Defective HIV-1 Proviruses Are Expressed and Can Be Recognized by Cytotoxic T Lymphocytes, which Shape the Proviral Landscape

Abstract

Despite antiretroviral therapy, HIV-1 persists in memory CD4⁺ T cells, creating a barrier to cure. The majority of HIV-1 proviruses are defective and considered clinically irrelevant. Using cells from HIV-1-infected individuals and reconstructed patient-derived defective proviruses, we show that defective proviruses can be transcribed into RNAs that are spliced and translated. Proviruses with defective major splice donors (MSDs) can activate novel splice sites to produce HIV-1 transcripts, and cells with these proviruses can be recognized by HIV-1-specific cytotoxic T lymphocytes (CTLs). Further, cells with proviruses containing lethal mutations upstream of CTL epitopes can also be recognized by CTLs, potentially through aberrant translation. Thus, CTLs may change the landscape of HIV-1 proviruses by preferentially targeting cells with specific types of defective proviruses. Additionally, the expression of defective proviruses will need to be considered in the measurement of HIV-1 latency reversal.

Keywords: APOBEC-mediated G-to-A hypermutations; HIV-1 cure; HIV-1 latent reservoir; HIV-1 proviral landscape; alternative splicing; cell-associated HIV-1 RNA; cold-target inhibition; cytotoxic T lymphocytes; defective HIV-1 proviruses; defective ribosomal product.

Figure 1. The landscape of HIV-1 proviruses may be dynamic

Correlation between the duration from diagnosis to HIV-1 proviral landscape analysis (enrollment) and the proportion of different subsets of HIV-1 proviruses containing (A) intact genome, (B) Ψ/MSD deletions/mutations, (C) hypermutations, (D) large internal deletions. r and P values are calculated by Pearson correlation. See also Figure S1.

Figure 2. HIV-1 proviruses can bypass MSD defects by activating novel SDSs and splicing into canonical SASs in vitro and ex vivo

(A) Spliced HIV-1 RNA transcripts from reconstructed patient-derived proviruses. Numbers on the right indicate the number of isolates from cloning of PCR products. (B–F) Sequences of canonical (Ocwieja et al., 2012; Schwartz et al., 1990a), known alternative (Ocwieja et al., 2012), and novel splice sites. (G) Spliced HIV-1 RNA from patient resting CD4⁺ T cells upon ex vivo activation. (H) Sequences of novel splice sites in participant 85. Purple box, short sequence repeats likely related to the 245 bp deletion encompassing MSD. See also Table S1–S5, Figures S2–S6.

Figure 3. HIV-1_hypermut can be transcribed ex vivo

Transcription of HIV-1_hypermut in vivo from rCD4s and upon ex vivo anti-CD3/CD28 activation. Resting CD4⁺ T cells from ART-suppressed individuals were treated with or without anti-CD3/CD28 activation for 3 days in the presence of enfuvirtide. Levels of non-hypermutated and hypermutated HIV-1 gag RNA were measured as the frequency of total HIV-1 gag (by qRT-PCR) times the proportion of hypermutated HIV-1 gag RNA (as measured by targeted gag deep sequencing). (A) Location of hypermutation hotspots (TGG) which will cause lethal missense start codon mutation or lethal nonsense mutations. (B–C) Proportion of hypermutated HIV-1 RNA. Green, nonhypermutated HIV-1 RNA. Red, hypermutated HIV-1 RNA containing lethal mutations. (D–E) Frequency of hypermutated HIV-1 RNA in resting and activated CD4⁺ T cells. Data represent mean ± SEM.

Figure 4. Cells containing HIV-1_def can be recognized by HIV-1-specific CTLs

Flow cytometry showing %degranulated CTLs (CD8⁺CD107a⁺) from 2 Gag-specific clones (A–B), 2 Nef-specific clones (C–D) and a CMV-specific clone (E) upon co-culture with autologous CD4⁺ T cells transfected with HIV-1_def plasmids. NL4-3 has an anchor residue mutation (K162Q) at the IK9 epitope making it not susceptible to CTL recognition (A). Each target was tested in independent quadruplets shown in mean ± SEM. P values were calculated by one-way ANOVA with Dunnett’s multiple comparison test. *P<0.05, **P<0.01, ***P<0.001, **** P<0.0001. See also Figure S7.

Figure 5. Cells containing HIV-1_def competes with cells containing intact proviruses for CTL recognition

(A) Negative control using CD4⁺ T cells transfected with the vector plasmid. (B) Positive control using CD4⁺ T cells transfected with the vector plasmid and loaded with cognate peptide. (C–E) Cold-target inhibition assay showing HIV-1-specific CTL recognition of cells containing HIV-1_def. Six paired replicates per condition were shown. P values were calculated by paired Student t-test. See also Figure S7.

Figure 6. Cells containing HIV-1_hypermut can be recognized by CD8⁺ T cells ex vivo

Resting CD4⁺ T cells were activated with anti-CD3/CD28 in the presence of enfuvirtide and co-cultured with autologous CD8⁺ T cells prestimulated with a Gag peptide mixture, a CMV peptide mixture or a HLA-A2-restricted SL9 peptide. Targeted gag RNA deep sequencing and qRT-PCR was used to measure the proportion and the frequency of HIV-1 RNA containing lethal hypermutations. (A–D) Proportion (A–C) and frequency (D) of HIV-1 RNA containing lethal hypermutations from CD4⁺ T cells upon co-culture with autologous CMV- (A), Gag- (B), SL9-(C) prestimulated CD8⁺ T cells. (E–G) Proportion of HIV-1 RNA with or without lethal hypermutations 5′ to wild-type or escaped CTL epitopes upon co-culture with autologous CMV- (E), Gag- (F), and SL9- (G) prestimulated CD8⁺ T cells. Data represent mean ± SEM. *, P value <0.05 by two-tailed Student t-test.

Figure 7. Proposed model of the scope of potential targets of HIV-1-specific CTLs compared with the size of the latent reservoir

The provirus population includes intact (yellow and pink) and defective (blue) proviruses as described (Ho et al., 2013). Following reversal of latency by endogenous stimuli or latency reversing agents, cells with certain types of HIV-1_def may express HIV-1 RNA and protein and may become targets for HIV-1-specific CTLs.