Influence of HAART on alternative reading frame immune responses over the course of HIV-1 infection

Abstract

Background: Translational errors can result in bypassing of the main viral protein reading frames and the production of alternate reading frame (ARF) or cryptic peptides. Within HIV, there are many such ARFs in both sense and the antisense directions of transcription. These ARFs have the potential to generate immunogenic peptides called cryptic epitopes (CE). Both antiretroviral drug therapy and the immune system exert a mutational pressure on HIV-1. Immune pressure exerted by ARF CD8(+) T cells on the virus has already been observed in vitro. HAART has also been described to select HIV-1 variants for drug escape mutations. Since the mutational pressure exerted on one location of the HIV-1 genome can potentially affect the 3 reading frames, we hypothesized that ARF responses would be affected by this drug pressure in vivo.

Methodology/principal findings: In this study we identified new ARFs derived from sense and antisense transcription of HIV-1. Many of these ARFs are detectable in circulating viral proteins. They are predominantly found in the HIV-1 env nucleotide region. We measured T cell responses to 199 HIV-1 CE encoded within 13 sense and 34 antisense HIV-1 ARFs. We were able to observe that these ARF responses are more frequent and of greater magnitude in chronically infected individuals compared to acutely infected patients, and in patients on HAART, the breadth of ARF responses increased.

Conclusions/significance: These results have implications for vaccine design and unveil the existence of potential new epitopes that could be included as vaccine targets.

Figure 1. HIV-1 genome (HXB-2 strain), and localization of the 47 alternative reading frames (ARF).

Figure depicts all 199 ARF-tested peptides. These include 13 forward ARF within frames 1, 2 or 3 (in green) and 34 reverse ARF within frames -1, -2 or -3 (in purple). HIV-1 classically defined encoding genes are shown in blue.

Figure 2. Alternate reading frame-encoded amino acids in circulating viral sequences.

A) Distribution of known mutant origins across viral ARF sequences attributable to a particular cause based on information associated with the sequence accession the NCBI nr protein database presented in Table S2. B) Composition of viral coding sequence computed as a percentage assigned to the coding sequence for the Env, Gag, Pol poly-proteins based on nucleotide base counts for a particular gene region compared to the total nucleotide count for the structural genes of the virus. Gag comprises 1,503 nt of the total of 6,878 nt of structural gene sequence; Pol comprises 3,139 nt of the total; Env comprises of 2,571 nt of the total. This is the distribution of origins within the genome that would be expected if originating events for the incorporation of ARF and their detection in circulating HIV-1 viral sequences were distributed randomly throughout the genome. C) The distribution of ARF incorporated into circulating viral sequences that was observed in our searches of NCBI nr protein database for ARF sequences in circulating HIV-1 viral sequences. The percentages were computed by dividing the number of BLAST hits with ARF sequence incorporated into a given gene region by the 123 total hits examined. D) A three-way alignment between the HXB-2 reference sequence for the Env region, the accession AAL78125.1 and the alternate reading frame encoded ORF 67. E) A three-way alignment between the HXB-2 reference sequence for the Gag region, the accession AEQ21252.1 and the alternate reading frame encoded ORF 3. F) A three-way alignment between the HXB-2 reference sequence for the Pol region, the accession CAF29000.1 and the alternate reading frame encoded ORF 23. All three-way alignments were generated by combining two pair-wise alignments created in Geneious, followed by manual editing. Note each accession is similar to both the HXB-2 reference sequence for the structural proteins and the alternate reading frame encoded sequence, but not to both sequences simultaneously within the same region of the sequence.

Figure 3. Breadth of ARF responses in acutely infected patients.

A) Number of detectable responses observed for each individual ARF peptide-pool tested. B) Number of ARF peptide pools that induced detectable responses in each acutely infected individual. Blue bars represent patients On HAART and red bars represent patients Off HAART. We were unable to follow patients #26, #27 and #28 before HAART interruption.

Figure 4. Overview of ARF responses in acutely infected patients.

Out of 22 ARF peptide pools tested, 16 induced detectable responses. Each individual graph depicts the intensity of responses measured in an ELISPOT assay against IFN-γ (SFU/million PBMC) for each peptide in the corresponding patient. Blue bars correspond to patients On HAART and red bars to patients Off HAART. Responders were only considered if the net response against the peptide pool was >50 SFU (cut-off) over background and greater than twice the background.

Figure 5. Breadth of ARF responses in chronically infected patients.

A) Number of detectable responses observed for each individual ARF peptide-pool tested. B) Number of ARF peptide pools that induced detectable responses in each chronically infected individual. Grey bars represent patients Before HAART and black bars represent patients On HAART. Patients #47 and #48 were only tested before HAART introduction and patient #49 was only tested after HAART.

Figure 6. Overview of ARF responses in acutely infected patients.

Out of 22 ARF peptide pools tested, 18 induced detectable responses. Each individual graph depicts the intensity of responses measure in an ELISPOT assay against IFN-γ (SFU/million PBMC) for each peptide in the corresponding patient. Grey bars correspond to patients Before HAART and black bars to patients On HAART. Responders were only considered if the net response against the peptide pool was >50 SFU (cut-off) over background and greater than twice the background.

Figure 7. ARF responses in one HIV-1 chronically infected patients before and on HAART.

In the pie charts each color represents a pool of immune response that the patient mounted against. Numbers inside the pie charts correspond to the magnitude of the response against the pool in SFU/million PBMC.