The general plan of the study
Between February 2015 and May 2016, 10 chronically HIV-1-infected patients were included in treatment arms 5 and 6 of the SPARC-06 Clinical Trial (NCT02961829) to receive an autologous MDDCT without (group 5) or with (group 6) a conditioning regimen consisting of nicotinamide and auranofin (39). We enrolled males ≥ 18 years under suppressive ART with undetectable viral load for more than two years without virologic failure (Table 1). CD4 + T-cell counts at enrollment were above 500 cells/mm3, and CD4+ T-cell count nadir was above 350 cells/mm3. MDDCT was performed in the study candidates after the first 48 weeks of the intervention period in three doses at baseline (time zero), fifteen days after the first dose, and 30 days after the first dose. There were two significant deviations from the protocol (see Results section). All patients had undetectable viral load upon MDDCT administration except for the two protocol violators (P21 and P30, Table 1), which were viremic and not receiving ART during MDDCT administration.
To avoid any effect of auranofin and nicotinamide on MMDDC function, auranofin was interrupted 24 weeks before and nicotinamide immediately before the MDDCT.
We designed personalized peptides for pulsing of MDDCs, following a multi-step approach for each study candidate. As a first step, each patient’s HIV-1 gag sequence was characterized from peripheral blood mononuclear cell (PBMCs) DNA through the generation of genomic sequences after clonal amplification of viral strains through Next Generation Sequencing (NGS) and single genome amplification. For single genome amplification, at least ten clones were generated per patient, as previously described (44). In the second step, the HIV DNA sequences were translated to amino acid sequences. Briefly, the three frames of translation were aligned to the start of Gag in the HXB-2 reference sequence to determine the correct reading frame using Clustal-Omega (45). The polypeptides obtained were edited by manual correction, i.e. tryptophans were used to replace spurious stop codons in the middle of the sequence. Clustal-Omega was also used to create a consensus sequence. Additional alignments were then made using the published aligned sequences to map the highly conserved regions of the consensus Gag sequences from each study subject.
In parallel, HLA haplotypes of the same individuals were sequenced. HLA A, B, C, and DR typing was done by sequence-specific oligonucleotide methodology using a LUMINEX 200 apparatus. Briefly, DNA was extracted from the buffy coat of EDTA- anticoagulated peripheral blood with the Qiamp DNA mini kit (Qiagen) and subjected to One lambda LAbtype SSO (Thermo Fisher) according to the manufacturers’ instructions.
Finally, the selection of HIV-1 Gag epitopes was performed by designing peptides from the autologous HIV gag sequence and selecting those predicted to be recognized by each individual’s MHC Class I and Class II. The epitopes were predicted using the NetMHCpan v4.0 server (46), which displays an overall accumulated prediction ranking score of 49–78 (47)), as calculated between 2018 and 2021 by The Immune Epitope Database (IEDB) (48)). A number of peptides were selected in positions encompassing amino acids 256–377 of the Gag polypeptide. Those peptides containing more than one cysteine were discarded wherever possible to avoid internal disulfide bonds. The number of peptides designed for each candidate depended on the patients’ affinity for multiple HLA haplotypes. Therefore, the higher the number of HLAs predicted to be bound by each peptide, the lower the number of Gag peptides for each candidate. Also, the number of peptides was chosen to keep the immunizing stimuli to a minimum. For example, peptides predicted to bind to multiple HLA haplotypes were preferred over those recognized by a single epitope: to minimize the synthesis costs and to ensure the maximum level of sequence conservation, preference was given to those peptides recognized by the haplotypes of more than one individual. For individuals for whom at least two peptides theoretically binding with high-affinity to more than one of their haplotypes were not found, a higher number of peptides, including peptides outside those above mentioned highly conserved regions, were designed, in order to increase the likelihood of immune recognition. In this manner, we designed 2 to 6 peptides (9-mers) per candidate, as shown in Supplementary Table 1. After MDDCT administration, NetMHCpan predictions were reanalyzed using the newly developed Custommune software (49)), which is validated based on documented biological data from the Los Alamos database (50)). The results were used to stratify patients’ ex-post based on the concordance between software predictions (51).
An automatic desktop synthesizer (PSSM 8 of Shimadzu) was used for the simultaneous solid-phase synthesis of all peptides using the Fmoc procedure. The final peptides were “unprotected” in TFA and purified by HPLC semi-preparation using an Econosil C-18 column (10 µm, 22,5 x 250 mm) and a system of two solvents: (A) trifluoroacetic acid (TFA) / H2O (1: 1000) and (B) TFA / acetonitrile (ACN) / H2O (1: 900: 100). The column was eluted at a flow rate of 8 mL/min with a gradient of 0 to 80% of solvent B for 45 minutes. The HPLC analysis was done using a binary HPLC system made by Shimadzu with a UV-vis detector SPD-10AV (Shimadzu), coupled to an Ultrasphere C-18 column (5 µm, 4,6 x 150 mm) that was eluted with system solvents A1 (TFA / H2O, 1: 1000) and B1 (ACN / H2O / TFA, 900: 100: 1) at a flow rate of 1.0 mL/min and a gradient of 10–80% of B1 for 10 minutes. The eluates of the HPLC columns were monitored for their absorbance at 220 nm. The molecular weight and the purity of the synthesized proteins were verified by electron spray (LC / MS-2010 Shimadzu). The number of peptides was determined by analyzing the amino acids (Shimadzu).
The collection of MDDC precursor cells was carried out by leukapheresis with the cell separator Terumo Cobe Spectra at the Blood Center of São Paulo, São Paulo, Brazil. The total blood volume was calculated for each participant, and 1.5 total blood volumes were processed, in peripheral venous access, with continuous flow, at a speed of 50–60 mL/min. After the end of the collection, the leukapheresis product was sent to the Retrovirology Laboratory of the Federal University of Sao Paulo for the separation of monocytes and subsequent differentiation into DCs. Mononuclear cells from leukapheresis products were separated by Ficoll-Hypaque Premium (GE Healthcare® BioSciences, PA, USA) density gradient centrifugation and were cryopreserved in aliquots of 5 × 107 cells/mL in liquid nitrogen using fetal bovine serum (FBS; Gibco Life Technologies®, CA, USA) with 10% dimethyl sulfoxide (DMSO; Merck, Darmstadt, HE, DEU) until further assays were performed. Before use, cells were thawed at 37°C in a water bath and seeded at 5 × 106/mL in 175 cm2 tissue culture flasks (Corning® - Merck, Darmstadt, HE, DEU) in RPMI 1640 medium (Gibco Life Technologies) for two h at 37°C in a 5% CO2 incubator to obtain adherence-isolated monocytes. After incubation, non-adherent cells were removed by washing, and the remaining cells (predominantly monocytes) were differentiated in MDDCs. To this aim, we compared two different procedures to select and optimize the most efficient protocol. In particular, we compared a protocol using IL-4 initially and GM-CSF followed by TNF, IL-1β, and IL-6 (henceforth, IL-4 protocol; Supplementary Fig. 1A) with a protocol using IFN-α initially and GM-CSF, followed by LPS (henceforth, IFN-α protocol; Supplementary Fig. 1B). Briefly, in the IL-4 protocol, adherent cells were cultured in AIM-V medium (Therapeutic Grade – Gibco Life Technologies) in the presence of 50 ng/ml recombinant human Granulocyte-macrophage colony-stimulating factor (GM-CSF; Cell-Genix®, NH, USA) and 50 ng/ml recombinant human IL-4 (Cell-GenixR, NH, USA) for five days, so as to obtain immature MDDCs (iMDDCs). On day 5, iMDDCs were pulsed with a pool of personalized HIV peptides that was added to the cells (0.2 µg/mL each peptide) overnight. Later, the cells were washed to remove unbound peptide particles and were cultured for an additional two days in an AIM-V medium supplemented with the maturation cytokines IL-6 (100 ng/ml), IL-1β (10 ng/ml), and TNF (50 ng/ml), plus GM-CSF (50 ng/ml) and IL-4 (50 ng/ml; all from Cell-Genix), so as to obtain mature MDDCs (mMDDCs) pulsed with autologous peptides (Supplementary Fig. 1A).
In the IFN-α protocol, adherent cells were cultured in AIM-V medium to which GM-CSF and 500 IU/mL IFN-α-2b (Miltenyi Biotec, Auburn, CA, USA) were added on days 0 and 1 to obtain iMDDCs. On day 2, a pool of personalized HIV peptides was added to the cells (0.2 µg/mL each peptide) overnight. The following day, six h before cell harvest, maturation of iMDDCs was induced by adding 5 IU/mL lipopolysaccharide (LPS-SM, Ultrapure InvivoGen, San Diego, CA, USA). After the incubation period, mMDDCs were recovered on ice and washed three times with sodium chloride solution (0.9% NaCl, USP grade; Hospira, Lake Forest, IL) (Supplementary Fig. 1B).
Before reinfusion in patients, both iMDDCs and the final MDDCT were subjected to quality controls and immunophenotyping by flow cytometry analysis. For validation tests of DC maturation protocols, cell mortality was assessed using the cell viability kit (BD Biosciences) according to the manufacturer’s recommendations. For assessing the viability of MDDCs before MDDCT administration, the amine-reactive fixable LIVE/DEAD stain (Gibco Life Technologies, OR, USA) was used, by staining for 30 min at 4°C. For further staining, after thorough washing with phosphate-buffered saline (PBS), cells were incubated for 20 min at 4°C and assayed using the following monoclonal antibody panel: CD11c APC or V450 (clone B-ly6), CD14 Pacific Blue (clone M5E2), HLA-DR FITC (clone G46–6), CD80 PE (clone L307.4), CD86 FITC (clone 2331 FUN-1) and CCR7 PerCP (clone 150503). All the mAbs were obtained from BD Biosciences®, CA, USA, except CCR7 (R&D Systems®, MN, USA). Subsequently, cells were washed with FACS buffer (0.2% albumin and 0.1% sodium azide in PBS) and then fixed with 2% paraformaldehyde in PBS. Data were acquired on an LSR Fortessa Flow Cytometer (BD Biosciences) using the DIVA software, and the analysis was performed with FlowJo vX 0.7 software (Tree Star®, OR, USA).
MDDCT immunogenicity analysis
The effect of MDDCT on specific anti-HIV immune response in treated patients was assessed in patients’ PBMCs collected from total blood on days 0, 15, 30 (i.e., upon each dose of therapeutic immunization), and 120 (i.e., post-immunization follow-up). Total blood was drawn before each MDDCT dose to avoid confounding effects of MDDCT on each baseline. For this purpose, cells were plated at a concentration of 1x106/mL in 96-well round-bottomed plates in 200 µl of RPMI 1640 medium plus 10% FBS. Cells were either left unstimulated or incubated with 1 µg/mL of personalized peptides for 48h in an incubator at 37°C and 5% CO2. Six hours before cell harvest, the positive control wells (C+) were stimulated with 1 µg/mL staphylococcal enterotoxin B (SEB, Merck OR, USA). After one h stimulation with SEB, 20 µg/mL of Brefeldin A (BFA, Merck) was added to the plate to block any further protein transport. Background controls (C−) consisted of PBMCs cultured in the absence of peptides and were used as blank, i.e., subtracted from the percentage of cytokine-producing lymphocytes obtained from wells incubated with the peptides. To evaluate immune responses/activation, PBMCs were first fixed and permeabilized using the Cytofix/Cytoperm and Perm/Wash kits (BD Biosciences CA, USA), following the manufacturer’s recommendations. Cells were then incubated with a fixable live/dead stain (Gibco Life Technologies OR, USA) to check their viability and stained with a panel of antibodies including α-CD3 (V450; clone UCHT1), α-CD4 (BV605; clone RPA-T4), and α-CD8 (APC-H7; clone SK1) as well as antibodies against the intracellular cytokines IFN-γ (PerCP-Cy5.5; clone B27), TNF (PE-Cy7; clone MAb11), and IL-2 (FITC; clone MQ1–17H12) (all from BD Biosciences CA, USA). Data acquisition was performed on an LSR Fortessa Flow Cytometer, using the DIVA software, and the analysis was performed with FlowJo vX 0.7. All samples and controls were analyzed in triplicate.
Proviral DNA quantitation in Rectal biopsies
Viral DNA was measured as an estimate of the viral reservoir by qPCR as previously described. (52)(53)(54)
Comparisons between two groups were conducted by chi-square testing, relative risk analysis and non-parametric Wilcoxon test. Comparisons between more than two groups were conducted by one-way ANOVA, followed by Dunnet’s post-test, or by two-way ANOVA followed by Tukey’s post-test. When appropriate, a logit transformation was applied to the data to restore normality before the statistical test. To allow for paired statistical analysis of the differences, patients with a missing value in a time point were excluded from that specific statistical analysis, but all available values were included when plotting the overall data. All analyses were conducted using the Prism v.6 software (GraphPad® Software Inc, CA, USA).