaccines: prospects and challenges

Documents