Novel Nested Peptide Epitopes Recognized by CD4+ T Cells Induced by HIV-1 Conserved-Region Vaccines.

CD4+ T-cell responses play an important role in the immune control of the human immunodeficiency virus type 1 (HIV-1) infection and as such should be efficiently induced by vaccination. It follows that definition of HIV-1-derived peptides recognized by CD4+ T cells in association with HLA class II molecules will guide vaccine development. Here, we have characterized the fine specificity of CD4+ T cells elicited in human recipients of a candidate vaccine delivering conserved regions of HIV-1 proteins designated HIVconsv. The majority of these 19 most immunogenic regions contained novel epitopes, that is, epitopes not listed in the Los Alamos National Laboratory HIV Sequence Database, which were able in vitro to stimulate vaccinees’ CD4+ T cells to proliferate and produce interferon-γ and tumor necrosis factor-α. Accumulation of HLA class II epitopes will eventually accelerate development of HIV-1 prophylactic and therapeutic vaccines.


Introduction
There is no doubt that T cells are an important component of the host immune defense against HIV-1 infection [1,2]. Adaptive cellular antiviral responses are mediated by CD8 + and CD4 + T cells, which recognize foreign peptides presented by self-major histocompatibility complex molecules, in humans designated human leukocyte antigens (HLA), of class I and II, respectively. Upon stimulation, CD8 + and CD4 + T cells produce a variety of soluble intercellular signaling molecules leading to cell death and inducing an antiviral state in cells in their immediate vicinity. The convention is to regard CD8 + T cells as the effector cells killing HIV-1-infected cells and CD4 + T cells as the helper cells to both CD8 + cytolytic cells and B cells producing antibodies, however, it is well documented that CD4 + T cells can also exert cytolytic activity [3][4][5][6][7]. Both CD8 + [8][9][10][11] and CD4 + [5,6,[12][13][14][15][16] T-cell responses have been associated with HIV-1 immune control and their broadly specific coordinated actions are likely to be important for protection.
CD8 + T cells are highly selective for peptides presented by HLA class I molecules, which has allowed for a relatively accurate experimental definition of optimal peptide lengths (usually 8 to 10 amino acids) and HLA-binding motifs [17,18]. In contrast, definition of common patterns for HLA-class II-associated peptides has been more challenging [19][20][21]. HLA class II binds nested peptides of up

IFN-γ ELISPOT Assay
Cryopreserved PBMCs were tested in an IFN-γ ELISPOT assay as described previously [44]. ELISPOT plates (S5EJ044I10; Merck Millipore, Bedford, MA, USA) were prewetted for 1 min with 15 µL of 35% ethanol and coated with anti-IFN-γ antibody at 10 µg/mL in PBS (clone 1-D1K; Mabtech AB, Nacka Strand, Sweden) at 4 • C for 16 h. Next day, plates were washed with PBS and blocked with R10 at 37 • C for at least 1 h. STCLs were added at 4 × 10 4 cells/well into triplicate wells for peptide pools and duplicate wells for individual peptides at a final concentration of 1.5 µg/mL each and incubated at 37 • C, 5% CO 2 overnight. Six no-peptide wells with cells and 0.45% DMSO served as a negative control. Cells cultured with 10 µg/mL PHA (Sigma Aldrich, Poole, UK) served as a positive control. Next day, wells were incubated with biotin-conjugated anti-IFN-γ mAb followed by alkaline phosphate-conjugated streptavidin (both from Mabtech AB) and color substrate BCIP/NBT Plus (Mabtech AB). The reaction was stopped after 5 min by washing with tap water and the plates were air dried. Next day, the spots were counted by an AID ELISpot Reader with version 5.0 software (AID GmbH, Straβberg, Germany). The results were expressed as median net SFU/10 6 PBMC after subtracting the median number of spot-forming units (SFU) in no-peptide wells from the test wells. Positive responses were defined as at least 50 SFU/10 6 PBMC above background and at least twice the background of wells without peptide stimulation.
The gating strategy was employed as follows: (i) FSC-A vs. FSC-H to gate out doublets, (ii) FSC vs. SSC wide gate to exclude cell debris, (iii) CD3 vs. LIVE/DEAD to gate on viable CD3 + T-cells, (iv) CD4 vs. CD8 to gate on single-positive CD4 and CD8 T cells and (v) IFN-γ and TNF-α CD4 + subsets.

Epitope Prediction
Peptides identified by the ICS analysis of peptide pool-or individual peptide-expanded STCL and the volunteer's HLA types were entered into the Epitope Location Finder (ELF) algorithm [46] and compared to previously reported epitopes retrieved from the T-helper/CD4 + Epitope Summary [47]. Potential HLA DRB1 restriction was suggested based on the match of anchor residues and motifs associated with the submitted HLAs.
Vaccines 2020, 8, x FOR PEER REVIEW 4 of 22 4 and compared to previously reported epitopes retrieved from the T-helper/CD4 + Epitope Summary [47]. Potential HLA DRB1 restriction was suggested based on the match of anchor residues and motifs associated with the submitted HLAs.

CD4 + T-Cell Epitope Mapping
The primary HIV-CORE 002 immunological readout employed 199 15-mer peptides overlapping by 11 amino acids, which spanned the entire HIVconsv protein. These HC001-HC199 peptides were divided into six pools P1-P6 and used to determine the frequencies of HIVconsv-specific responses in vaccine recipients in a fresh ex vivo IFN-γ ELISPOT assay [44]. For mapping of stimulatory 15-mer

CD4 + T-Cell Epitope Mapping
The primary HIV-CORE 002 immunological readout employed 199 15-mer peptides overlapping by 11 amino acids, which spanned the entire HIVconsv protein. These HC001-HC199 peptides were divided into six pools P1-P6 and used to determine the frequencies of HIVconsv-specific responses in vaccine recipients in a fresh ex vivo IFN-γ ELISPOT assay [44]. For mapping of stimulatory 15-mer peptides, cryopreserved PBMCs collected between 10 weeks and 1 year after the last vaccine administration were first expanded in vitro for 10 days using individual peptide pools to establish a short-term cell line (STCL), which was then tested against each 15-mer of the original pool in an IFN-γ ELISPOT assay. To narrow down optimal peptide length and identify CD4 + T cells, single stimulatory 15-mer 'parental' peptides used to expand PBMCs for 10 days and the resulting SCTLs were incubated with progressively truncated peptides in an ICS assay and the cytokine production by CD3 + CD8 + and CD3 + CD4 + T cells was monitored. For 11 individuals showing positive CD4 + T-cell responses, optimal epitope cores were predicted in silico based on the subjects' HLA class II molecules as well as assessed experimentally. Full subjects' HLA class I types are given in Table 1, while only the potentially restricting molecules are given in the text. The results are reported below in the order of the 15-mer peptide, in which they appear in the HIVconsv immunogen, and the HIV-1 protein of origin and HXB2 amino acid location are also provided below the peptide schematics.
HC003 EVIPMFTALSEGATP (Gag 167-181) One subject 418 (DRB1*04:01 DRB1*07:01) developed CD4 + T-cell responses to the HC003 peptide with stimulatory shorter peptides EVIPMFTALSEGAT (ET14), EVIPMFTALSEGA (EA13), EVIPMFTALSEG (EG12), EVIPMFTALSE (EE11), and EVIPMFTALS (ES10) (Figure 2). Peptide PMFTALSEGAT is reported in the LANL-HSD without restriction as well as the PEVIPMFSALSEGATP peptide, which is restricted by DR1. The Epitope Location Finder (ELF) of the LANL-HSD predicts two epitope cores EVIPMFTALS and FTALSEGAT to bind DRB1*04:01, and VIPMPT and LSEGAT to bind to DRB1*07:01 with the anchor residues underlined. peptides, cryopreserved PBMCs collected between 10 weeks and 1 year after the last vaccine administration were first expanded in vitro for 10 days using individual peptide pools to establish a short-term cell line (STCL), which was then tested against each 15-mer of the original pool in an IFNγ ELISPOT assay. To narrow down optimal peptide length and identify CD4 + T cells, single stimulatory 15-mer 'parental' peptides used to expand PBMCs for 10 days and the resulting SCTLs were incubated with progressively truncated peptides in an ICS assay and the cytokine production by CD3 + CD8 + and CD3 + CD4 + T cells was monitored. For 11 individuals showing positive CD4 + Tcell responses, optimal epitope cores were predicted in silico based on the subjects' HLA class II molecules as well as assessed experimentally. Full subjects' HLA class I types are given in Table 1, while only the potentially restricting molecules are given in the text. The results are reported below in the order of the 15-mer peptide, in which they appear in the HIVconsv immunogen, and the HIV-1 protein of origin and HXB2 amino acid location are also provided below the peptide schematics.
6 Figure 3. HC017/VID 418-definition of CD4 + T-cell epitopes in HIV-1 protein Gag. 15-mer-expanded cryopreserved PBMC of volunteer 418 were stimulated with truncated peptides (left) in ICS and production of IFN-γ (green) and TNF-α (orange) by CD4 + T cells was determined. Arrows in the graph and amino acids S and K indicate the stimulatory peptide termini.
6 Figure 3. HC017/VID 418-definition of CD4 + T-cell epitopes in HIV-1 protein Gag. 15-mer-expanded cryopreserved PBMC of volunteer 418 were stimulated with truncated peptides (left) in ICS and production of IFN-γ (green) and TNF-α (orange) by CD4 + T cells was determined. Arrows in the graph and amino acids S and K indicate the stimulatory peptide termini.
HC102/VID 403-definition of CD4 + T-cell epitopes in HIV-1 protein Vif. HC102 peptide-expanded SCTL from volunteer 403 were stimulated in an ICS assay using truncated peptides (left) and CD4 + T cells producing IFN-γ (green) and TNF-α (orange) were enumerated using flow cytometry. Amino acid R indicates the stimulatory peptide termini.

Discussion
In the course of this work, we have defined novel peptides stimulatory for CD4 + T cells elicited by a candidate HIV-1 vaccine in healthy, HIV-1/2-negative volunteers in Oxford, UK. This vaccine delivered the first-generation conserved regions of HIV-1 proteins employing clade consensus amino acid sequences, which were assembled into a chimeric immunogen designated HIVconsv [40]. Both the use of conserved regions frequently subdominant in natural infection and utilization of potentially artificial consensus sequences might have contributed to the relatively high discovery rate of novel, previously unreported epitopes absent from the LANL-HSD. In contrast, reported CD4 + T-cell epitopes reported in Gag p24 EEKAFSPEV (Gag 160-168) and EKAFSPEVIPMFSAL (Gag 161-175) restricted by DRB1*01:01 were present in the HIVconsv immunogen (peptide HC001), but were not immunogenic even in volunteers expressing the correct restricting DRB1 allele. This may be due to the epitope location at the very N-terminus of the HIVconsv protein, only preceded by the translational start methionine, which might have interfered with peptide processing or led to epitope destruction [48][49][50].
Our results concur with previous studies in that each STCL expanded in vitro by a 10-day culture with a 15-mer peptide recognized one or multiple clusters of shorter peptides likely each centered around a peptide core, the interaction of which with the HLA class II groove is usually characterized by a motif of anchor residues [5,6,31,51,52]. However, it remained challenging to define the 'optimal' binding/stimulatory peptide length as the strongest stimulatory peptides were often 11 or 12 residues long, yet sometimes shorter versions clearly induced robust reactivities, too. The 11/12-amino acid lengths may be optimal by the virtue of the flanking region sequences for both binding to the restricting HLA class II molecules and interacting with the T-cell receptors on these CD4 + cells as documented by the near 60 MHC class II structures. Despite this increasing amount of structural information, there is no simple solution to the 'optimal epitope length' conundrum. While optimal peptide length may exist for a combination of defined HLA class II allele with a particular T-cell receptor, it is not likely to be the minimal stimulatory peptide length. Thus, a class II epitope may be best described collectively by definition of the longest and shortest (may be shorter than core) peptide length of nested stimulatory peptides, identification of the binding epitope core guided by a binding motif and determination of the strongest stimulatory peptide length. Although much more stringent, similar plasticity is emerging for the definition of HLA class I epitopes [17], but it is likely much more pronounced for class II epitopes given the open ends of the peptide-binding cleft. Bearing in mind that all stimulatory lengths may contribute to and some modulate negatively the immune response [33][34][35][36][37][38], these parameters may vary depending on the readout: production of IFN-γ, TNF-α, IL-2, or cytotoxicity, and so forth, and that multiple epitopes may be overlapping. For these reasons, in Table 2, we listed all tested peptides stimulatory for at least 1% of expanded CD4 + T-cell lines and indicated the relative balance in the production of the two measured cytokines.
HIV-1-specific CD4 + T cells display multiple functions [5,6,31,51]. In the present study, CD4 + T cells were capable of proliferation and production of at least IFN-γ and TNF-α, two important antiviral and proinflammatory cytokines. Bearing in mind in vitro expansion, these two functions were not always performed by equal frequencies of responding cells. What governs these quantitative and qualitative differences remains unclear, but the presentation of the HIVconsv immunogen to the immune system, here by the combination of intramuscular DNA, nonreplicating simian adenovirus, and MVA delivery, likely critically impacted on the type of induced responses by analogy to different viruses or bacteria inducing different types of immune memory [1,7,53,54]. Thus, for an efficacious vaccine, at the very least the immunogens and means of their delivery will need to be optimized to maximize induction of protective T-cell traits [1,54,55]. For HIV-1, this may also include success at inducing T-cell populations with access to B-cell follicles, the sites of persistent HIV-1 reservoir and replication [56]. As for epitope mapping, some responses may be easier/more difficult to detect by one or the other cytokine and the stimulation facilitates/complicates precise epitope mapping. Also, the interaction of the CD4 coreceptor with MHC class II greatly reduces the number of HLA-TCR engagements required for T-cell activation [57] and substantially increases cytokine production by CD4 + T cells [58], and this interaction may vary for different responses.
A previous study, which mapped comprehensively CD4 + T-cell responses across the whole proteome in HIV-1-positive individuals, identified immunodominant responses to the Gag and Nef proteins [31]. This was not the case following HIVconsv vaccination, which induced prominent Pol epitope responses directly proportional to the relative length of the Pol segment in the HIVconsv sequence (Table 3). These results would support the ability of a potent vector delivery to (re)focus immune responses including CD4 + T cells toward regions that are generally subdominant in natural HIV-1 infection [59,60]. A different study reported over half of tested variant peptides not being recognized by CD4 + T cells [51], emphasizing even for the more promiscuous HLA class II molecules peptide selectivity and the importance of focusing on and matching vaccine responses to less variable regions of the HIV-1 proteome. Finally, 11 helper T-cell peptides with HLA-DR supermotif, which were 82% conserved in clade B isolates, were identified and their stimulatory ability were confirmed in HIV-1-positive patients [21]. One of the critical CD4 + T-cell traits for achieving protection against HIV-1 is T-cell frequency. The combination of intramuscularly delivered simian adenovirus prime and MVA boost (with or without prior DNA) is on the forefront of currently tested regimens of nonreplicating vaccine vectors in humans [44,[61][62][63]. In the present work, we did not evaluate the true frequencies of CD4 + T-cell responses with individual specificities in peripheral blood, but rather focused on their specificities, which was greatly facilitated by the use of in vitro expanded T-cell cultures. We were also unable to unequivocally identify the HLA class II restricting molecules beyond the tissue typing of vaccine recipients and performing binding predictions using the ELF algorithm. Finally, it would have been interesting to investigate the clonality of the CD4 + T cells responding to individual peptide clusters or whether CD4 + T cells display generally higher or lower clonality compared to CD8 + T cells. Regardless of these limitations, our effort contributes to the collection of CD4 + T-cell determinants in HIV-1, which will drive iterative vaccine improvements especially in the light of the increasingly recognized important roles of CD4 + T cells in the generation of protective anti-HIV-1 responses.

Conflicts of Interest:
The authors declare no conflict of interest other than T Hanke is a coinventor on patents WO06123156, WO98/056919, PCT/US2014/058422, and EP14846993.5. C Brander is cofounder and CSO of AELIX Therapeutics. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.