Transcript - Murdoch Research Repository

SCREENING FOR HIV RESPONSES USING OPTIMAL EPITOPES
PREDICTED BY HLA-VIRAL SEQUENCE POLYMORPHISM
ASSOCIATIONS.
Western Australian
HIV Cohort Study
Roberts SG1, Almeida CM1, Bronke C1, Ahmad I1, Al Damuk A1,Cooper D1, Corkery M1, Keane N1,
3
1
1,2
1,2
Heckerman D , Chopra A , Mallal S , John M
1Centre
for Clinical Immunology and Biomedical Statistics, Institute of Immunology and Infectious Diseases, Murdoch
University, Perth, Western Australia, 2Department of Clinical Immunology and Immunogenetics, Royal Perth Hospital, Perth,
Western Australia, 3Microsoft Research, Microsoft Inc, Redmond, USA
HIV escapes immune recognition by mutating critical amino
acid residues in known HLA-restricted epitopes. Such changes
may limit the ability of ex-vivo assays to detect and map the
epitopes subject to such selection in-vivo. We used HLAassociated polymorphisms in HIV generated from an analysis of
a combined cohort of 800 anti-retroviral naïve, HLA-diverse,
predominantly subtype B-infected individuals from the US and
Western Australia to predict and map HIV-specific T cell
responses(see figure 1). A novel epitope prediction program
‘Epi-pred’ was used to predict optimal length epitopes around
sites of HLA-allele specific polymorphism and these were tested
in IFN-γ ELISpot assays, taking into account the autologous
HLA genotype and the autologous viral sequence of 200 US
cohort individuals. The ELISpot assay was optimised and
automated for high-throughput testing of multiple HLAcustomised plates1 (see figure 2).
Figure 1.
Method used to
select patients
peptides to
perform
ELISpots on.
HIV sequences
(n=800)
Statistical
analysis
3A
HLA-B*2705
Nef
Protein
Start
position
HLA
Epitope
Env
209
A*0101
SFEPIPSHY
Env
310
Positive
responses
(n)
1
Gag
406
Nef
9
Nef
178
B*1801
KEVLVWKF
1
Pol
8
Multiple
FPQGKAREF
1
Pol
901
B*2705
KRKGGIGGY
1
A*0101/Cw*0401 GPGPGRAFY
A*0302
1
RAPRKKGC
WK
1
A*0101/A*0302 SVVGWPAVR
1
Table 1. Seven novel epitopes were seen in the 29
patients tested.
Non690
SFU/106 adapted
PBMCs epitope
KRKGGIGGY
SAG-----3’
All associations/epitopes classified as individual cellular “hypotheses” to test with PBMCs.
• CD8+ T-cell epitopes (8 to 11-mer) predicted by ‘Epi-pred’
• Known epitopes from LANL
5’----- EKI
RLRPGGKKKY
KLK -----3’
Pollyallelic peptide panel
Known epitopes/novel escape
variants: Medium priority in
screening assays
Test non-adapted and adapted epitopes.
178-185
HLA-B*1801
5’-----DPE
No difference in
responses
10
Adapted
SFU/106 epitope
PBMCs
560
SFU/106
PBMCs
0
SFU/106
PBMCs
2800
SFU/106
PBMCs
KEVLMWKF
Env 209-217
Non-adapted
epitope
Adapted
epitope
HLA-A*0301
1900
NonSFU/106 adapted
PBMCs epitope
KEVLVWKF
HLA-B*1501
RLRPGGRKKY
Novel epitopes/escape variants:
High priority in screening
assays
Non-classical escape
DSR-----3’
KRKGGIGEY
Gag 20-29
3C
Tat 42-50
5’-----HNF
Consensus sequence scanned with ‘Epi-pred’ with non-adapted and
adapted amino acids substituted at sites of HLA association
3B
Classical escape
Pol 901-909
874 HLA
associations
Known epitopes/escape
variants: Low priority in
screening assays
Novel epitopes
Figure 3. Responses to adapted and non-adapted epitopes detected in ELISpot assays. (3A) Classical
escape – the adapted epitope elicits a lower response compared to the non-adapted epitope. (3B) Nonclassical escape – the adapted epitope elicits a higher response than the non-adapted epitope. (3C) No
difference in responses seen.
ACTG (n=555) +
WA cohort (n=245)
HLA-alleles
(n=800)
On average, 11 epitopes were
tested for each patient and of
these 18% elicited an IFN-γ
response.
In an analysis of
the first 29 individuals tested
in this system, seven putative
novel epitopes were detected
(see table 1).
In many
instances, the HLA-driven
change led to loss of reactivity
as predicted for classical CD8+
T-cell escape, however more
complex patterns of reactivity
were seen, particularly in Nef
epitopes (see figure 3).
Adapte
d
epitope*
5’-----ITK
GLGISYGRK
KRR-----3’
840 SFU/106
PBMCs
Non-adapted
epitope
900 SFU/106
PBMCs
Adapted
epitope
HLA-A*0101
5’----- PKV
SFEPIPSHY
CAP -----3’
SFEPIPSIY
20
SFU/106
PBMCs
Nonadapted
epitope
740
SFU/106
PBMCs
Adapted
epitope
GLGISYGRR
HIV mutates residues that affect HLA binding , TCR
recognition or intracellular processing allowing the virus to
escape detection by the host immune system (see figure 4 & 5).
If escape is extra-epitopic, then test non-adapted
Nef 71-81 and HLA-B*3501
ELISpot assay
Nef 90-97 and 83-91 and HLA-B*0801
5’-----YKGALDLSHFLKEKGGLEGL-----3’
Coat with INF-g
capture Ab
Figure 2. Overview of the ELISpot
method. Each of the steps illustrated
can be set up as a separate method
using the Biomek FX.
RPQVPLRPMTF
Add peptide & cells
Add
biotinylated
detection Ab
570
PBMCs
200 SFU/106
PBMCs
900 SFU/106 PBMCs
Adapted
epitope
“imputed”
FLKEMGGL
2500 SFU/106
PBMCs
Adapted
epitope
“imputed”
FLKENGGL
420 SFU/106
PBMCs
Adapted
epitope
“imputed”
FLKEQGGL
1540 SFU/106
PBMCs
Adapted
epitope
“imputed”
Non-adapted
epitope
Adapted
epitope
Figure 4. Amino acid change in anchor
position
Add TMB
Acknowledgements
Study population was drawn from ACTG 5142/5128, Beckman Coulter,
“Bill & Melinda Gates Foundation”, National Institutes of Health, National
Health & Medical Research Council and CCIBS staff
[email protected]
[email protected]
http://www.ccibs.org
Non-adapted
epitope
FLKEEGGL
5’-----YKGALDLSHFLKE(E/M/N/Q)GGL
EGL-----3’
Add
streptavidin
enzyme
Accredited for compliance with ISO/IEC 17025 interpreted
for research using CITAC Guide CG2, for HLA sequence
based typing, viral sequencing and ELISpot analysis.
Accreditation number 15785
RPQVPLRPMTY
SFU/106
3540 SFU/106
PBMCs
720 SFU/106 PBMCs Neo-epitope
Figure 5. Mutations affecting TCR recognition and
intracellular processing.
The screening strategy was based on a genetics directed
approach, in which specific sites and epitopes were tested
based on their in-vivo polymorphism. The results have enabled
us to identify possible novel epitopes that warrant further
investigation using confirmatory assays. At a population level,
such screening takes into account the most prevalent HLA
genotypes and HLA-restricted responses in that population
efficiently. These results provide further insights into CD8+ Tcell responses against HIV and have implications for HIV
vaccine design.
References
1. Almeida et al. 2009. Automation of the ELISpot assay for high-throughput
detection of antigen-specific T-cell responses. J Immunol Methods; 344:1-5