CTL and the Controlof HIV-1 Replication
Otto Yang, M.D.
Or
Lessons learned from mixing HIV-1 and CTL
AIDS VACCINE 2007
Seattle, Washington
Evidence that CTL are protective in HIV-1 infection:
• Temporal associations with reduction in acute viremia and AIDS
• SIV-macaque model: CD8+ T cell depletion leads to sharply increased viremia, and recovery with decreased viremia
• HIV-1 sequence evolution in CTL epitopes in vivo: selective pressure (escape)
New technologies offer amazingly precise characterizations of CTL
• ELISpot: rapid, robust, validated measurements using few cells
• ICS: ability to examine phenotypes and functional markers in remarkable detail
• Such assays tell us about CTL: how many, what specificity, and what functional potentials
• How do these correlate to antiviral activity? These assays offer a detailed picture of the effector end of the CTL interaction, but what about CTL-HIV-1 interaction and the upstream events in infected cells?
Questions about the antiviral activity of HIV-1-specific CTL
• Do they kill HIV-1-infected cells?• Do they suppress HIV-1 replication (via
what mechanisms?), and what factors influence efficiency?
• Why do they fail in vivo, and how can a vaccine prevent that failure?
T Cell Receptor (On CTL)
T Cell Receptor (On CTL)
CTL recognize cells displaying foreign protein epitopes
Peptide-HLA Complex(On Infected Cell)
CTLRelease of
Cytolytic GranulesAnd
Signal for Apoptosis,Cytokines
Target Cell
HIV-1 replicates rapidly, posing a challenge to CTL
CD4+T Cell
Initial Infection
CD4+T Cell
First VirionProduced
~1 Day
Cell Death(Last Virion Produced)
~1 Day
Killing Would
Result in:Prevention of
Virion Production
Reduction ofVirion
Production
HIV-1-infected cells early in viral replication are susceptible to CTL
010203040506070
1 2 3 4Day Post-Infection
0
5
10
15
20
HIV
Tite
r (1
0^6/
ml)
Susceptibility to CTLViral Titer
O.O. Yang and R.P. Johnson et al, 1996, J Virol, 70:5799
HIV-1-specific CTL can potently inhibit viral replication
Infect CD4+T Lymphocytes
With HIV-1Culture
With CTL
CultureAlone
0.1
1
10
100
1000
10000
3 5 7 10Days
H9-B14H9-B14 + EnvCTL(B14)
Adapted from O.O. Yang and B.D. Walker et al. J. Virol 1997, 71:3120
Suppression of HIV-1 by CTLcan be irreversible
0.1
1
10
100
1000
10000
3 5 7 10 12 14 17 19 21 23 25Days Post Infection
Infected Cells Infected Cells + CTL
Remove CTL
Adapted from O.O. Yang and B.D. Walker et al. J. Virol 1997, 71:3120
CTL can suppress HIV-1 via cytokines and direct cytolysis
1
10
100
1000
10000
100000
1000000
0 4 6 8 11Day after Infection
H9H9+CTL(B14)H9-B14 + CTL(B14)H9+H9-B14+CTL(B14)
SolubleSoluble
+Lytic
Adapted from O.O. Yang and B.D. Walker et al. J. Virol 1997, 71:3120
CTL recognizing different epitopes can vary in their antiviral efficiency
-1
0
1
2
3
4
5L
og U
nits
Sup
pres
sion Control
(mismatch)RT (IV9-A2)
Gag (SL9-A2)
Nef (KL9-B60)
RT Gag NefNote that T.J. Tsomides and H.N. Eisen et al (JEM 1994, 180:1294) showed that IV9 is presented ~12/cell and SL9 is presented ~400/cell.
Unpublished data
Both Nef-mediated HLA-I downregulation and epitope kinetics
affect CTL antiviral activity
0
1
2
3
4
5
Log
Uni
ts S
uppr
essi
on
Late SL9 (in Gag)/WT NefLate SL9 (in Gag)/Defective NeEarly SL9 (in Nef)/Defective N EARLY +
NO Nef effect
LATE +NO Nef effect
LATE +Nef effect
Adapted from A. Ali and O.O. Yang et al. J. Virol 2004, 78:561
WT Nef
0
1
2
3
4
5
6
7
2 4 6Day Post-Infection
The impact of Nef on CTL antiviral activity can be quantitated
Adapted from S. Adnan and O.O. Yang et al. Blood 2006, 108:3414
M20A Nef
0
1
2
3
4
5
6
7
2 4 6Day Post-Infection
6.3 (no CTL)
5.5 (+CTL)
6.3 (no CTL)
0.9 (+CTL)
Nef Impact Ratio= (6.3-5.5) ÷ (6.3 -0.9) = 0.15
WT Nef
0
1
2
3
4
5
6
7
2 4 6Day Post-Infection
The impact of Nef on CTL antiviral activity can vary dramatically
Adapted from S. Adnan and O.O. Yang et al. Blood 2006, 108:3414
M20A Nef
0
1
2
3
4
5
6
7
2 4 6Day Post-Infection
6.3 (no CTL)
3.2 (+CTL)
6.3 (no CTL)
2.6 (+CTL)
Nef Impact Ratio= (6.3-3.2) ÷ (6.3 -2.6) = 0.84
Nef does not affect the antiviral activity of HLA-I C-restricted CTL
0.0
0.2
0.4
0.6
0.8
1.0
1.2N
ef Im
pact
Rat
io
B
C
A
ALL A/B ALL CAdapted from S. Adnan and O.O. Yang et al. Blood 2006, 108:3414
Other epitope factors may also affect Nef susceptibility (?)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Nef
Impa
ct R
atio
ALLEARLY(A/B)
ALLLATE(A/B)
(Unpublished data)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Nef
Impa
ct R
atio
B*07
B*57
A*02
ALLA*02
ALLB*07
ALLB*57
GagPolEnv
NefRev
HIV-1-specific CTL lose antiviral function with senescence
0.0
0.5
1.0
1.5
2.0
Log
Uni
ts S
uppr
essi
on
Control hTERT
Adapted from M. Dagarag and O.O. Yang et al. J Virol 2003, 77:3077
Control hTERT
Perforin
Gra
nzym
e B
From O.O. Yang and C. Uittenbogaart et al. Virol 2004, 332:16
Summary of factors determining CTL antiviral efficiency
• Epitope levels (adequacy for recognition)• Epitope kinetics (timing of recognition in
relation to viral production)• Impact of Nef-mediated HLA-I
downregulation• Effector status of CTL• What is the relative importance of these
factors???
0 10 20 30 40 50 60 70 80 90
RPAEPVPLQL
-S--------
--T-------
--E-------
---Q------
--------H-
-------F--
--V-------
--E----F--
-------FH-
--Q-----L-
Epi
tope
Seq
uenc
e
CTL killing of peptide-loaded versus HIV-infected cell
Peptide 10µg/mlInfected C
An example that antiviral assays can give different results than standard CTL assays
(Implications for escape and cross-clade studies?)
0 10 20 30 40 50 60 70 80 90
RPAEPVPLQL
-S--------
--T-------
--E-------
---Q------
--------H-
-------F--
--V-------
--E----F--
-------FH-
--Q-----L-
Epi
tope
Seq
uenc
e
CTL killing of peptide-loaded versus HIV-infected cell
Peptide 10µg/mlInfected C
Clade A1 consensusClade A2 and F2 consensus
Clade F1 consensus
MS Bennett and OO Yang, Submitted
CTL detection does not predict virus suppression
1
2
3
4
5
6
3 4 5 6 7Days
Control +CTL
RPAEPVPLQL(Consensus B/C/G)
1
2
3
4
5
6
3 4 5 6 7Days
Control +CTL
1
2
3
4
5
6
3 4 5 6 7Days
Control +CTL
1
2
3
4
5
6
3 4 5 6 7Days
Control +CTL
-S--------(Consensus A1)
--T-------(Consensus A2/F2)
--E-------(Consensus F1)
MS Bennett and OO Yang, Submitted
Implications for analysis of current vaccine trials?
• ELISpot is the major validated surrogate marker for CTL immunogenicity of vaccines
• Most likely, ELISpot responses are necessary but not sufficient to reflect the antiviral activity of CTL (recognition of infected cells)
• If current vaccine trials fail to demonstrate clinical benefit despite “immunogenicity,” it may be premature to discount a CTL-based vaccine approach based on lack of protection in vaccine “responders”
Multiple mechanisms have been observed and proposed to explain failure in vivo
• Abnormal differentiation and activation
• Exhaustion/senescence• Hypofunctional profile• Lack of CD4 help• Epitope escape mutation
Can a single process explain these diverse events (Ockham’s razor)???
10
100
1000
10000
100000
1000000
0 5 10
Day
HIV (SLYNAVATL) + CTL
Single amino acid changes in epitopes can allow HIV-1 to completely avoid CTL
10
100
1000
10000
100000
1000000
10000000
0 5 10
Day
HIV (SLYNTVATL) + CTL
HIV-1Infection
Immunodominance of Poorly Conserved
Epitopes
A hypothesis….Early Escape andPoor Containment
Early Massive Irreversible CD4+
T Cell Loss
Original Antigenic
Sin
Decreasing CTL Quantity and Quality
Eventual Complete Failure
Abnormal CTL Activation and Differentiation
Ongoing Escape
Ongoing HIV-1 Replication
HIV-1Infection
Immunodominance of Poorly Conserved
Epitopes
A hypothesis….
Early Reduction of HIV-1
Preserved CD4+ T
Helper Cells
Effective CTL
Containment of HIV-1
Targeting of Conserved Epitopes
Vaccine Priming
What strategies can prevent escape?
All randomly generated epitope mutants
Viable mutants
Unrecognizable mutantsEscape
Mutants
A mutation must be both non-recognized and viable to allow escape
All randomly generated epitope mutants
Viable mutants
Unrecognizable mutants
NoEscapeMutants
Bypassing immunodominance to focus CTL on very constrained epitopes?
All randomly generated epitope mutants
Viable mutants
EscapeMutants
Viable mutants
Unrecognizable mutants
Focus immunity on constrained sequences?
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
From entropy calculations generously provided by B Korber