Subjects and PBMC isolation

Subjects were drawn from a large prospective cohort of 613 volunteers living with HIV to evaluate clinical, laboratory, immunologic and viral markers of disease progression (IAVI protocol C) [26]. From this cohort, 13 volunteers from clinics in Uganda, Rwanda, Zambia and South Africa were selected based on being i) apparently healthy in clinical latency, ii) anti-retroviral (ARV) therapy naïve and iii) willing to donate up to 100mL of blood at an additional visit which ranged at between 1203 and 5319 days post estimated date of HIV-1 acquisition (EDA) (Table 1). Blood samples from HIV-1 uninfected South African blood transfusion volunteers were also obtained through Contract Laboratory Services Ltd., Johannesburg, South Africa. Peripheral blood mononuclear cells (PBMC) were isolated by density centrifugation, frozen at 10 million viable PBMC per vial in 10% V/V dimethyl sulphoxide (DMSO) in heat-inactivated foetal calf serum (HIFCS) (all Merck Life Science Ltd., UK) and stored in vapour phase liquid nitrogen.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Information for subjects living with HIV.

https://doi.org/10.1371/journal.pone.0260118.t001

PTE peptides and mapping strategy

Four pools of HIV-1 Gag, Nef, Env and Pol PTE 15mer peptides and 1408 individual peptides were obtained through the NIH HIV Reagent Program, Division of AIDS, NIAID, NIH, Germantown, USA, contributed by DAIDS/NIAID. Peptides were dissolved in DMSO such that each peptide once pooled was at a concentration of 222μg/mL and stored at -80°C. Individual peptides were combined into a total of 174 pools of 18 to 26 peptides per pool such that each individual peptide was present in three different peptide pools: a 3-dimensional (3-D) peptide matrix. Matrix pools remained separated into matrices for each of the four HIV-1 proteins (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. PTE peptide sets and peptide pool details.

Sets of individual 15mer peptides and pools of peptides for HIV-1 Gag, Nef, Env and Pol proteins were supplied by the NIH HIV Reagent program, USA.

https://doi.org/10.1371/journal.pone.0260118.t002

Peptides were diluted in RPMI medium with 10% V/V heat-inactivated calf serum (HIFCS) (R10) and used at a final ELISpot assay concentration of 1μg/mL per peptide and <0.5% V/V DMSO. Duplicate peptide pools and quadruplicate DMSO/R10 only wells as negative controls (mock) were stored frozen in 96 well deep well plates in the arrangement required for IFNγ ELISpot assay. Phytohemagglutinin (PHA) (Merck Life Science Ltd., UK) as a positive control was prepared and added at the point of ELISpot assay set up at 10μg/mL final assay concentration.

Expansion of CD8 T-cells and IFNγ ELISpot assay

CD8 T-cells were expanded from PBMC and IFNγ ELISpot assays performed using a method similar to that described elsewhere [24]. Typically, 6 to 12 million thawed viable PBMC were resuspended in 10mL R10 with 50 units IL-2/mL (R10/50) and 0.5μg/mL CD3/CD4 bispecific antibody (kindly provided by Professor Johnson Wong, Harvard Medical School, USA). Cultures were doubled in volume with R10/50 at days 2 and 4, transferring cultures to additional flasks.

On day 7, approximately 2/3 of or up to 100 million cells were removed, washed 3 times with 20mL PBS and resuspended in RPMI medium with 20% HIFCS (R20) without IL-2 at half the removed culture volume. Cells were rested for 22 to 26 hrs and IFNγ ELISpot assay conducted on day 8 as described previously [27, 28]. In brief, 100μL mock (DMSO/R10), 3-D matrix HIV-1 PTE peptide pools and PHA were added to IFNγ ELISpot plates (Mabtech AB, Sweden) followed by 50μL cells (1 to 2x10⁵ cells/well). Following 16 to 24 hours incubation (day 9), spots were developed by sequential incubation with biotinylated anti-human IFNγ antibody (Mabtech AB, Sweden), Streptavidin/biotin/peroxidase complex (Life Technologies Ltd., UK) and finally AEC/buffer peroxidase substrate (Merck Life Science Ltd., UK), with wells washed with PBS/0.5% v/v tween-20 (Merck Life Science Ltd., UK) between steps. Spots were enumerated using an ELISpot reader (AID Autoimmun Diagnostika, Germany). During days 10 and 11, individual peptides to test from the complete 1408 peptide set were identified by deconvolution of the responding peptide matrix pools and diluted in R10 in the required 96 well plate format. Peptides were added to wells of IFNγ ELISpot plates on day 11.

Also, on days 7 and 9, the remaining 1/3 of culture was doubled in volume with R10/50. On day 10, cultures were recovered, washed three times with PBS and rested as above. On day 11, cells were recovered, counted and added to ELISpot wells as before to assess CD8 T-cell responses to selected individual PTE peptides. Any remaining cells were resuspended in chilled (+2 to +8°C) 10% DMSO/90% HIFCS and 1 mL aliquots with up to 30x10⁶ cells per vial were cryopreserved in liquid nitrogen. Frozen cells were thawed and washed in R10 medium, rested overnight in R20 medium before testing in IFNγ ELISpot assay.

For an IFNγ ELISpot assay to be considered valid, the mean of mock wells must be ≤50 and for PHA >500 spot forming units (SFU) per million cells. A response to a peptide or peptide pool stimulus was considered positive if the mean of replicate wells was ≥4 times the mean mock response and the mean mock-subtracted response was ≥38 SFU per million cells [24, 27, 28].

A list of well-defined HIV-1 CD8 T-cell epitopes was acquired from the LANL HIV immunology database: http://www.hiv.lanl.gov/content/immunology (HIV Sequence Compendium 2018). This list, often referred to as the “A-list” consists of 281 epitopes across HIV-1 proteins and includes the sequence of each epitope and its location on the HIV-1 HXB2 protein sequence [29]. All sequences for peptide responses identified in the present study and the A-list peptides were collated and aligned against the HXB2 protein sequence. Alignments were performed using the protein sequence alignment tool from: https://www.ncbi.nlm.nih.gov/ (National Center for Biotechnology Information (NCBI)).

Isolation of CD8 T-cells from PBMC and IFNγ ELISpot assay

To compare ELISpot responses of expanded CD8 T-cells and CD8 T-cells isolated directly from PBMC without expansion, 6 of the 13 subjects living with HIV were selected (Table 1 subjects 1, 6, 8, 9, 10 & 13) based on additional frozen PBMC vial availability and positive responses to Nef peptide pools. CD8 T-cells were expanded from PBMC for 10 days and washed as before. Based on prior known PBMC recoveries per vial, 4 to 6 vials from each of these 6 subjects were then thawed and both expanded CD8 T-cells and PBMC were rested for 23 hrs in R20. So that all CD8 T-cells underwent similar conditions just prior to ELISpot assay, CD8 T-cells were isolated from all cell cultures by negative selection using a CD8 T-cell isolation kit (Miltenyi Biotech Ltd., UK) following manufacturer’s instructions. Cells were tested in IFNγ ELISpot assay as before, with 200,000 cells per well with one of the following stimuli; mock, PHA, a pool of 138 15mer peptides overlapping by 11 amino acids matched to the Cytomegalovirus (CMV) pp65 protein (Genscript Biotech, Netherlands) or the 7 Nef peptide pools of the 1^st dimension of the Nef matrix. Peptides were again used at a final assay concentration of 1μg/mL per peptide. Nef was selected as the PTE set with the fewest peptides, a full Nef and/or other HIV-1 protein matrix ELISpot assay could not be conducted due to the limits on frozen PBMC vial availability and CD8 T-cell recovery from PBMC.

Flow cytometric analyses of expanded and PBMC CD8 T-cells

In a separate study [30] of HIV-1 pathogenesis currently in progress within the IAVI protocol C cohort, IFNγ production by CD8 T-cells in response to CMVpp65 or HIV-1 Gag, Nef, Env or Pol PTE peptide pool stimuli was assessed by intra-cellular cytokine flow cytometry. Both thawed and overnight rested PBMC and 7-day expanded, washed and rested CD8 T-cells from the same subject samples (n = 28) were assessed for % CD3+CD8+ T-cells expressing intra-cellular IFNγ.

Thawed and overnight rested PMBCs and expanded and rested CD8 T-cells were washed in R10 media and 0.75–1.0 1x10⁶ cells were stimulated for 6 hours with HIV-1 Gag, Nef, Env, Pol PTE or CMVpp65 peptide set pools along with mock and positive control Staphylococcal enterotoxin B, (SEB, Merck Life Science Ltd., UK) in the presence of 12.5μg/mL Brefeldin A (Merck Life Science Ltd., UK) and 0.002mM monensin (BioLegend Ltd., UK). Cells were washed in PBS, stained with LIVE/DEAD Fixable Near-IR dead cell stain (Life Technologies Ltd., UK) according to manufacturer’s instructions, washed with 2% V/V HIFCS in PBS and incubated with anti-CD3 SK7, anti-CD8 RPA-T8, anti-CD4 SK3 and anti-CD19 HIB19 (Becton Dickinson, San Jose, USA) monoclonal antibodies at 4°C in the dark for 30min. Cells were then permeabilised with BD Cytofix/Cytoperm kit (Beckton Dickinson, USA) according to the manufacturer’s instructions and incubated with anti-IFNγ B27 antibody (Becton Dickinson, USA) at 4°C in the dark for 30 minutes and washed. S1 Table provides details of antibodies and dilutions used. Cells were analysed using a FACSymphony flow cytometer and data were analyzed and presented using FlowJo software version 10.6 (Becton Dickinson, USA).

HIV-1 proviral plasmid construction: IRES constructs

Generation of HIV-1 IMC [21, 31] and the optimisation of replicating Renilla luciferase reporter HIV-1 utilising a modified encephalomyocarditis virus (EMCV) internal ribosome entry site (IRES) element, collectively referred to as IMC-LucR.6ATRi [22, 23] has been described previously. All plasmids were generously provided by Dr Christina Ochsenbauer, University of Alabama, Birmingham, USA (Table 3). All but virus number 4 (NL-LucR) were derived from subjects living with recently acquired HIV-1 (transmitted-founder isolates).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Details of LucR.6ATRi IMC reporter viruses tested in the viral inhibition assay.

https://doi.org/10.1371/journal.pone.0260118.t003

Phylogenetic analysis of HIV-1 isolates

Single genome near full length nucleotide sequences for each viral gene (Gag, Pol, Vif, Vpr, Vpu, Tat, Rev, Env and Nef) were translated to amino acid sequences using the Los Alamos National Laboratory genecutter tool (https://www.hiv.lanl.gov/content/sequence/GENE_CUTTER/cutter.html). The translated proteomes were concatenated and aligned on Geneious Prime version 2020.0.3 followed by phylogenetic analysis in MEGA version X [32, 33]. The initial phylogenetic tree for the heuristic search was obtained by applying the Maximum Parsimony method after which the evolutionary history was inferred by using the Maximum Likelihood method and General Reverse Transcriptase model that allowed for some sites to be evolutionarily invariable [34]. 100 permutations were performed and a tree with the highest log likelihood selected. The tree was drawn to scale, with branch lengths measured in the number of substitutions per site. Pairwise distances between pairs of sequences were derived using the General Reverse Transcriptase model involving 222 amino acid sequences and a total of 3429 positions.

Cell lines and generation of LucR IMC

HEK 293T/17 cells were obtained directly from the American Type Culture Collection (ATCC, Manassas, VA, USA) (CRL-11268). TZM-bl dual-reporter cell line was obtained directly from the NIH HIV Reagent program, Division of AIDS, NIAID, NIH: TZM-bl Cells, ARP-8129, contributed by Dr. John C. Kappes and Dr. Xiaoyun Wu. Both cell lines were maintained in Dulbecco’s modified Eagle’s medium (DMEM) with 10% V/V HIFCS, 10 mM hepes, 100 U/mL penicillin and streptomycin (all Merck Life Science Ltd., UK) at 37°C and 5% CO₂.

Virus stocks of HIV-1-infectious molecular clones were generated by proviral DNA transfection of 293T/17 cells as previously described [21, 22, 31, 35]. Briefly, 5 × 10⁶ 293T/17 cells were seeded in a T75 culture flask one day prior to transfection. 12μg of DNA in DMEM was transfected using FuGENE 6 (Promega Ltd., UK), according to manufacturer’s instructions. After 6 hr, transfection medium was replaced with fresh medium. Viral supernatants were harvested from 293T/17 cells at 48 hrs, supernatants clarified at 250g/10 min, 0.45μm filtered and frozen at -80°C.

The fifty-percent Tissue-Culture Infectious Dose (TCID₅₀) per mL of virus stocks were determined as previously described [35]. Serial dilutions of viruses were added to flat bottom 96 well plates in quadruplicate followed by 100,000 TZM-bl cells in the presence of 10μg DEAE-dextran (Merck Life Science Ltd., UK) per mL. Plates were incubated at 37°C and 5% CO₂ for 48 hr. 100μL culture medium was carefully removed from each well and replaced with luciferase reporter gene assay system reagent (BriteLite Plus Assay System, PerkinElmer Ltd., UK). After 3 min incubation, luminescence was measured using a Tecan Infinite^® reader (Tecan Ltd., Switzerland).

LucR IMC viral inhibition assay

CD8 T-cells were polyclonally expanded from PBMC as above for 7 days. CD4 T-cells were expanded in a similar manner with 0.5 μg/mL CD3/CD8 bispecific antibody and IL-2 (Professor Johnson Wong, Harvard Medical School, USA). Cultures volumes were doubled at days 3 and 6 with R10/50.

On day 7, 1x10⁶ viable CD4 T-cells were centrifuged in 15mL tubes at 250g/10 min and supernatants decanted with most of the residual volume aspirated. Cells were resuspended in the remaining residual volume and one of ten HIV-1 LucR.6ATRi IMC viruses added at a multiplicity of infection (MOI) of 0.1 (Table 3). Cells and virus were spinoculated at 1800g for 2 hours at ambient temperature and cells resuspended at 1x10⁶ cells/mL in R10/50. 100 μL CD4 T-cells (100,000 cells) were placed in quadruplicate in 96-well flat bottom white plates. On the same day, expanded CD8 T-cell cells were resuspended at 1x10⁶ cells/mL in fresh R10/50 and cultured in 24-well culture plates. Both the CD8 and infected CD4 T-cells were incubated in parallel for 3 days, when CD8 T-cells were recovered from wells, centrifuged, supernatants decanted and cells resuspended in an equal volume of R10/50. 100μL CD8 T-cells or R10/50 media were added to the infected CD4 T-cells in duplicate and incubated for an additional 5 days. On day 8 post-CD4 T-cell infection, 100 μL of supernatant were carefully removed from wells and 100 μL Renilla-Glo^™ Luciferase Assay System (Promega Ltd., UK) added. Plates were incubated for 3 min at room temperature, followed by measurement of luminescence activity using a Tecan^® Infinite reader. Luciferase activity was determined as a measure of viral infection and replication measured in relative light units (RLU). CD8 T-cell-mediated inhibition was expressed as reduction in log₁₀ RLU of CD4/CD8 T-cell co-cultures, compared with cultures of infected CD4 T-cells alone. A value of ≥ 0.8 log₁₀ inhibition was considered positive [20].

Assessment of HIV-1 replication in expanded CD8 T-cells from subjects living with HIV

In the separate study of HIV-1 pathogenesis currently in progress within the IAVI protocol C cohort [30], culture supernatants were sampled from day 10 expanded CD8 T-cells of 71 ARV-naïve subjects living with HIV (median pVL 13600/mL, range <39 to 1.46x10⁶, IQR 1995 to 92500). In order to assess possible productive replication of subjects’ own HIV-1 in expanded CD8 T-cells, supernatant p24 gag protein contents were determined by commercial ELISA following the supplier provided protocol (Perkin Elmer Ltd., UK, p24 ELISA NEK050). As a comparison with cultures where productive HIV-1 replication is expected, day 13 culture supernatants were included from 7 day expanded CD4 T-cells from a HIV-1 uninfected subject that were infected in vitro with one of three HIV-1 isolates at a MOI of 0.01. The isolates were not HIV-1 LucR molecular constructs but were analogous to the NL, CH077 and 247F_V2 constructs described in Table 3.

Statistics

Data were analysed using GraphPad Prism version 8.4.3 for Windows, GraphPad Software, San Diego, USA.

A two-tailed non-parametric Wilcoxon matched pairs signed rank test was used to determine significant differences between two sets of paired values. Non-parametric Kruskal-Wallace test was used to determine significant differences within multiple data groups followed by Dunn’s multiple comparison to identify differences between individual groups. A two-tailed non-parametric Spearman test was used to compute correlation coefficients between two variables. Raw data values were unadjusted for all statistical analyses. The threshold for significance was defined as p < 0.05. So as not to exclude datapoints from display that include a true value of 0, where data are displayed in figures on a log₁₀ scale, IFNγ ELISpot response values of less than 1 SFU per 10⁶ cells were assigned a value of 1 and IFNγ flow cytometry responses of less than 0.01% of CD8 T-cells were assigned a value of 0.01.

Ethics

Work was approved by the local ethics review boards, including the Rwanda National Ethics Committee, the Uganda Virus Research Institute Science and Ethics Committee (Currently the UVRI Research Ethics Committee) and the Uganda National Council of Science and Technology, the University of Cape Town Health Science Research and Ethics Committee, the University of Zambia Research Ethics Committee and the Emory University Institutional Review Board. Written informed consent was obtained from all participants.

Polyclonal CD8 T-cell expansion

For the 13 subjects living with HIV, starting with 1 vial of cryopreserved PBMC (median 9.4, minimum 4.2 million PBMC), 7-day bi-specific antibody expansion resulted in median 121.4 million viable cells (Table 4). For 12 of these subjects, approximately 2/3 of the 7-day cells were removed, washed and rested for >22 hours. For subject 7 with the fewest expanded cells (38 million), all 7-day cells were washed and rested. Median viable cell numbers following resting were 98.5% of those removed on day 7. This provided sufficient cells in all 13 subjects for the first 3-D matrix peptide pool ELISpot requiring 38 to 76 million cells for 100,000 to 200,000 cells per ELISpot well in duplicate. The remaining 1/3 of day 7 cells (median 44.5 million) were cultured for a further 3 days and yielded median 97.7 million viable cells on day 11 following washing and >22 hours resting. These cells were tested in ELISpot with single peptides, selected based on responses to the first 3-D matrix ELISpot. CD8 T-cells were expanded from a second PBMC vial for single peptide ELISpot for subject 7 with insufficient cells expanded on day 7. For 7 HIV-1 seronegative subjects a median 12.1 million starting PBMC resulted in median 176.2 million cells after 7 days of expansion. Cells from these subjects were not assessed in day 11 single peptide ELISpot due to lack of any responses to matrix pools.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. CD8 T-cell expansion and number of PTE peptides and epitope regions recognised.

https://doi.org/10.1371/journal.pone.0260118.t004

PTE peptide IFNγ ELISpot responses

CD8 T-cells removed directly from bi-specific antibody/IL-2 culture without washing and resting are highly activated and result in high IFNγ ELISpot magnitudes without further stimulation. Across all 13 subjects living with HIV and 7 seronegative subjects, minimal release of IFNγ was detected in ELISpot assay of washed and rested expanded CD8 T-cells in the absence of further stimulation (maximum 7.5 SFU per million cells with mock stimulus). CD8 T-cells were functional in all subjects demonstrated by minimum PHA responses of 1235 SFU per million cells. No responses to the 1408 PTE peptides arranged in 174 3-D matrix pools were detected in CD8 T-cells expanded from 7 HIV-1 seronegative subjects (highest 15 SFU per million cells, mock-subtracted across 1218 matrix pool stimulations in duplicate). Responses to HIV-1 matrix PTE peptide pools were detected in all 13 subjects living with HIV. These were deconvoluted to narrow down the possible peptides recognised with actual recognised peptides identified in the second ELISpot on day 11.

A median of 197 (26 to 480) single PTE peptides were tested in the second ELISpot and a total of 462 PTE peptide responses (309 individual PTE peptides) were identified with a median of 33 peptide responses (5 to 65) per subject (Table 4 and S2 Table). The number of peptides recognised across the HIV-1 proteins tested was 126 Gag (27.3% of total), 103 Nef (22.3%), 54 Env (11.7%) and 179 Pol (38.9%). A single epitope region may account for more than one observed peptide response when variants of the same peptide are recognised or the amino acid sequences of two or more recognised peptides overlap [36]. Recognised PTE peptides were arranged by amino acid location relative to HXB2 protein sequences and were assigned to one epitope region where peptide sequences overlapped by 8 or more amino acids. As a result, the subjects’ median 33 PTE peptides recognised could be explained by recognition of a median of at least 16 (3 to 41) epitope regions.

Alignment of peptide responses against a well-defined list of CD8 T-cell epitopes referred to as the A-list (HIV Sequence Compendium 2018) demonstrated that the responses detected using this expansion methodology are in line with responses reported in other studies. Of the 462 15mer peptides recognised, 114 (90.5%) of 126 Gag, 100 (97.1%) of 103 Nef, 40 (74.1%) of 54 Env and 130 (72.6%) of 179 Pol peptides overlapped by at least 8 amino acids with the HXB2 sequence location of at least one A-list epitope. Overall 384 (83.1%) of the 462 peptides recognised were located in previously defined optimal CD8 T-cell epitope regions (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Locations of PTE peptides recognised compared with A-list epitopes across HIV-1 proteins.

The x-axes represent HXB2 HIV-1 protein amino acid sequence position for each of the four represented proteins. The upper y-axes display the number of times an amino acid at each HXB2 amino acid sequence position is present in peptides recognised by the present study subjects (open area) compared with the A-list of known epitopes (shaded area). The lower y-axes display the study subjects and locations of peptide responses (black dash) as well as A-list epitope locations (shaded dash). (A) HIV-1 Gag protein (B) HIV-1 Nef protein (C) HIV-1 Pol protein (D) HIV-1 Env protein.

https://doi.org/10.1371/journal.pone.0260118.g001

Subject 8, tested at 48 months post estimated acquisition, was selected for further assessment of peptide recognition at an earlier time point based on availability of PBMC vials, a relatively high number of peptides recognised (48) and apparent control of HIV-1 replication in vivo evidenced by sustained low plasma viral loads (maximum 1035/mL) and stable CD4 counts (minimum 472/μL) at time points following 1 year of estimated date of acquisition. Table 5 demonstrates that CD8 T-cells from this subject recognised 32 PTE peptides at 30 months with 19 Gag, 9 Pol and 4 Nef peptides. All 32 peptides were again recognised at 48 months with an additional 16 peptides: 10 Gag, 1 Nef, 3 Pol and 2 Env. 10 were variants of peptides recognised at month 30 with 9 of these being Gag peptides. Env peptides were not recognised at month 30, with responses to 2 peptides apparent at month 48. As well as a broadening of the peptides recognised, the median response magnitudes for each peptide increased from 285 (interquartile range (IQR) 103–482) to 857 (IQR 501–1475) SFU per million cells for the 32 peptides recognised at both time points, with median magnitude of 487 (IQR 249–1091) for all peptides recognised at month 48.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. PTE peptide recognition by subject 8 at two post HIV-1 acquisition time points.

https://doi.org/10.1371/journal.pone.0260118.t005

IFNγ responses of unexpanded and expanded CD8 T-cells

Thawing of 4 to 6 PBMC vials per subject living with HIV (n = 6) provided 16.4 to 37.5x10⁶ PBMC after 23 hours rest. Negative selection led to 6.5 to 14.5x10⁶ CD8 T-cells (median 9.9x10⁶ cells or 36.9% of recovered PBMC), sufficient for ELISpot analyses of a limited number of CMVpp65 and 1^st dimension matrix Nef peptide pools, rather than all 3 dimensions. IFNγ ELISpot responses (SFU/10⁶ cells) to CMVpp65 or Nef matrix peptide pools were similar between 10-day expanded CD8 T-cells and CD8 T-cells isolated directly from PBMC (medians 158 and 130 SFU/10⁶ cells, respectively, p = 0.427) (Fig 2A) with a significant and strong positive correlation (r = 0.814, p <0.001) (Fig 2B). Similar findings were apparent in a separate study group when PBMC and expanded CD8 T-cells were assessed by intra-cellular cytokine flow cytometry following stimulation with CMVpp65 or HIV-1 Gag, Nef, Pol and Env PTE peptide pools. %IFNγ+ CD8 T-cells were similar between expanded CD8 T-cells and CD8 T-cells isolated directly from PBMC (medians 0.417 and 0.347%, respectively, p = 0.314) (Fig 3A) with a significant positive correlation (r = 0.691, p <0.001) (Fig 3B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. IFNγ ELISpot responses to CMVpp65 or HIV-1 Nef peptide pools in expanded CD8 T-cells or CD8 T-cells isolated from PBMC.

A) SFU per million cells with medians (bar) and interquartile ranges and B) correlation between data sets with a non-parametric Spearman test used to compute the correlation coefficient (r) between the two the datasets. For log₁₀ scale display only, values of <1 SFU were assigned a value of 1. Statistical analyses were conducted using unaltered data.

https://doi.org/10.1371/journal.pone.0260118.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Flow cytometric analyses of IFNγ+ events of CD3+CD8+ T-cells stimulated with CMVpp65 or HIV-1 PTE peptide pools in expanded CD8 T-cells or PBMC.

A) percent IFNγ+ events with medians (bar) and interquartile ranges and B) correlation between datasets with a non-parametric Spearman test used to compute the correlation coefficient (r) between the two datasets. For log₁₀ scale display only, values of <0.01 SFU were assigned a value of 0.01. Statistical analyses were conducted using unaltered data.

https://doi.org/10.1371/journal.pone.0260118.g003

Expanded/Cryopreserved CD8 T-cells

Day 10 expanded and rested CD8 T-cells were cryopreserved in liquid nitrogen vapour. There was no significant difference in ELISpot responses of these cells to HIV-1 Gag, Nef, Pol or Env peptide pools following thaw and overnight rest compared with expanded CD8 T-cells tested prior to cryopreservation (p = 0.854) with a significant positive correlation (r = 0.853, p <0.001) (S1 Fig).

LucR IMC replication in expanded CD4 T-cells and inhibition by autologous CD8 T-cells

All 10 IMC replicated in expanded CD4 T-cells from all subjects with log₁₀ RLU values between 3.5 and 5.6 over media only background. Inhibition of IMC replication in CD4 T-cells above the positive cut-off of 0.8 log₁₀ reduction in RLU values was evident following co-culture with CD8 T-cells from all subjects with median of 7 IMC inhibited (range 4 to 10) (Fig 4A and S3 Table). CD8 T-cells from two subjects living with HIV-1 clade C (6 and 9) inhibited all ten IMC. Arranging data by both subject clade and test HIV-1 IMC clade indicated that for subjects living with the same HIV-1 clade there was no significant difference in inhibition of test HIV-1 IMC of different clades (within clade A subjects: p = 0.167, clade C: p = 0.920 or clade D: p = 0.182) (Fig 4B). Grouping all inhibition values by subject HIV-1 clade demonstrated that 66%, 78.3% and 40% of IMC were inhibited above the 0.8 log₁₀ positive inhibition value by subjects living with HIV-1 clade A, clade C or clade D, respectively. Log₁₀ Inhibition values were significantly higher for subjects living with clade A (median 0.93, p = 0.027) and C (median 1.17, p < 0.001) compared with clade D (median 0.69). Inhibition values for subjects living with HIV-1 clade C were marginally higher than clade A, though not significantly (p = 0.054) (Fig 4B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. CD8 T-cell mediated inhibition of HIV-1 replication in autologous CD4 T-cells.

A) Inhibition for each subject living with HIV, arranged by subject HIV-1 clade, B) Inhibition arranged by subject HIV-1 clade and test HIV-1 LucR IMC clade. Each of the 3 left panels, inhibition values for test viruses of the same clade are grouped together for subjects living with the same HIV-1 clade (A, C or D clades). In the right panel, inhibition values for all test viruses are grouped by subject HIV-1 clade. Bars represent the medians with interquartile ranges, the dotted line represents the assay positivity value (>0.8 log10).

https://doi.org/10.1371/journal.pone.0260118.g004

Peptide recognition and HIV-1 inhibition breadth

There was a significant positive correlation between the number of HIV-1 isolates inhibited and the number of peptides recognised by subjects, including variants of the same peptide (r = 0.664, p = 0.016) (Fig 5A). There was also a positive but not significant correlation between the breadth of inhibition and the minimum number of epitope regions recognised by CD8 T-cells (r = 0.483, p = 0.096) (Fig 5B). Positive but not significant correlations were also apparent between inhibition breadth and the number of Gag (r = 0.355, p = 0.231), Nef (r = 0.503, p = 0.082), Env (r = 0.378, p = 0.201) and Pol (r = 0.420, p = 0.53) peptides recognised.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Correlations between the number of HIV-1 isolates inhibited and A) the total number of PTE peptides recognised and B) the total number of PTE epitope regions recognised.

A non-parametric Spearman test was used to compute the correlation coefficients (r) between the datasets.

https://doi.org/10.1371/journal.pone.0260118.g005

Phylogenetic analysis of viruses in the VIA panel

The IMC were derived from the major nodes in HIV-1 subtypes A, B, C and D including one subtype A1/D recombinant. As most subjects from which these sequences were derived acquired HIV-1 subtype A, C, D or unique recombinant forms of these subtypes, there was insufficient representation of subtype B sequences to evaluate the breadth of representation of the subtype B IMC. The selected IMC also represent a diversity of distances from the protocol C consensus sequence and are therefore representative of the transmitted/founder (T/F) variants in this cohort (S2 Fig).

Assessment of HIV-1 replication in expanded CD8 T-cells from subjects living with HIV

Day 13 cultures of expanded CD4 T-cells infected in vitro with HIV-1 isolates demonstrated HIV-1 replication with release of HIV-1 p24 Gag protein into culture supernatants of between 1.3x10⁵ and 3.2x10⁵ pg/mL. Minimal release of p24 Gag protein that would be derived from subjects’ own acquired HIV-1 was detected in day 10 expanded CD8 T-cells from 71 ARV-naïve subjects living with HIV, where no values were above the assay positive cut-off value of 5.8pg/mL (mean OD₄₉₀ of media only wells + 0.05). 12 of 71 supernatants did record OD₄₉₀ values below the supplier assay positive cut off value but above the media only wells, which would be equivalent to p24 Gag concentrations between 1.1 and 2.8pg/mL (S3 Fig).