Direct identification of HLA‐presented CD8 T cell epitopes from transmitted founder HIV‐1 variants

Cytotoxic T lymphocytes (CTLs) are a critical arm of the immune response to viral infections. The activation and expansion of antigen specific CTL requires recognition of peptide antigens presented on class I major histocompatibility complex molecules (MHC‐1) of infected cells. Methods to identify presented peptide antigens that do not rely on the pre‐existence of antigen specific CTL are critical to the development of new vaccines. We infected activated CD4+ T cells with two HIV‐1 transmitted founder (TF) isolates and used high‐resolution mass spectrometry (MS) to identify HIV peptides bound on MHC‐1. Using this approach, we identified 14 MHC‐1 bound peptides from across the two TF isolates. Assessment of predicted binding thresholds revealed good association of the identified peptides to the shared HLA alleles between the HIV+ donors and the naïve PBMC sample with three peptides identified through peptide sequencing inducing a CD8 T‐cell response (p < 0.05). Direct infection of naïve CD4 cells by HIV TF isolates and sequencing of MHC‐I presented peptides by HPLC‐MS/MS enables identification of novel peptides that may be missed by alternative epitope mapping strategies and can provide valuable insight in to the first peptides presented by an HIV‐infected CD4 cell in the first few days post infection.

computational analysis of HIV sequences or ELISpot based epitope mapping [6]. However, computational analysis does not account for the complexity of the antigen processing and presentation pathways, while ELISpot approaches can only identify already known epitopes for which there are pre-existing cognate CD8 T cells. Furthermore, these strategies are not designed to identify epitopes that can be found in HIV-1 TF variants.
A critical component of inducing an effective adaptive, CD8 T-cell immune response is the presentation of 'self' peptides in the context of major histocompatibility complex (MHC) class I molecules. The down regulation of peptide presentation is a key strategy by which viruses avoid immune detection and a number of HIV-1 gene products have been implicated to interfere with MHC class I antigen presentation with the strongest evidence favouring roles for HIV-1 Nef and Tat in this process [7].
The identification of HIV epitopes recognized by the immune system has historically been achieved through ELISpot epitope mapping strategies. For these strategies, synthetic peptide libraries are required, which are limited by their predicted or idealized designed epitopes and the cost, a major factor in particular if an autologous peptide strategy is required. This can in part be overcome by using consensus peptides or cohort-specific potential T-cell epitope pools [8]; however, both of these strategies rely on algorithms that mine only existing epitopes. Further drawbacks for both autologous and predictive peptides are that neither is subject to the post-translational modifications that can occur with in vivo peptide presentation [9,10].
Direct identification of peptides presented by HLA class I and II molecules associated peptides by LC-MS/MS can be an invaluable tool both for enabling T-cell epitope discovery, in addition to informing vaccine design and delivery strategies [11]. The use of LC-MS/MS enables the direct characterization and quantitation of HLA-restricted peptides derived from both HLA Class I and Class II MHC complexes [12]. This technique has been used to characterize the HLA class I-associated immunopeptidome in C1866 cell line and primary human CD4 cells infected with HIV-1 for identification of viral presented antigens that were broadly recognized by HLA-matched HIV-1-infected individuals [13]. Furthermore, strategies to distinguish co-precipitating proteins from HLA-bound ligands were established in the context of HIV-1 [14] and identification of viral peptides containing a protein splice junction was reported [15]. It was recently shown that detection of HIV-1 viral peptides early in infection (6 h [11,16]. Env-specific peptides derived from HIV-infected cells were recently observed in antigen presenting cells [17].
Utilizing a sequential elution step with anti-MHC Class I antibodies, it is possible to create a fractionation of MHC Class I associated peptides that can then be processed for identification using LC-MS/MS. Additional resolution of the MHC class can also be achieved using monoclonal antibodies specific for each subclass. Moreover, the flow through from these affinity columns can be used to detect and quantify

Statement of significance
The research described here has implications for understanding the early kinetics of viral immunity by utilizing HPLC-MS/MS for direct sequencing from cellular HLA complexes for CD8 epitopes. The ability to model early viral infections in naïve PBMC samples with specific HLA profiles will enable a more detailed evaluation of the relationship of HLA haplotypes to epitope presentation in acute viral infections. the amount of expressed HIV antigens as described recently for Vaccinia virus-infected dendritic cells [18,19].
In this study, we set out to identify novel HIV-1 epitopes presented on MHC-I from TF HIV-1 variants using HPLC-MS/MS sequencing from an HIV negative, HLA-matched PBMC donor, and confirm immunogenicity of the identified peptides using ELISPOT analysis with PBMCs isolated from the HIV patients from which the founder viruses were derived ( Figure 1). Use of TF variants that have not undergone immune evasion mutations and that possess traits that confer increased ability to be transmitted at the population level would yield vaccine relevant epitopes.

Kinetics of CD4+ T cell infection by primary HIV-1 isolates
For optimal detection of MHC bound peptide by LC-MS/MS sequencing, it is imperative to ensure that a large percentage of cells are productively infected [12].  Figure S1).  Figure 2A, p < 0.05), which reflects intrinsic differences in the viral replicative capacity of these two viruses [20].
Concurrent assessment by p24 intracellular flow cytometry demonstrated that 39.1% and 51.5% of both R880F and R463F infected cells, respectively, were alive, CD3+, CD4+ and p24+ ( Figure 2B and Figure   S1) indicating productive viral infection at sufficient frequency to proceed with peptide sequencing.

Validation of identified epitopes by ELISpot
To further understand the relationship between the peptides identi-  Table 2). Three Among the peptides identified by sequencing that elicited responses, two had been previously identified by mapping strategies (GQWVHQNF and RVMGTQMNY) [20]. Peptide SQVHGTNIM, predicted to be associated to B*15:03 and C*02:10, was not identified by either autologous epitope mapping or listed on the LANL database and so may be a newly identified CD8 T-cell epitope.  [13]. Although the current methodologies are reagent and cell intensive, the peptides identified by directly sequencing from acutely infected samples offer an insight into the early epitopes presented to the immune system following HIV-1 infection. Furthermore, the unbiased nature through which these peptides are identified allows for the discovery of novel epitopes whose functional properties can then be analysed.

DISCUSSION
A further advantage is that this method does not require the use  [20]. This association between both in vivo and in vitro viral replication kinetics [20] with the number of epitopes that we identified through a HPLC-MS/MS approach further suggests that epitopes identified may be physiologically relevant. The finding that more epitopes were identified from the rapid progressor whose TF variant grows to higher titres in vitro suggests a role for viral replica-     PBS and pelleted at 300 × g for 10 min and resuspended in 5 mL of 2× lysis buffer as indicated above for >1 × 10 7 cells and pipetted gently up and down until the lysate has a homogenous appearance (greater than five times).

Peptide elution and MS
MHC-peptide elution and MS to identify HIV specific peptides was conducted as previously described [12]. Briefly, lysates of infected cells were cleared by two subsequent centrifugation steps at 500 × g  threshold was defined by decoy database searches implemented in the regarding search engines at a general false discovery rate of 5% [13].
All identified peptides provided as supplementary file. Full data set for HIV and self-peptides available at dataspace.iavi.org.

ELISpot
A single PBMC vial for volunteer 170038 (514 days post EDI) and 175042 (510 days post EDI) was assessed by Human IFN-γ 96 well ELISPOT kit (Cellular Technology Limited, Cat# hIFNg-1M/2) as per manufacturer's instructions. Responses were assessed by Student t- test for difference to mock tested wells (p < 0.05) [23].